Interactive engineering simulators and learning tools. Pick an app, design the system, simulate its behaviour and learn the theory — all in the browser.
Classified unit conversion, always at hand — results update live as you type.
Classified unit conversion, always at hand — results update live as you type.
Design, validate and simulate hydraulic circuits with ISO 1219 symbols. Includes a full power pack builder, cartridge and proportional valves, live pressure/flow computation and 60+ ready examples.
Build pneumatic circuits with compressors, FRL units, 3/2–5/3 valves, cylinders and Festo-style electropneumatic control. Validate, simulate and animate the circuit in real time.
Feature-based 3D part design in the browser — CATIA-style specification tree, multi-profile sketches (also on any model face), pads, pockets, holes, fillets, measure & section tools, ISO drawing + STL export. No install, and a full CAD course in the Learning Hub.
IEC 60617 schematic editor + live DC simulation in the browser: Kirchhoff/Ohm solved on every change, animated current flow, voltage-coloured nets, meters on any point, relays, diodes, LEDs, op-amps, a tested sample-circuit gallery and an IC pinout library. No install.
Design and test planar mechanisms in a dark studio UI: links, gears, cams, sliders, springs, dampers, hydraulic jacks and motors with real kinematic/dynamic/static multibody simulation (semi-implicit Euler + sequential-impulse constraint solver), joint-reaction readouts, coupler traces, 2D drawing plus an orbitable 3D view, and a 3-level library of 49 solver-tested sample mechanisms from four-bars to full suspension corners.
Why a small force can lift tons: pressure in a confined fluid and the force-multiplying piston principle.
Hydraulics transmits force and motion through a confined, nearly incompressible fluid — usually mineral oil. The whole technology rests on Pascal's law (1653): pressure applied to a confined fluid is transmitted undiminished in all directions and acts with equal force on equal areas, at right angles to them.
If pressure p acts on two pistons of different area, the forces are proportional to the areas. A hand pump of 1 cm² pushing on 100 bar (10 MPa) produces 1 kN; the same pressure on a 100 cm² press plunger produces 100 kN ≈ 10 tonnes of force. That is the hydraulic lever — but energy is conserved: the small piston moves 100× further than the large one.
How pumps create flow (not pressure!), displacement volume, and why pressure only appears against resistance.
A hydraulic pump converts mechanical rotation into flow. The key spec is the displacement volume Vg (cm³/rev) — the volume pushed per revolution:
The pump does not "make" 200 bar. It pushes oil; pressure rises only because the flow meets resistance (a load, a restriction). With no resistance the outlet stays near zero. This is why every system needs a pressure relief valve: with the cylinder at its end stop, flow has nowhere to go and pressure would rise until something breaks — the relief opens and returns oil to tank.
Push and pull forces, extension speed from flow, the annulus area effect and regenerative (differential) circuits.
A double-acting cylinder converts pressure back into linear force. Two sides, two areas: the full piston area AK and the much smaller annulus AR on the rod side.
Piston Ø 80 mm, rod Ø 45 mm → AK = 50.3 cm², AR = 34.4 cm². At 160 bar: push force ≈ 80.4 kN, pull force ≈ 55 kN. Feeding 40 l/min: extend speed ≈ 0.13 m/s, retract ≈ 0.19 m/s — the cylinder retracts faster on the same flow because the annulus is smaller.
Connect P to the cap end and the rod side simultaneously: oil from the rod side joins the pump flow into the cap end. The result is a much faster extension at reduced force — perfect for rapid approach before a working stroke. The Studio's intermediate examples include this circuit with electric changeover.
The three classic pressure valves — what each one protects, regulates or triggers, and how to tell their ISO symbols apart.
Pressure valves all compare line pressure with an adjustable spring, but they answer three different questions.
Normally closed, piloted from its inlet, opens to tank at the setting. It limits maximum system pressure and protects the pump. It does not regulate during operation — it just caps peaks (and dumps all pump flow at setting at end of stroke).
Normally open, piloted from its outlet: it starts closing when outlet pressure exceeds the setting, holding the downstream branch at reduced pressure. Used when one branch (e.g. a clamp) needs less force than the main system.
Normally closed, piloted from inlet like a relief, but its outlet feeds another actuator instead of tank: step 2 starts only after step 1 reached the pressure that proves it finished (e.g. clamp → drill). It needs an external drain because its outlet is pressurised.
Positions and ways (4/3, 4/2…), center spool types, actuation methods — and how the square-box ISO symbol encodes all of it.
A directional control valve (DCV) routes flow between ports. "4/3" means 4 ports × 3 positions. Ports: P (pressure), T (tank), A and B (actuator). Each square box shows one spool position with the internal connections drawn inside it.
Manual lever, push-button, roller (limit), spring, pneumatic pilot (triangle) and solenoid (rectangle with diagonal). Detents hold a manually shifted spool.
Controlling actuator speed with throttle valves — why placement matters, and when you need pressure compensation.
Speed follows v = Q/A, so controlling flow controls speed. A simple throttle (needle valve) creates a pressure drop that depends on flow — hence its control is load dependent.
Pair the throttle with an antiparallel check valve and you get one-way speed control: throttled extend, free retract (or vice versa) — the classic "Flow control valve with bypass" in the Studio palette.
A compensator spool keeps a constant Δp (≈ 5–10 bar) across the adjustable orifice, so the set flow stays constant regardless of load. That is what guarantees identical speeds under varying force.
Viscosity and ISO VG, why 70 % of failures are dirt, cleanliness codes, and when a system needs a cooler.
Hydraulic oil transfers power, lubricates, seals gaps and carries heat away. Its most important property is viscosity, specified as ISO VG (mm²/s at 40 °C). Typical choice: VG 32–46; too thin → internal leakage and wear, too thick → cavitation and sluggish response at cold start.
Roughly 70–80 % of hydraulic failures trace back to contaminated oil. Cleanliness is coded by ISO 4406 (e.g. 19/17/14 particles >4/6/14 µm). Servo valves need ~2 classes cleaner than gear pumps. Filters are rated by the β-ratio: β10 = 200 means 99.5 % of 10 µm particles are captured.
Every pressure drop becomes heat: Ploss = Δp × Q / 600 kW. Relief-valve throttling at end of stroke is a frequent heater. If the tank cannot dissipate it (~0.6–1.0 kW per 100 l tank by natural convection), add a return-line cooler — in the Studio you can place oil coolers and a heater with temperature sensors.
Why pneumatics, compressor types, the bar/psi units, and why dry air matters (pressure dew point).
Pneumatics uses compressed air — clean, fast, overload-safe and available everywhere in a factory. Typical working pressure is 6 bar. Air is compressible: that makes pneumatics springy (great for clamping and end-stop cushioning) but less precise than hydraulics for exact positioning.
Filter, regulator, lubricator — what each stage does, ISO 8573-1 air classes, and condensate handling.
Between network and machine sits the service unit — in the Studio it is one symbol combining all stages (insert it with "Service unit").
Separates water droplets and particles (typ. 5–40 µm) in a cyclone chamber; condensate collects in the bowl and drains manually or automatically. Air quality classes follow ISO 8573-1 (particles / dew point / oil), e.g. class 7.4.4 for standard shop air.
A diaphragm poppet holds the secondary pressure constant even when the primary side fluctuates, and vents overpressure (relieving type). This is the knob the operator sets — usually 6 bar.
Adds a fine oil mist for old-style cylinders and tools. Modern machinery is designed lubrication-free (special seals and materials); if a lubricator is installed, it must never run dry — once oiled, always oiled.
Single vs double acting, force at 6 bar, cushioning — and why air exhaust throttling is the standard speed control.
One port; air extends the piston and a return spring retracts it when vented. Used for short strokes, clamps, ejectors. Force is reduced by the spring.
Air drives both directions. Force at 6 bar (0.6 MPa):
End-position cushioning throttles the exhaust over the last millimetres so heavy loads do not slam into the end caps.
With compressible air, meter-in throttling causes stick-slip: the piston hesitates, pressure builds, then it jumps. The standard is meter-out (exhaust-air) flow control at each cylinder port — the trapped air cushions and stiffens the motion. A quick-exhaust valve does the opposite: it vents directly at the cylinder for maximum return speed.
When space or stroke length rules the design: rodless cylinders (magnetic or band coupling between piston and outer carriage — stroke equals the unit length, not twice it), compact and short-stroke cylinders, guided drives, rotary actuators (vane or rack-and-pinion) for swivelling, and bellows/actuated grippers. All follow the same force formula on their effective area.
Port numbering (1, 2, 3, 4, 5 — ISO 5599), monostable vs bistable, air-piloted control and center functions.
Pneumatic valve ports follow ISO 5599: 1 = supply, 2 & 4 = working lines, 3 & 5 = exhausts, 12/14 = pilot lines. So a 5/2 valve has 5 ports and 2 positions.
Controls single-acting cylinders or acts as a signal element: NC (normally closed, supply blocked at rest) or NO. Push-button, roller lever (limit switch), or air-pilot actuation.
Mid position with all ports closed (holds cylinder), both chambers exhausted (free movement), or both pressurised (approximate stop in position).
AND (two-pressure valve), OR (shuttle valve), time delays — and how classic sequences like A+ B+ B− A− are built.
Two inlets, one outlet; an internal ball connects the pressurised inlet to the outlet. Either push-button OR the other starts the cylinder (e.g. control from two positions).
Output only when both inlets are pressurised (it passes the lower pressure). Used for two-hand safety/start conditions: start only when button AND end position are present.
A 3/2 valve + flow control + small air reservoir: after the pilot arrives, the chamber fills through the throttle until switching pressure is reached — an adjustable pneumatic timer (0…30 s typical).
Cylinder A extends (clamp), then B extends (work), B retracts, A retracts. Realised with roller valves at each end position reporting "step done" to the next valve, or with a cascade/stepper module that only enables the active step — eliminating trapped opposing signals, the classic trap in pneumatic sequencing.
Free-air consumption of cylinders, Cv/Kv sizing of valves, tube sizing, and the hidden cost of leaks.
Per stroke, the cylinder volume is filled from (pgauge+1) atmospheres down to 1 — consumption in normal litres:
Example: Ø 63 mm × 400 mm stroke DA at 6 bar: ≈ 1.25 l chamber × 7 ≈ 8.7 NL per extend, plus rod-side ~7.9 NL — about 16 NL per full cycle. Multiply by cycles/min for compressor sizing.
Valve flow capacity is given as nominal flow (NL/min) or C V / K V; undersized valves starve the cylinder and slow it down. Main lines are typically sized for ≤ 6–8 m/s mean velocity, drops and pressure loss ≤ 0.3 bar to the farthest consumer.
The grammar of fluid-power schematics: boxes, circles, triangles, solid vs dashed lines, and composite symbols.
ISO 1219-1 defines the graphic symbols for fluid-power circuits — the "alphabet" every schematic in the RGZ studios follows.
Triangles on the symbol edge mark flow direction of energy converters. Actuation glyphs attach to valve boxes: springs (zigzag), solenoids (rectangle with diagonal), manual, roller, pneumatic pilot (outline triangle), detents (notches).
Tank, motor, pump, coupling, relief and breather — how the component groups of a power pack work together and how the Studio models it.
The hydraulic power unit (HPU) is the heart of every system. In the Studio you place it with one click ("Power pack"), but inside that symbol lives a complete component group — worth knowing part by part.
Bladder, piston and diaphragm types, the pre-charge rule, and the four classic jobs of an accumulator station.
A hydro-pneumatic accumulator stores oil against compressed nitrogen. Types: bladder (fast response, most common), piston (large volumes, high pressure), diaphragm (small sizes).
The nitrogen pre-charge defines the working window. Filling must never drop below p₀ (bladder damage) and the system must never compress the gas to its limit. Common rule: p₀ ≈ 0.6–0.8× minimum working pressure (0.9× for damping tasks). Usable volume follows the gas law:
Why pumps scream and die young: vapor bubbles at the suction, implosion damage, and how to prevent them by design.
If absolute pressure in the suction line falls near the oil's vapor pressure, vapor bubbles form. Carried to the pressure side, they implode in microseconds, creating micro-jets up to ~1000 bar that erode nearby metal (pitting). Symptoms: the pump gets loud (gravel rattling), flow falls, and the suction elements wear fast.
Aeration = air entering (leaky suction fitting, foaming return above oil level). It makes oil spongy and can cause micro-dieseling: compressed air bubbles ignite the oil-air mixture locally like a diesel engine — scorching seals and degrading the oil.
Where the pressure goes: line friction, velocity classes for suction/pressure/return, and picking hose and tube diameters.
Every restriction converts pressure to heat. Line losses follow Darcy–Weisbach:
Note the brutal exponent: loss rises with velocity² — and halving the diameter multiplies the loss ~32×. Fittings, elbows, valves and filters add their own resistances (Σζ).
From the target velocity: d = √(4Q / πv). Example: 60 l/min in a pressure line at 5 m/s → d ≈ 16 mm inner bore — the next standard size up wins. Hoses also age faster at temperature peaks and must meet the pressure rating with a safety factor ≥ 4:1 against burst.
Minimum bend radius, torsion, slack, clamps and cleanliness — the installation rules that decide a hose assembly's lifetime.
Most premature hose failures are installation failures. The rules come straight from EN 853/856 and manufacturer practice:
Flush new pipes, cap open ends immediately, and clean hose assemblies after cutting/crimping — rubber and metal residue from assembly is a classic early-failure cause for valves and pumps downstream.
NG6/NG10 porting patterns, subplates and sandwich valves — how industrial valve stations are physically assembled.
Industrial DCVs bolt onto standardized interfaces defined by ISO 4401 (DIN 24340): size 03 = NG6 (¼"), 05 = NG10 (⅜"), 07 = NG16, 08 = NG25, 10 = NG32. The pattern fixes exactly where P, T, A, B (and X, Y, L) ports sit, with locating pin and bolt holes — so any brand's valve fits any pattern-correct plate.
Bore from force, check the load with reserve, then verify the piston rod against buckling on long strokes.
Round up to the next ISO bore (25, 32, 40, 50, 63, 80, 100, 125, 160 mm…). Then check the retraction force too when it drives the load. Speeds follow from flow: v = Q/A — consider whether the annulus-side retraction speed suits the process.
A long, slender piston rod in compression can buckle — the governing case for presses and lift cylinders with free stroke. Effective buckling length depends on the mounting (guided/fixed ends), and manufacturers publish permissible load vs. stroke tables per rod diameter. Reduce the risk with: a thicker rod, guides/trunnion or clevis mounting nearer the load line, or reduced pressure.
A field-proven sequence: pressure first, then flow — plus how to find internal leakage and test cylinders for drift.
Permanent test points (M16×2) at pump, relief, valve P/A/B make a 2-minute pressure reading possible without spilled oil. The Studio's test-point and gauge components mirror exactly that practice.
How proportional spools differ from on/off valves, and the amplifier parameters gain, ramp, deadband and dither.
On/off valves know two states; proportional valves position the spool continuously against a spring in proportion to the command current (typ. 4–20 mA or 0–10 V), metering flow (and thus direction and speed) in one valve. Servo valves go further: nozzle-flapper or jet-pipe pilot stages with micron-grade spools — maximum dynamics, but they demand very clean oil.
What a hydraulic fluid must do, how additives deliver it, and when to choose HFA/HFC/HFD fire-resistant fluids.
Never mix fluid types; changing types requires full flushing and filter/seal review.
Excavator or injection molding machine? Why mobile systems look different: load-sensing, DCV blocks, and 24 V control.
The physics is identical — but priorities differ sharply between a factory power unit and a mobile machine.
Piston vs screw vs scroll, duty-cycle thinking, and how receivers buffer consumption and let the compressor rest.
Compressed air leaves hot and wet. An aftercooler drops most water immediately; a refrigerated dryer then sets the +3 °C pressure dew point typical for general industry (adsorption dryers for −20…−70 °C).
The tank buffers demand spikes so the compressor cycles less: a machine drawing 60 NL once per minute can run a much smaller compressor behind a correctly sized receiver. Keep the pressure band (load/unload settings) as narrow as hysteresis allows — every extra bar of band is wasted energy (~7 % more energy per bar).
Ejectors, vacuum level vs flow, suction-cup holding force, and pneumatic grippers for handling.
A venturi ejector accelerates supply air through a nozzle; the jet entrains air from the vacuum port and exhausts both through a silencer — simple, moving-part-free vacuum to about 85–95 % of atmospheric pressure. Multi-stage ejectors reach the same level with less air.
Holding force comes from ambient pressure pushing the cup against the part — at −0.6 bar relative, a Ø 40 mm cup holds theoretically ≈ 75 N. For shear loads on vertical surfaces use double safety and bellows/oval cups on textures. Porous workpieces (cardboard, chipboard) need high-flow ejectors — leak rate, not vacuum level, governs.
Solenoids, reed switches, latching circuits and seal-in logic — the bridge from pure pneumatics to real machine control.
An electrical push-button energizes the valve's solenoid coil (typically 24 V DC): the valve shifts, the cylinder moves. At the end position a reed switch on the barrel (or an inductive proximity sensor for parts) reports the position back — closing the loop for the next step.
Detail you will use constantly: wire the relay's own NO contact in parallel with the start button. A pulse latches the relay ON ("seal-in"); the stop button (NC, in series) breaks it. That is exactly how the Studio's intermediate electropneumatic examples hold the cycle between start and stop.
The three enemies — dirty air, moisture, leaks — and a quick diagnostic sequence for sluggish cylinders.
The twelve formulas that solve 90 % of fluid-power engineering tasks — from force and flow to power and air consumption.
DC control circuits, Ohm's law, what a solenoid coil actually does with 24 V, and how to measure correctly with a multimeter.
Every electro-hydraulic and electro-pneumatic machine is two systems in one: a power section (oil or air) and a signal section (electricity). The signal side almost always runs on 24 V DC — safe to touch, short-circuit tolerant and standard for industrial sensors and valves.
Voltage U (volts) pushes, current I (amperes) flows, resistance R (ohms) opposes. A typical valve coil: 24 V, 30 Ω → it draws 0.8 A and converts ≈ 19 W into the magnetic force that strokes the spool (plus heat — a coil is rated for 100 % duty and gets warm in service, that is normal).
Current through the winding builds a magnetic field that pulls an iron armature a few millimetres. That armature strokes the valve pilot or spool. No current → the return spring drives the valve home. This crisp ON/OFF behaviour is what makes solenoid valves the standard "muscle switch" of fluid power.
NO vs NC contacts, reed switches on cylinders, inductive / capacitive / optical proximity sensors and electro-pneumatic pressure switches.
Control starts with information: input devices delivered by people (pushbuttons, selector switches) and sensors delivered by the machine (positions, pressure, presence).
A pushbutton has one of two resting states: NO (normally open — closes when pressed) and NC (normally closed — opens when pressed, used for STOP and safety chains). Contacts are numbered per standard: function digit first (1 = NC, 2 = NO), then the pole index — so 13–14 is the classic NO pair.
Output styles: PNP (switches +24 V to the input — European standard) and NPN (switches 0 V). Three wires: brown +24 V, blue 0 V, black signal.
How a small signal switches a big load, contact numbering, pick-up vs drop-out delay, and the three-wire latching pattern every classic control is built from.
A relay is an electrically operated switch: a small coil (24 V, <1 W) pulls an armature that moves isolated contacts. It separates control from load, multiplies one signal into many contacts, and stores information. A contactor is the same idea scaled for power loads (motors, pumps).
Put the relay's own NO contact in parallel with the START button. Press once → the coil energises and seals itself in through its contact. A NC STOP in series releases it. Exactly the same pattern appears everywhere: as ladder logic in a PLC, or as a double-solenoid valve's two coils (SET / RESET).
The 5/2 solenoid valve as the interface between electricity and air: direct vs control-side actuation, coil ratings, overrides and a complete press example.
Gate between the worlds: the solenoid directional valve. Its coil moves a small pilot; the pilot air strokes the main spool (on larger valves) or the coil drives the spool directly (small 3/2, 5/2 valves).
Press S1 → relay K1 energises → its contact drives coil Y1 → the 5/2 switches → the cylinder extends against the part. Release S1 → Y1 drops out → spring return on valve and cylinder. Add a NC guard-door contact in the rung and you have the skeleton of a safe press control.
Building a controlled A+ B+ B− A− cycle with reed contacts and relays: step guards, signal storage and on-delay timing — the ladder logic of classic machines.
A sequence control executes steps in a fixed order — each step is only allowed when the previous one reports completion. That report comes from reed switches (positions) or pressure switches, and the "brain" is a small set of relays.
Exercise classic A+ B+ B− A− (clamp, feed, retract, unclamp): relay K1 runs when START and home signals a0 AND b1 are true; K1's contacts drive coil Y1 (A extends). Sensor a1 fires K2 → B extends; b1 fires K3 → B retracts; b0 fires K4 → A retracts. Every rung is a tiny AND: condition from the sensors, permission from the previous relay.
Three classic pure-pneumatic specials: dwell timers from restrictor + volume + 3/2, pressure-triggered next steps, and the little valve that makes cylinders fast.
Not everything needs a PLC. Three small components deliver timing, pressure logic and speed directly in the air lines.
A restrictor + small air volume + 3/2 valve: the signal slowly fills the volume; when threshold pilot pressure is reached, the 3/2 switches. Adjustable 0–30 s. Connect it normal-closed for an on-delay, reversed for off-delay. Typical job: clamp dwell, filtered start signals (ignore bounces), end-position acknowledge.
A 3/2 that switches when line pressure exceeds its setting — the pneumatic pressure switch without electricity. Circuit pattern: "drill down until the counter-pressure from the block signals part-contact, then advance slowly".
A tiny disc valve mounted directly on the cylinder port: flow goes in normally through the small throttle, but exhaust dumps to atmosphere at the port instead of traveling back through the metre of tube to the valve. The classic way to retract a cylinder at full speed.
Why A+ B+ B− A− jams a naive control, how to split a sequence into groups that never collide, and how reversing valves clear blocked pilot signals.
Single-actuator controls are forgiving — every sensor signal has exactly one meaning. With two or more cylinders the same signal can be needed twice, and the classic problem appears: signal overlap.
Sequence A+ B+ B− A− with limit valves wired straight to the 5/2 pilots: when B leaves b0, its signal arrives at the "extend B" pilot — but B still stands still, so its own limit signal b0 is still pressed and holds the "retract B" pilot. The valve now feels both pilots at once: nothing moves, the machine hangs.
Solution of the Festo course: split the motion chain into groups that contain no cylinder twice. A+ | B+ B− | A− → three groups (or two in smarter splits). A cascade of supply-switching 4/2 valves feeds pilot air to one group at a time — signals of inactive groups are simply dead, so nothing can block.
The lighter alternative: a roller-lever reversing valve on the offending signal line, flipped by the cylinder itself. It exists exactly to chop the long signal that would otherwise overlap. Co-ordinated motion (feed + clamp + eject) is designed the same way: draw the displacement-step diagram, mark every signal, resolve every overlap on paper.
How air travels from the compressor room to the machine without losing pressure or carrying condensate — pipe sizing, slope, drop engineering.
The cleanest compressor cannot help a badly piped plant: undersized lines eat pressure (and money), and condensate follows the air straight into the valves.
Size by total consumption plus simultaneity and growth. As orientation: 6–10 m/s in mains; every bar of Δp at the tool is ≈ 7 % wasted compressor energy. Materials: galvanised steel, stainless, aluminium profile systems or PE grid lines — modern aluminium nets assemble without threading and never rust.
Power section meets signal section: direct and relay-switched solenoid DCVs on single- and double-acting cylinders, with the Boolean view of the classic exercises.
Electro-hydraulics pairs oil's force with electricity's logic. System structure (Festo model): power section (pump, valves, cylinder), signal section (sensors, relays/PLC), and the interface — the solenoid on the valve.
Contacts in series = AND (guard AND start), in parallel = OR (jog from panel OR remote). NC gives NOT. The XOR circuit (two stations, exactly one active) from the assembly-line exercise is two AND-NOT branches — worth drawing once by hand.
Clamp → press → release without a PLC: signal storage with double-solenoid valves, electrical set/reset, speed stages and pressure- or path-dependent sequences.
Real machines chain motions: clamp the part, feed the tool, retract, unclamp. Two families of memory solve it.
Clamp → press is typically pressure-dependent: pressure switch B1 closes only when clamp pressure is truly reached (part present, no slip) — then Y2 feeds the press. Retraction is path-dependent: an inductive sensor at the end position starts the return. Add a throttle check for a controlled working feed and an on-delay for dwell at depth; add a counter and an AUTO latch for continuous cycling until N parts are done.
Gear, vane and piston motors, torque from Δp and displacement, speed from flow — and where rotary actuators beat cylinders.
A hydraulic motor is a pump run backwards: pressurised oil in → shaft torque out. Same families, same physics — the displacement Vg is again the master spec:
Pressure rises only with the load torque — an unloaded motor spins at nearly zero pressure. Leakage oil (case drain — the third small line!) must run directly to tank unpressurised, or the shaft seal dies. For reversing drives use a 4/3 with closed centre plus cross-line reliefs protecting both legs against overrunning-load pressure spikes.
Bourdon-tube gauges, electronic pressure and flow sensors, temperature monitoring — the instruments that turn troubleshooting from guessing into reading.
Everything you have calculated in the previous lessons (p, Q, v, force) is only verifiable if you can measure it — that is why professional power units have test points at every critical leg.
Measure at defined points in a defined operating state, write the values down, compare against commissioning values. A relief drifting 20 bar down or a pump losing 30 % flow is visible months before failure — if you measured.
Hysteresis, inversion range, response threshold, pressure curves and step response — reading a proportional valve the way its designer does.
A proportional valve is fully described by its characteristic curves. Learn to read them and selection + commissioning become engineering instead of mystery.
Step-response time to 90 % and the −3 dB (or −90°) frequency tell how fast the valve follows: typical direct-acting 20–80 ms, piloted faster stages exist for demanding axes. Bode plots in datasheets pair amplitude and phase — for loop tuning, phase margin is the interesting one.
Max flow grows with √Δp over the metering edge — doubling rated Δp gives only ~40 % more authority. Overspeed conditions and pressure shocks need their own protections (counterbalance, reliefs) — proportional spools are metering edges, not safety valves.
From the manually switched feed drive to closed-loop positioning: the signal chain, speed profiles, and when closed loop is worth its sensor.
The Festo training path builds the same feed drive three times: manual hand lever → switching solenoid valves with limit switches → proportional. The proportional version wins on smoothness, programmability and reproducible speed profiles.
Setpoint (potentiometer, PLC analog output) → amplifier (ramps, gains, dither) → proportional valve → cylinder/motor speed. Open loop: the machine trusts the chain. Closed loop: a position or speed sensor reports the truth back to a controller that compares, computes error Δe and corrects the command continuously.
Speed profiles with ramps avoid jerking loads and compressive shocks; a slowdown point before every end position (taught via PLC cams) gives soft, fast cycles. Watch energy: proportional metering throttles → heat; over-centre loads need counterbalance valves; minimal leakage circuits protect warm-up accuracy. Modern axes move the whole loop INTO the valve (on-board electronics, digital fields).
Displacement-step diagrams, function charts, circuit and terminal diagrams — the paper trail that makes machines designable, testable and repairable.
A machine that exists only in someone's head is unrepairable. The Festo documentation set has four views — each answers a different question.
One line per actuator, steps on the x-axis, path/stroke on the y-axis: WHO moves WHEN. Drawn before any component is chosen — it is the contract of the control design, and reveals signal overlap at a glance.
The ladder-friendly tabulation: steps down, actuators and sensors across, X marks active elements per step; the direct bridge into relay/PLC programming (SFC-style).
The ISO symbol schematic (fluid) and the electrical ladder diagram (control). Elements numbered by circuit groups: 0 = supply, 1 = first drive, 2 = second…; every element's mark in the drawing = its mark in the machine.
For the cabinet builder: every wire numbered, every terminal labelled; the solenoid in rung 3, wire 14, terminal X2:7 — one identity from drawing to screwdriver.
The non-negotiable rules of fluid-power and electrical maintenance: depressurise, lock out, verify — and respect the pinhole jet.
Switched off ≠ de-energised. Accumulators hold hundreds of bar for hours; compressed-air receivers too; suspended cylinder loads can descend on their own. The ritual before any intervention: Lock out the start → release pressure at the designed bleed point → verify zero on the gauge → only then work (LOTO — lock-out, tag-out).
A pinhole-leak jet at 200 bar penetrates skin without pain at first and injects oil under it — a surgical emergency, not a band-aid case. Never trace a leak with the hand; use cardboard, gauge checks, or the shut-down method.
Control circuits at 24 V DC are body-safe; their knowledge risk remains (a 24 V mistake still drops a press). Machine mains, transformers and large capacitor banks belong to qualified hands. Effects of current on the body are current-path and time dependent — the professional rule-set (isolation, covers, RCDs, color-of-cables discipline) exists in every plant for a reason.
Step 1 of every hydraulic design: list the load, choose the working pressure, and calculate the cylinder bore with reserve. Worked numbers included.
This is the first lesson of the Hydraulic Designer Track. Each track lesson follows the same pattern: design decision → formula → worked numbers → build it in the Studio. Follow them in order and you will have dimensioned a complete, realistic machine by the end.
We will design a hydraulic press/lift: it must raise a 3 200 kg load (~31.4 kN) by 400 mm in about 4 s, then hold and lower it controlled. Keep this brief — the next lessons reuse it.
Pressure is a design choice, not a given: higher pressure → smaller, cheaper components, but more noise, leakage stress and filtration demands. Industrial standard bands:
63 – 100 bar140 – 210 bar250 – 350 barFor our lift we pick p = 160 bar — a good middle band with standard NG6 components.
Never round down. The standard series is 25, 32, 40, 50, 63, 80, 100, 125, 160 mm. With Ø 63 the static pressure at the load is only pstatic = 31 400 / 31.2 cm² ≈ 101 bar — comfortable reserve under our 160 bar network.
Standard rods are ~0.5–0.7 × bore: choose rod Ø 36 mm → annulus area AR = 31.2 − 10.2 = 21.0 cm², area ratio φ ≈ 1.49. With a 400 mm push stroke the rod is not slender (check the manufacturer's buckling table — 36 mm rod at 400 mm free stroke is rated far above 31 kN). Details: lesson “Selecting a hydraulic cylinder”.
Turn the motion requirement into a pump flow, then size suction, pressure and return lines so velocity and pressure drop stay in the green bands.
Design decision #2: how much oil must move, and through which pipes? Continuing our project: Ø 63/36 × 400 mm, lift in ≈ 4 s.
Retract comes free: the same 20 l/min on the annulus (21 cm²) gives v ≈ 0.16 m/s — the classic faster return. If the cycle time is unacceptable you either raise Q or accept it; that decision belongs here, not later.
0.5 – 1.5 m/s (cavitation-safe)2 – 4 m/s3 – 6 m/s up to ~150 bar, 5 – 7 m/s aboveEvery metre of line and every fitting eats pressure (Δp ∝ v²). On a 20 l/min system a badly chosen DN 6 pressure line can burn 15–25 bar as pure heat. Rule: total pressure-line losses ≤ 5–10 % of working pressure at full flow. The Studio colours undersized lines red in the pipe-sizing report — treat that report as your design checker, not decoration.
The 4/3 valve is a design statement: what may the machine do when stopped, starting, and mid-stroke? Select left, center AND right blocks like a professional.
Most beginners pick a “4/3 valve” and stop thinking. A designer picks three blocks — left, center, right — because each switching position is a machine state with consequences.
Our press must: extend (work) → stop+hold → retract. Three states = 3 positions. Double-acting cylinder = 4 ports. → 4/3 valve. (2-position valves suit bang-bang motions without intermediate stop.)
In the Studio's valve inspector you can now choose the left and right envelopes independently:
Solenoid (electrohydraulic sequences, our project), manual lever (jog/maintenance), pilot pressure (large values or logic). Rule: choose the spring arrangement first — spring-centered with 2 solenoids for 3 positions, spring-offset with 1 solenoid for 2 positions.
Decide where the throttle lives, then commission it with a repeatable 4-step procedure instead of turning screws until it “looks right”.
Our 3.2 t load must not free-fall during lowering. That single sentence decides the speed-control topology — the throttle goes on the rod side (meter-out for the retract/lower motion).
Throttling the rod side of our φ≈1.5 cylinder means the trapped rod-side pressure can reach prod ≈ pcap·φ + F/AR. Check hose and seal ratings on the return side — a “return line” in a meter-out lowering circuit is a pressure line!
Set the relief with the 1.15× rule, decide when a reducing valve is required, and add an accumulator for hold-at-pressure and power-failure cases.
Design decision #5: protecting the system at its edges — overpressure, branch pressure, holding force after the pump stops.
Decide the ceiling first, then live 15 % below it. Every component downstream (hoses, seals, gauges) is selected against the relief setting, not against “typical” pressure.
If our machine also clamps a delicate part (needs only 60 bar of clamp force), a pressure reducing valve in the clamp branch beats a separate power pack by a mile. It is normally open and throttles when downstream pressure exceeds the setpoint — different animal from the relief (normally closed). The Studio examples “Reduced-pressure clamp branch” and “PRESSURE CLAMP” demonstrate both.
Press-hold, emergency descent, pump-failure completion of a safety motion: all need stored energy. Quick sizing for our hold case:
Example: leak compensation for 60 s at ~5 ml/min → Vusable ≥ 0.3 l; with p₀=65 bar, p₁=100, p₂=160 bar → V₀ ≈ 0.75 l → choose a 1 l bladder accumulator. Always with: safety block (relief + drain), nitrogen charging rig, and the isolation valve before any service.
The last 20 % of design work that prevents 80 % of commissioning pain: power balance, heat balance, instrumentation points and a clean BOM.
Your schematic works in simulation. Is it a good machine? Designers answer three numbers and one document before releasing anything.
All throttled energy becomes heat. Worst case for our press: dwell phases with pump over relief = full 4.7 kW into the oil. Natural tank dissipation ≈ 0.5–1 kW per 100 l. With an 80 l tank we need a return-line cooler ≈ 4 kW, or better: tandem center + load-sensing so the pump idles during dwell. Simulation truth: run your circuit 5 minutes of dwell and watch the oil-temperature metric — over 60 °C sustained means redesign, not a bigger cooler.
Minimum professional set: gauge isolator point after pump (M), at valve P, at both cylinder ports, temperature + level switch in the tank, pressure switch where control cares about pressure (our clamp). Every point costs €5 and saves a service hour later. Masses of red gauges in the examples are a habit, not a style.
From your final Studio schematic, write every element with: quantity, type, key spec (Ø 63/36×400, 20 l/min @ 1450 rpm, NG6 4/3 tandem 24 V DC…), and settings (relief 160 bar, throttle position from H4, accumulator pre-charge from H5). Add hose diameters from H2. This document IS the machine for purchasing and service.
The pneumatic counterpart of H1/H2: bore from force with the 1.3 rule, air consumption for the FRL and compressor check, and exhaust throttling done right.
Pneumatic Designer Track. Project: a pick-and-place pusher must shove a 25 kg box 200 mm in 0.8 s, 30 cycles/min. Friction and return spring eat margin — design margin in pneumatics is not optional.
F = 25 kg · ~0.3 friction µ · g ≈ 74 N… tiny. The speed is the real driver — but start with force anyway:
Choose one or two sizes above the arithmetic when in doubt — air is cheap compared to a stalled cycle.
v ≈ 250 mm/s required. That is fast for pneumatics: size the valve ≥ the cylinder's demand (nominal flow at 6 bar in the datasheet; for Ø 25 at 6 bar & 250 mm/s ≈ 150–250 NL/min — a standard 5/2 with G⅛" ports suffices). Tubing: 6×4 mm for cylinders ≤ Ø 32, 8×6 mm up to Ø 63 — keep runs under ~2 m or speed decays measurably.
Exhaust-air flow controls on both ports. Meter-in on compressible air gives stick-slip: pressure builds behind the piston until static friction breaks, then it jumps. The Studio demonstrates this: set a supply-side throttle and watch the piston stutter; move it to the exhaust side and the motion smooths.
The method that scales from two cylinders to twenty: displacement-step diagram → sensor placement → PLC rungs — built and simulated end-to-end in the Studio.
Design decision of every automation machine: the sequence. Amateurs wire by intuition and debug for days; professionals draw the step diagram first and the circuit almost falls out of it.
Machine: feeder pushes a part under a press, press stamps, both retract. Write the motion chain: A+ B+ B− A− (A = feeder, B = press). Draw two lines (A and B) over 4 steps — who moves when. This 2-minute drawing is the contract everything else must satisfy.
Each transition needs proof the previous motion finished: reed switches a0 a1 b0 b1 at the four end positions. Rule: one sensor per state change you depend on — never infer B extended from “enough time passed” if quality or safety depends on it (use time only for dwell, and a pressure switch when force matters).
Trace every pilot signal's lifetime in the diagram: does any sensor keep its signal into a step where it would command the opposite? In A+ B+ B− A− with bistable valves, b0 lingers — classic overlap. Solutions: cascade air supply by groups, electrical edge evaluation, or — the modern choice — a PLC evaluating conditions each scan.
Give each step a memory bit (Studio PLC syntax — one condition, one action per rung):
S1 = START AND a0 AND b0 — step 1 enable: home positions proven.S1 -> V1 left — A extends (valve V1 to extend).S2 = a1 → S2 -> V2 left — when feeder arrives, press extends.S3 = b1 → S3 -> V2 right — press bottom sensor retracts the press.S4 = b0 → S4 -> V1 right — press home → feeder back; cycle re-arms at a0.No trapped signals, no air-group plumbing — sensors are PLC inputs, solenoids are outputs, and deleting PLC1 really stops the machine (test it!). That is exactly how a real machine dies when its controller is pulled.
PLC1 (temporarily!): the sequence must now stop after the first motion. Undo the delete — you have proven the control chain end-to-end.b1 -> T1 start … T1 -> V2 right).
The heart of every system: tank, motor, pump and relief valve in one unit.
The heart of every system: tank, motor, pump and relief valve in one unit.
The power pack converts electrical energy into hydraulic flow. In the Studio it is one symbol exposing P (pressure) and T (tank) ports, but physically it contains the tank, the electric motor, the pump, the pressure relief valve and usually the breather, level gauge and suction strainer.
The motor drives the pump at constant speed (4-pole ≈ 1450 rpm at 50 Hz). The pump draws oil from the tank through the suction line and delivers a fixed flow to P. When the actuators stop consuming flow, pressure climbs until the relief valve opens and dumps oil back to T — limiting the maximum system pressure.
Specify by pump flow (l/min), relief setting (bar, the system ceiling!) and motor power (kW). Tank volume is typically 2–4× the pump flow per minute for stationary units. Relief = 1.1–1.15× the highest working pressure.
The workhorse: simple, cheap, robust fixed-displacement delivery.
The workhorse: simple, cheap, robust fixed-displacement delivery.
Two meshing gears in a tight housing transport oil in the tooth gaps from suction to pressure side. It is the most common industrial pump because of its tolerance to dirt, low price and simple construction.
As the gears unmesh on the suction side, the expanding volume draws oil in; the oil travels around the housing wall (never between the teeth!) and is squeezed out where the gears mesh again. Flow per revolution is fixed by the tooth volume (displacement Vg).
Sizes 0.25–250 cm³/rev, continuous pressures to ~250 bar (300 peak). Volumetric efficiency 85–95 %, drops with wear and with low-viscosity hot oil. Noisy above ~2000 rpm.
Swashplate control: flow follows demand instead of the relief valve.
Swashplate control: flow follows demand instead of the relief valve.
An axial-piston pump whose displacement is adjustable while running — usually by tilting the swashplate. A compensator piston moves the plate against a spring, so the pump itself regulates its output.
When system pressure reaches the compensator setting, a small control spool admits pressure to the destroking piston: the swashplate flattens and the pump delivers only leakage/compensation flow. Pressure stays constant while flow follows the consumers — no relief throttling, far less heat.
Pressures to 350 bar continuous. Controls: pressure compensator, load-sensing, power limiting, electric proportional. More expensive and more sensitive to contamination (ISO 4406 class ~19/17/14 or better) than gear pumps.
Muscle-powered pressure: for service, rescue and remote tools.
Muscle-powered pressure: for service, rescue and remote tools.
A lever-driven piston pump, often two-stage (large piston for fast prefill, small piston for high pressure). No electricity needed — the standard backup when the power pack fails or is unavailable.
The lever strokes a small piston; check valves rectify the motion so each stroke pushes a small volume toward the consumer. Two-stage designs switch automatically at a set pressure: high flow below ~20 bar, then a small plunger continues up to 700+ bar.
Typical: 1–20 cm³ per stroke, 160–700 bar. Oil reservoir integrated. Handle effort ≤ 350 N at full pressure.
A pump in reverse: pressurized oil in, shaft torque out.
A pump in reverse: pressurized oil in, shaft torque out.
The gear motor converts flow back into rotation. Externally identical to a gear pump, internally modified (balanced bearings, case drain) because it starts under load and may rotate both directions.
Pressure acts on the tooth surfaces: the imbalance between the pressure and suction sides creates torque. Flow rate sets the speed; the pressure difference sets the torque. The third small port (case drain) returns internal leakage to tank without pressure.
To 250 bar, 500–4000 rpm, modest starting torque (85–90 % of theoretical). Always connect the case drain directly to tank — back pressure kills the shaft seal within hours.
The heavyweights: high pressure, high starting torque, precise low speed.
The heavyweights: high pressure, high starting torque, precise low speed.
Axial-piston motors (bent-axis or swashplate) and radial-piston motors deliver the highest torque densities. Radial designs (multi-lobe cam) rule LSHT duty: huge torque at a few rpm.
Pressure on the pistons creates force components against the swashplate or cam ring; the geometry converts them to shaft torque. Variable versions change displacement on demand — high displacement for starting, low for speed.
To 450 bar; starting torque up to 95 %; radial LSHT units to several 10 000 Nm. Require clean oil (servo-grade filtration for high-speed axial types).
The universal linear actuator: force in both directions, asymmetric speeds.
The universal linear actuator: force in both directions, asymmetric speeds.
A piston on a rod inside a tube, with a port on each chamber. Cap (piston) side area AK is larger than the annulus AR on the rod side — the source of unequal forces and speeds.
Pressure on the cap side extends the rod; pressure on the annulus retracts it. Same flow → faster retraction (smaller area). Same pressure → larger extend force (larger area). The ratio φ = AK/AR (1.25–2) must enter every sizing calculation.
Bores 25–400 mm, rods 0.4–0.7× bore, strokes to 3+ m. Add end cushioning for kinetic energy above ~0.1 m/s with heavy loads; consider buckling for long push strokes (Studio inspector computes both).
One port, one direction: returned by spring, load or gravity.
One port, one direction: returned by spring, load or gravity.
Only the cap chamber is pressurized; the rod side is vented. Retraction comes from a return spring, the load itself (press rams, car lifts) or external force. Plunger rams have no piston at all — the rod IS the pressure surface.
Oil in → extend (minus spring force in spring-return types). The 3/2 valve releases the chamber to tank → the spring or gravity pushes oil back. Simple, cheap, and needs no valve outlet management on the retract side.
Force = p·A minus spring/gravity. Spring-return strokes limited to ~100 mm (spring packaging); plunger rams to 1 m+. Always provide a breather on the vented side — a plugged breather stalls the return.
Long stroke from a short body: staged sleeves extend one after another.
Long stroke from a short body: staged sleeves extend one after another.
Several nested stages (sleeves) extend in sequence — the retracted package is roughly 1/3 of the full stroke. Available single-acting (dump trucks) and double-acting.
Pressure first moves the largest sleeve (its cap area is biggest — highest force at lowest pressure); when it bottoms, the next sleeve moves. Forces step DOWN and speeds step UP with each stage, because the active area shrinks.
Strokes to 10 m from ~3 m package. The largest stage sees the full stroke force; side loads are critical — the last (thin) stage tolerates almost none. Check tipping clearance and stage-seal accessibility.
A rod on both ends: equal areas, equal forces, equal speeds both ways.
A rod on both ends: equal areas, equal forces, equal speeds both ways.
The piston rod exits both end caps. Both chambers have identical annulus areas, so extend and retract are perfectly symmetric.
Because A is equal on both sides, F and v are identical in both directions — and a double-rod cylinder synchronizes naturally when two chambers are connected in series (volume in = volume out).
Force is rod-side class only (no full-bore cap area!). Choose for symmetry, not for power. Excellent rigidity against bending since both ends are guided.
Swiveling instead of stroking: torque through a limited angle.
Swiveling instead of stroking: torque through a limited angle.
Limited-rotation actuators: a vane in a circular chamber (≤ 270°, one or two vanes) or a piston driving a rack that turns a pinion (to 360°+). They replace cylinder + linkage when the motion is directly rotary.
Pressure on the vane face creates torque directly; in rack-and-pinion units, a short-stroke cylinder drives the rack, and the pinion converts stroke to angle. Adjustable end stops and cushioning are integral.
Torque from a few Nm (bench grippers) to > 100 kNm (valve actuation in pipelines). Noisy T-fittings shorten vane seal life — use straight-through flow.
The standard cylinder valve: four ports, three positions, three design blocks.
The standard cylinder valve: four ports, three positions, three design blocks.
Four ports (P, T, A, B) and three spool positions. Each position connects the ports in a defined pattern — in the Studio you can now configure the left, center AND right block independently, from parallel to crossover, tandem, float, closed and regenerative.
A sliding spool with lands opens and closes port connections. Solenoids stroke it (electrohydraulics), springs return/center it. Flow forces grow with √Δp across the metering edge — the practical limit is thermal, so watch rated flow at the actual Δp.
NG6 (to ~80 l/min), NG10 (~120), NG16+ per ISO 4401. Center block = machine state at rest: tandem unloads the pump while holding the load; closed locks everything; float frees the actuator.
Two-position workhorses: on/off control for DA and SA actuators.
Two-position workhorses: on/off control for DA and SA actuators.
4/2 valves drive double-acting cylinders with no mid position; 3/2 and 2/2 valves serve single-acting cylinders, pilot lines and function switching. Cartridge (seat) versions shift leak-free.
One solenoid strokes the spool; a spring returns it (monostable) or friction/detent holds it (bistable). Spool valves tolerate tiny leakage; seat valves seal absolutely — choose seats when a load must never creep.
2/2 seat valves: zero leakage, full flow both directions, the standard safety isolator. Spool 4/2 NG6 to 80 l/min. Always check "overlap at switching" for tight clamp circuits.
Direction AND speed in one valve — spool position follows the command current.
Direction AND speed in one valve — spool position follows the command current.
A solenoid pair with position feedback strokes the spool proportionally to a 4–20 mA / 0–10 V command. Metering edges are shaped so flow (hence actuator speed) scales with spool travel.
The amplifier converts the command to a coil current; the spool balances magnetic force against a spring. Dither (100–200 Hz ripple) keeps the spool micro-sliding and kills friction hysteresis; ramps in the amplifier soften reversals.
Hysteresis ≤ 1–3 %, deadband 10–30 % (overlapped) or near-zero (feedback spools). Demands filtered oil (ISO 4406 18/16/13 or better). Check natural-frequency margin: valve ~4× faster than the axis.
The safety guardian: caps system pressure, dumps flow to tank at setting.
The safety guardian: caps system pressure, dumps flow to tank at setting.
Normally closed, piloted from its own inlet. When inlet force exceeds the adjustable spring, the poppet/spool opens to tank. Direct-operated for small flows; pilot-operated for the main system protection.
Pilot stage (small poppet + spring) opens first; the tiny pilot flow creates a pressure drop across a spool orifice, lifting the main stage. That two-stage trick keeps the opening characteristic flat: full pump flow over setting with little overshoot.
Set 1.1–1.15× max working pressure. Cracking vs full-flow pressure differ ~10 % (direct) / ~3 % (piloted). Every bar of relief-flow becomes heat: P = Δp·Q/600 kW. Vented versions enable remote/unload control.
Normally OPEN and piloted from its outlet: holds a branch at reduced pressure.
Normally OPEN and piloted from its outlet: holds a branch at reduced pressure.
The mirror image of the relief: open at rest, it starts closing when its OUTLET reaches the setting, throttling the supply to hold downstream pressure constant — even with fluctuating inlet.
Outlet pressure pilot the spool against the spring. Exceed the setting → spool travels toward closed; below → fully open. Relieving versions also dump downstream overpressure to tank (important with thermal expansion in clamped volumes!).
Outlet ranges typically 5–100 bar. Needs an external drain when relieving. Hysteresis ~1–3 bar; pilot flow 0.5–1 l/min continuous.
Pressure logic in oil: step 2 starts only when step 1 has truly finished.
Pressure logic in oil: step 2 starts only when step 1 has truly finished.
Like a relief valve, but its outlet feeds another actuator instead of tank. It opens when its inlet (the completed function’s pressure rise) proves the first step is done — a hydraulic IF/THEN.
During cylinder 1 extend, pressure is below the setting: the valve is closed, step 2 gets nothing. Cylinder 1 reaches its end stop/clamp force → pressure rises to the setting → valve opens → cylinder 2 moves. External drain is required because the outlet is pressurized.
Set above the first step’s max running pressure (plus ~10–15 % margin). Combination with check valve = free reverse flow. Avoid chatter at the transition with small volumes (damped spools exist).
The vertical-load valve: no runaway, no drift, controlled lowering.
The vertical-load valve: no runaway, no drift, controlled lowering.
A pilot-assisted relief valve mounted on the load side of a cylinder or motor. It stays CLOSED under the load pressure and opens only when the opposite side is actively driven — with back-pressure preventing runaway.
Load pressure alone cannot open it (setting chosen ~1.3× load pressure). Driving the other chamber pilots it open through the pilot piston (pilot ratio 3:1…10:1): the lowering pressure then modulates a smooth relief. Integral check valve gives free flow in the load-lifting direction.
Pilot ratio high (8–10): responsive, stable for static loads; low (3:1): for unstable/cycling loads. Mount directly on the cylinder port — hose-burst protection demands zero flexible line between valve and actuator.
The Studio implementation of load-holding valves — with a-side/b-side spring orientation.
The Studio implementation of load-holding valves — with a-side/b-side spring orientation.
The brake valve component models the counterbalance/brake function inside the simulator, including which port the spring side belongs to (a-side vs b-side spring) so the free-flow direction matches your circuit.
Free flow through the check in one direction; throttled/pilot-controlled relief in the other. The spring-side choice swaps which cylinder chamber is protected — exactly like choosing the mounting side of the physical valve.
Set the cracking pressure ~1.3× the expected load pressure; choose the spring side toward the cylinder port that carries the load. Verify in both directions: free lift, controlled lower.
One-way traffic: free flow one way, blocked the other.
One-way traffic: free flow one way, blocked the other.
A spring-loaded poppet or ball opens with a small cracking pressure (0.5–5 bar) in the free direction and seals metal-to-metal in the blocked direction.
Flow lifts the poppet against the spring; reverse pressure seats it harder (the sealing is pressure-assisted). Variants: without spring (any mounting orientation), with dashpot (anti-hammer), as cartridge inserts.
Cracking pressure: low for suction/bypass duty, higher when it must hold against vibration. Δp at rated flow 1–6 bar. Seat leakage essentially zero — the standard choice where a load must not creep.
Locked until permitted: leak-free load holding with a pilot key.
Locked until permitted: leak-free load holding with a pilot key.
A check valve that can be opened from the OUTSIDE: a pilot piston driven by the opposite line pressure lifts the poppet, releasing otherwise trapped oil.
At rest it traps the cylinder leak-free (better than any spool valve). Driving the opposite side pressurizes the pilot port → the pilot piston (area ratio 3–4×) opens the held chamber → the load lowers in a controlled way through the driving valve’s metering.
Single and double versions (double locks BOTH cylinder chambers — standard on outriggers). Ratio 3:1 to 4:1 typical; high held pressures need staged/dampened poppets against shock opening.
An OR gate in oil: the higher-pressure input wins the outlet.
An OR gate in oil: the higher-pressure input wins the outlet.
Two inlets, one outlet. A free ball or spool seals the lower-pressure inlet and connects the higher-pressure one to the outlet — hydraulic OR logic in one fitting.
Pressure at inlet 1 shifts the ball against inlet 2’s seat: outlet = inlet 1. When inlet 2 becomes higher, the ball travels back. Response is fast; the idle side is vented through the active side if it leaks to tank.
Both inlets must be pressure-free when unused, or return-flow paths defined. Small Δp (~1–3 bar) between inputs needed for reliable switching.
A variable orifice: speed control whose price is load dependence.
A variable orifice: speed control whose price is load dependence.
An adjustable restriction (needle in seat, or slotted spool). Flow follows the orifice law — it depends on BOTH opening and pressure drop, which is why plain throttles change speed with load.
Reducing the opening raises Δp across the edge; the actuator gets less flow. Heat exacts payment: the pressure drop × flow converts to warmth. Double-acting use needs two throttles or one bypass-equipped unit.
Fine threads give resolution (turns-from-closed calibration!). Sharp-edged orifices are viscosity-insensitive; long-screen throttles drift with oil temperature.
The professional speed control: set flow stays constant whatever the load does.
The professional speed control: set flow stays constant whatever the load does.
An adjustable orifice IN SERIES with a compensator spool that keeps the Δp across the orifice at a constant 5–10 bar. Constant Δp + constant opening → constant flow, independent of load and supply pressure.
The compensator balances orifice inlet vs outlet pressure on its spool faces with a spring (the Δp setpoint). Rising load? Compensator opens. Rising supply pressure? Compensator throttles harder. Excess supply energy is burned at the compensator — or avoided with a bypass/3-way version.
Flow ranges from 0.05 l/min to hundreds; minimal control Δp ~10 bar. Provide a free-flow check if reverse flow needed. Priority (3-way) versions dump excess flow without heating.
One flow in, two equal flows out — synchronized cylinders without electronics.
One flow in, two equal flows out — synchronized cylinders without electronics.
A floating spool with metering notches splits the inlet into two (nearly) equal part flows regardless of branch pressures — or combines them in reverse. In the Studio it now models this faithfully: 50/50 metering including load-independent synchronization.
Each part flow crosses a fixed and a variable orifice; the spool shifts so both branches see equal Δp — hence equal flow. If one branch stalls, its pressure rises to the relief (a real spool divider is not a pressure equalizer: stall one side and the other side lifts double!).
Division accuracy 2–5 %; better with symmetric design and mid-loads. For long-stroke absolute sync use re-phasing (end-point correction valves) — dividers accumulate error over many strokes.
The hydraulic battery: oil stored against compressed nitrogen.
The hydraulic battery: oil stored against compressed nitrogen.
A pressure vessel with a rubber bladder: nitrogen pre-charge p₀ inside, oil outside. Working pressure compresses the gas; the stored volume is available on demand within milliseconds.
Below p₀ the bladder seals the oil port (it must NEVER be forced below). Between p₁ (min working) and p₂ (max), the usable volume follows Boyle’s law. Fast discharge is adiabatic, slow charging isothermal — the capacity differs.
Sizes 0.1–50 l, to 330 bar. Pre-charge with NITROGEN ONLY (oxygen explodes with oil mist!). Check p₀ yearly; bladder fails fast if cycled below p₀.
Isolation, relief and drain: the mandatory hardware around stored energy.
Isolation, relief and drain: the mandatory hardware around stored energy.
DIN/ISO rules require every accumulator to be isolatable, depressurisable to tank, and protected by its own relief valve. The safety block packages all three functions plus the gauge port.
An isolation ball valve separates the accumulator from the system; a drain (manual or solenoid) vents the oil side to tank; a relief limits pressure AT the accumulator even if the system relief is set higher or isolated.
Block relief ≤ accumulator rating. Before ANY work on the circuit: isolate, drain to tank, verify 0 bar at the gauge. Nitrogen pre-charge rigs are the only permitted way to set p₀.
High-pressure fortress: protects valves right after the pump.
High-pressure fortress: protects valves right after the pump.
A filter built for full system pressure, placed between pump and valve group. Its μ-rating and β-ratio decide what cleanliness your proportional/servo valves live in.
Flow passes outside-in through the element; a bypass valve opens if the element clogs (no-starvation rule) — unless a no-bypass version protects servo valves absolutely. A clogging indicator (visual/electrical) reports β-limit.
β10 = 200 means 99.5 % capture at 10 μm. Size for 1.5–2× pump flow at cold viscosity. Servo valves demand 3–5 μm absolute; gear pumps tolerate 25 μm.
The two quiet janitors: return filtration for all flow, strainers for pump life.
The two quiet janitors: return filtration for all flow, strainers for pump life.
The return filter sees 100 % of circulating oil at low pressure — the standard plant-wide cleanup point. The suction strainer is a coarse screen (100–150 μm) protecting only the pump, NOT the system.
Return oil flows through the element to a clean tank; bypass at ~1.5–3 bar prevents back-pressure. Suction strainers sit below oil level; Δp limits are tight (≤ 0.05 bar at cold start) or the pump cavitates.
Return filters: β10–25 with electrical clog indicator as standard. Strainers are NOT a substitute for filtration — they only stop rocks. Never starve the suction side with a fine filter.
Every throttle watt leaves as heat — the cooler exports it.
Every throttle watt leaves as heat — the cooler exports it.
Air-blast units (fan over finned block) go into return or bypass lines; water-oil plate coolers squeeze heat into cooling water. Both extend oil and seal life — temperature is the oil’s aging clock.
Sizing starts from the heat balance: Ploss ≈ 20–30 % of installed power in typical machines (100 % in relief dwell!). The cooler must dissipate that at the allowed oil temperature (≤ 60 °C shop standard) at the worst ambient.
Mount with a bypass check (cold-start protection) and after filtering. Fan noise scales with blade size at equal duty — two smaller fans are quieter.
Cold oil cavitates and sticks: gentle pre-heating protects the pump and valves.
Cold oil cavitates and sticks: gentle pre-heating protects the pump and valves.
A screw-in electric element (or plate heat exchanger) that brings the tank to minimum start viscosity. It must stay submerged, and its surface loading must stay low to avoid cracking the oil.
Below ~10–15 °C, VG 46 oil thickens so much that suction Δp and valve response suffer — cavitation follows. The heater thermostat (typ. set 20–30 °C) pre-warms; a thermostat interlock can block the main motor until the tank is ready.
Surface load ≤ 1–2 W/cm² (oil scorching starts local coking at hot spots). Use with level protection — a dry element burns out in minutes and contaminates the tank.
Flexible and replaceable — but only as strong as its routing rules.
Flexible and replaceable — but only as strong as its routing rules.
Multi-layer: inner tube, 1–6 wire reinforcements (braid/spiral), cover. Rated by bore (DN), working pressure with ≥4:1 burst safety, and impulse endurance.
Under pressure the hose shortens up to 4 % and stiffens — routing must allow movement without twisting (torsion destroys wire layers fastest). Minimum bend radius and straight exits after fittings are non-negotiable.
Choose by pressure (with peaks!), fluid compatibility, temperature and bend radius. Never combine bending with torsion in the same hose; protect against abrasion with sleeves/spacing; date-code and replace by age.
Seamless precision tube, bent once — the cheap, clean default for fixed runs.
Seamless precision tube, bent once — the cheap, clean default for fixed runs.
Cold-drawn seamless steel (or stainless) tube to DIN 2391/EN 10305, connected by cutting rings, flares or welds. Cheaper than hose, zero aging, no diffusion — the reference for stationary machines.
Bend radii ≥ 3× OD, clamp spacing by diameter (L = √(d) rules), and stress-free alignment prevent fatigue cracks. Galvanic isolation in clamps avoids corrosion points.
Wall thickness from Barlow formula with fatigue factor; target velocities per line class (suction 0.5–1.5, pressure 3–6 m/s). Deburr and FLUSH all cut tube — assembly swarf kills valves downstream.
M16 test couplings: the 10-second pressure window into a live machine.
M16 test couplings: the 10-second pressure window into a live machine.
Quick-connect pairs (poppet or flat-face) and the tiny M16×2 test points with captive check. Flat-face versions minimize spill and air inclusion — the professional default.
Test points seal by an internal valve opened only by the mating test hose: connect, read, disconnect — no oil loss, no shutdown. Quick couplings for attachments lock by sleeve; rated for full system pressure.
Flat-face for frequent cycles & cleanrooms; poppet for budget. Test points at every diagnostic location: pump, relief, valve P/A/B, cylinder ports.
Drilled blocks replace pipes: compact valve stations with gallery channels.
Drilled blocks replace pipes: compact valve stations with gallery channels.
An aluminum/steel block with ISO 4401 surfaces on the faces and drilled galleries inside. Cartridge (logic) elements push the concept further: big flows inside the block, pilot valves on top.
Galleries connect P, T, A, B between stations; sandwich valves stack between DCV and block. Intersecting drillings are closed by tapered plugs. Rule: generous galleries (low Δp!), no crossing collisions, test points planned, expansion plugs allowed.
Check Δp < 5–10 % at design flow through each gallery. Cartridge logic elements handle 1000+ l/min — way beyond any spool valve, with poppet tightness.
Screw-in seat valve: electrically switched, zero-leakage isolation in manifolds.
Screw-in seat valve: electrically switched, zero-leakage isolation in manifolds.
A cartridge pushed into a manifold cavity: coil ON/OFF strokes a poppet. Normally closed forms act like safety gates inside blocks — with seat tightness where spools would leak.
The coil pulls an armature; the poppet lifts against spring and pressure. Damped two-stage versions avoid water hammer at large flows. The Studio’s cartridgeSolenoidValve22 with its energized property models exactly this actuation.
Voltage 24 V DC standard; duty 100 %; manual override pin for service. Watch coil current → PLC output capacity (0.5–2 A).
The mechanic’s window: coiled tube, pointer — and correct scale choice.
The mechanic’s window: coiled tube, pointer — and correct scale choice.
A flat C-tube straightens under pressure, moving a pointer over the dial. Glycerine filling damps vibration and pulsation — unfilled gauges rattle themselves into early death.
Choose full scale so working pressure sits at ~75 %. Reading = relative (bar gauge). With gauge isolators (6-port selector), one gauge serves many points without continuous exposure.
Classes 1.0–2.5 % FS. Protect from pulses with snubber or isolator. A dead-needle gauge on a hammering relief circuit is a diagnostic signal in itself.
Pressure becomes a contact: the B1 of every electrohydraulic ladder.
Pressure becomes a contact: the B1 of every electrohydraulic ladder.
A piston/diaphragm element that trips a microswitch at an adjustable threshold — with adjustable hysteresis on quality units. It turns "clamp pressure reached" into a PLC input.
Pressure strokes the sensing element against a spring; the trip point = spring preload. Reset at a lower value (the dead band) — in the Studio it now models proper Schmitt behavior (pickup at setting, dropout at ~95 %).
Ranges 5–400 bar; repeatability ~1 %; hysteresis adjustable 3–10 %. Mount where the pressure proves the function (at the clamp port, not the pump!).
The oil’s guardian: alarm at 60, shutdown at 70 — before seals cook.
The oil’s guardian: alarm at 60, shutdown at 70 — before seals cook.
A bimetal or PT100-based switch in the tank with fixed/adjustable thresholds. Above 80 °C, oil oxidation rate doubles per 8–10 K and seals age in dog years — hence the strict guard rails.
Contact changes state at the setpoint (with a small built-in hysteresis). In the Studio it switches on the simulated oil temperature (metrics) with a −2 °C drop-side — exactly the field device behavior.
Typical setpoints: warning 60 °C, cooler start 45–50 °C, shutdown 70–80 °C. Combine with level switch so a low tank cannot fake-cool system damage.
The truth about delivery: pump wear and leakage visible in l/min.
The truth about delivery: pump wear and leakage visible in l/min.
Rotation-based meters — turbine for low-viscosity steady flow, positive displacement (gear/oval wheel) for oil. They turn "feels slow" into a measured number.
The rotor speed is proportional to flow; displacement meters count actual volumes (best accuracy for oil, tolerate the DP they need). Portable test meters link two points with pressure gauges: pump health under load in minutes.
Accuracy 0.5–2 %; pressure drop 0.5–3 bar at rating. Always compare readings at same temperature — viscosity shifts turbine response.
1450 rpm into torque: bellhousing, flexible coupling, alignment discipline.
1450 rpm into torque: bellhousing, flexible coupling, alignment discipline.
Standard IEC motors (4-pole ≈ 1450 rpm at 50 Hz) drive the pump through a bellhousing and an elastic jaw coupling. The assembly is rigid, quiet and self-aligning within limits.
The coupling tolerates small misalignments and damps torsional shocks (relief cracking is a torsional event!). Motor sizing: shaft power ≈ p·Q/(600·η) with service factor 1.1–1.25.
Never oversize blindly: oversized motors run inefficiently far below rated load. Check rotation direction at first start (dry reverse running scars gear pumps for life).
Where oil meets logic: 24 V DC solenoids and how the Studio simulates real control.
Where oil meets logic: 24 V DC solenoids and how the Studio simulates real control.
Every solenoid valve is a border between power and signal. In the Studio the border is real: PLC rungs, relay contacts, timers and sensor inputs shift the valves — delete PLC1 and the machine honestly stops.
A PLC scan evaluates rungs (conditions) and writes outputs (solenoids via their target valve). Latched rules "COND -> V1 right" hold actions; combinatorial "Q1 = COND" recompute every scan. Timers add dwell, sensors (pressure, reed) close the loop.
One coil per direction command, one relay/memory per state. Solenoid default de-energized = safe start. Check coil current vs PLC output rating.
Plumbing grammar: branch lines properly — and seal spare ports with blind plugs.
Plumbing grammar: branch lines properly — and seal spare ports with blind plugs.
Junctions split or merge lines; the blind plug seals an unused port pressure-tight. In the Studio an unplugged valve port otherwise "vents to tank" invisibly — the plug makes sealing a deliberate design act.
T-pieces feed multiple consumers from one supply; sharp Ts add measurable Δp. The blind plug completes the symbol honesty: what you cap is what the simulation treats as closed.
Seal every unused P/A/B port with rated plugs, never tape. In manifolds, expansion plugs close future stations — same principle, bigger scale.
Storage, cooling, de-aeration, settling: the quiet multitasker under the pump.
Storage, cooling, de-aeration, settling: the quiet multitasker under the pump.
More than a bucket: sized volume, suction/return zones separated by a baffle, breather, level gauge, cleanout covers and drain. Good tanks make oil last; bad tanks churn bubbles and dirt.
Volume ≈ 2–4× pump flow/min gives residence time for air release and settling. Return oil enters submerged below surface (no foam!); suction draws from the quiet zone; the breather filters incoming air as level cycles.
Fill level above suction strainer at all attitudes. Breather rating ≥ pump flow in air equivalent. Slope the bottom to the drain; paint-free interior (phosphate coating).
The dash-dot line that says "these parts are bolted together".
The dash-dot line that says "these parts are bolted together".
ISO 1219 marks physically grouped components (power pack contents, manifold contents, FRL stages) with a dash-dot enclosure. It is a drawing convention that carries real assembly information.
Inside the frame: shared hardware, one article number. In the Studio the manifold/HIC boundary lets you group cartridges the same way — your schematic then matches the real manifold on the bench.
Use enclosures generously: any purchased unit with one part number deserves one. Label the frame with the unit tag (B10, A2…).
The quiet cousins: crescent-sealed tooth sets for low-noise duty.
The quiet cousins: crescent-sealed tooth sets for low-noise duty.
An inner ring gear drives an outer ring with one more tooth, separated by a crescent. The design runs notably quieter and with less pulsation than external gears.
Teeth unmesh and mesh against the crescent seal: suction and pressure zones grow and shrink smoothly, so flow ripple is small. Tight internal clearances demand cleaner oil than external gears.
Pressures to 300+ bar in premium lines; prized for servo-adjacent duty and low-noise halls. Sensitive to cavitation — watch suction conditions.
Sliding vanes in an eccentric ring: smooth, quiet, often adjustable.
Sliding vanes in an eccentric ring: smooth, quiet, often adjustable.
A slotted rotor carries vanes that slide against a cam ring. Offset rotor-to-ring distance sets displacement — shifting the ring (variable type) changes volume per revolution.
Chambers between vanes grow at suction and shrink at pressure. A pressure-compensated variable vane pump destrokes at its setpoint: constant pressure, flow on demand — a budget alternative to piston compensators.
To 160–210 bar; quiet; medium dirt tolerance. Variable versions to ~160 bar with compensator kits.
Silent continuous transport: spindles rolling oil without pulsation.
Silent continuous transport: spindles rolling oil without pulsation.
Two or three meshing screw spindles form sealed axial chambers that travel steadily from suction to pressure — no pulsation, near silent, outstanding for lubrication circuits.
Axial piston forces are balanced; delivery is perfectly smooth. Tolerates moderate viscosity range; absolutely intolerant of dry running.
To 250 bar (3-screw), best ≤ 40 bar for lube duty; quietest pump family; fixed only.
The last 20 mm matter: braking the piston inside the cylinder.
The last 20 mm matter: braking the piston inside the cylinder.
A cushioning spear closes the main outlet near stroke end, forcing oil through an adjustable throttle — the piston decelerates on an oil brake instead of slamming the cap.
The spear enters the port, outflow crosses the cushion screw → pressure rises in the trapped volume → smooth deceleration. Set the screw so the rod arrives with ~0 speed without bouncing.
Use when m·v is meaningful (fast or heavy axes). Check max cushion pressure against port ratings; adjustable vs fixed cushions.
A spacer inside the cylinder that saves rods and bearings on long strokes.
A spacer inside the cylinder that saves rods and bearings on long strokes.
A stop tube lengthens the effective piston-to-bushing distance, reducing bushing loads from moments and side forces. Cheap insurance for long-stroke, high-moment installations.
Side load F·e concentrates where piston and bushing act as lever-points; the stop tube moves the piston farther from the bushing at full extension → lower bearing pressure.
Rule of thumb: 25 mm stop tube per 250 mm stroke beyond ~600 mm. Better: eliminate side load with self-aligning couplings and proper mounts.
The rod reports where it is: integrated stroke transducers for closed loops.
The rod reports where it is: integrated stroke transducers for closed loops.
A magnetostrictive wire in a pressure-proof rod bore reads a ring-magnet attached to the piston: absolute position, 1–5 μm class resolution, no contact.
A current pulse travels the wire; the magnet’s field twists the return — the time-of-flight = position. Output analog 4–20 mA/0–10 V or bus (SSI, IO-Link).
Strokes to 2+ m, pressure rating of the cylinder unchanged. The sensor port needs a clean connector; protect the pigtail.
Micron-class metering for the fastest axes — and the cleanest oil requirement.
Micron-class metering for the fastest axes — and the cleanest oil requirement.
A torque-motor-driven pilot (flapper-nozzle or jet pipe) positions a lapped spool with micron precision. Step responses in the low milliseconds, hysteresis < 0.5 % — expensive, precise, fragile.
The pilot stage meters a tiny control flow that strokes the main spool proportionally to current — with mechanical or electrical feedback closing the position loop inside the valve itself.
Absolute filtration 3–5 μm (β≥200). Rated flow at Δp = 70 bar over the metering edges (datasheet convention!). Null drift tracked with hours/temperature.
Gain, ramps, deadband, dither: the four knobs that shape every proportional motion.
Gain, ramps, deadband, dither: the four knobs that shape every proportional motion.
The amplifier converts a command voltage into a controlled coil current — with the parameters that make or break motion quality: gain (scaling), ramp up/down (slewing), deadband compensation, dither.
Command → reference (with ramps) → error amp → PWM current driver. Current control makes coil force independent of coil heating; dither (100–200 Hz ripple) keeps the spool micro-sliding to avoid stiction.
Set ramps to kill pressure shocks first; then gain; THEN deadband. Dither amplitude highest value with no visible actuator tremble.
Piloted from the accumulator: dump the pump when the store is full.
Piloted from the accumulator: dump the pump when the store is full.
Externally piloted relief-family valve: pilot pressure from the accumulator (or another line) switches it open → pump unloads to tank at near zero pressure; a check isolates the charged consumer.
Below the setting the pump charges the accumulator through the check valve; at setting, pilot pressure trips the unload spool → pump idles while the accumulator feeds the circuit. Hysteresis band prevents chattering.
Set unload ~10–15 bar above max consumer pressure; pair with safety block relief. The energy-efficient classic for intermittent demand.
Gravity feeds the giant: filling press cylinders from an elevated tank.
Gravity feeds the giant: filling press cylinders from an elevated tank.
A large check valve between an overhead tank and the press ram: during fast approach the ram draws oil directly by gravity; during pressing it closes; during return, pilot pressure opens it for decompression and refill return.
Approach: cylinder volume grows faster than the pump can feed → prefill opens, oil floods in. Press: pump pressure closes it, full tonnage builds. Return: the pilot opens it so the huge volume escapes to the tank.
Sized by approach speed × area (hundreds of l/min). Must open softly (anti-cavitation/decompression staging) — slam causes audible bangs and line damage.
Anti-shock pair across a motor: no pressure spike survives a reversal.
Anti-shock pair across a motor: no pressure spike survives a reversal.
Two reliefs + two checks in antiparallel across a hydraulic motor: each direction has its own relief path and makeup feed. The standard protection for rotating inertia.
Closing the DCV with a spinning motor turns the motor into a pump: the starved side sucks through the check, the pressurized side relieves through its valve — deceleration without cavitation or spikes.
Relief setting ~10–20 % above working pressure; makeup from tank via checks (never block anti-cavitation paths!).
The parachute: locks the load if a line ruptures.
The parachute: locks the load if a line ruptures.
A small cartridge in the cylinder port that stays open in normal flow and slams shut when return flow exceeds a trigger rate — fusing the circuit mechanically at a burst.
Normal lowering flow passes; a burst accelerates flow → Δp on the sensing edge rises → the valve element seats → the load stops trapped at the cylinder port. Reset by lifting pressure.
Trigger ~1.5× max professional lowering flow. Must sit directly ON the cylinder (internal leakage = slow descent, not drop). Test by procedure, not by rupture hope.
One gauge, six truths: selector isolators for commissioning panels.
One gauge, six truths: selector isolators for commissioning panels.
A rotary selector valve connecting one gauge to 2–6 measurement points — the gauge lives protected and offline until a measurement is dialed in.
Rotate to a point: the isolator admits that line to the gauge, venting the previous. Between readings the gauge stands at atmosphere: no pulsation fatigue, no stolen volume.
Never leave it on a hammering point unattended. Mark points on the panel map (A1 = pump, A2 = clamp…).
Low oil lies: float guards stop the pump before the pump stops itself.
Low oil lies: float guards stop the pump before the pump stops itself.
Float switches at min. level plus bimetal temperature contacts — the standard tank interlocks. Low level means vortex air into the suction; high temperature cooks seals.
Magnets in the float trip reed contacts at thresholds; wired as NC chains so a broken wire reports safe. Temperature contacts warn (60 °C) and interlock (70–80 °C).
Also size the visual level/temperature gauge — the redundant eye always wins. Mount accessible for cleaning (sludge kills floats).
The tank inhales too: filter what enters, or invite dust storms inside.
The tank inhales too: filter what enters, or invite dust storms inside.
Every pump stroke cycles the level and breathes air through the cap. The breather element (3–10 μm) or desiccant cartridge (silica gel + filter) protects oil from dust and humidity.
Desiccant beads change color as they saturate with moisture; pleated elements capture particles. Undersize the breather and the tank puffs oil mist around the cap.
Air rating ≥ 1.5× pump flow. Replace on schedule or color; in humid/coastal sites desiccant is mandatory.
The honest isolator: full bore, quarter turn, visible ON/OFF by handle.
The honest isolator: full bore, quarter turn, visible ON/OFF by handle.
A drilled ball rotated by the handle: open = full bore (negligible Δp), closed = bubble-tight with seats. Two-way and three-way (selector/L or T port) versions for gauges and service.
Soft or metal seats seal; the handle ends in the flow direction indicator. Not for throttling — seats erode in a crack position.
Rated by pressure class (PN 350/500 in hydraulics). Lockable dies exist for lockout-tagout discipline.
High pressure from low: the area-ratio transformer in oil.
High pressure from low: the area-ratio transformer in oil.
Two linked pistons: low pressure drives a big piston, the small one delivers pressure × (Abig/Asmall). Oscillating versions pump continuously; single-stroke for clamps.
Force balance: pout = pin × ratio. The trade is volume: out flow = in flow / ratio (power conserved minus losses). No relief needed on the HP side when correctly sized.
Ratios 1.5–20. HP side needs HP-rated fittings & hoses. Combine with accumulator for sustained HP flow.
Let the cylinder suck: feeding the starved chamber during overruns.
Let the cylinder suck: feeding the starved chamber during overruns.
Check valves from tank/return into working lines, opening whenever local pressure collapses. They close the "pumping void" that inertial loads create — the partner piece to cross-line reliefs.
Overrunning load: driven side may cavitate as the actuator outruns the supply. Makeup checks admit return-side/tank oil into the starving chamber → no vapor bubbles, no diesel knock.
Cracking as low as possible (0.2–0.5 bar). Size for the full chase flow, not the nominal bleed.
Finger power for big valves: 5–30 bar pilot circuits of mobile machines.
Finger power for big valves: 5–30 bar pilot circuits of mobile machines.
A small manual pressure-reducing valve per axis axis: joystick angle → proportional PILOT pressure → main spool stroke. The operator commands hydraulics through hydraulics.
Each axis has two reducing cartridges; the pilot signal pushes the main valve against its centering springs proportionally. An independent low-pressure pilot supply (gear pump + relief) feeds all joysticks.
Pilot supply typically 20–40 bar, 5–15 l/min. Hysteresis/linearity of the joystick cartridge = the machine’s fingertip feel.
The end-position ear: metal nearby, contact closed, no touching.
The end-position ear: metal nearby, contact closed, no touching.
A coil oscillator damped by metal in its field; the output flips at the rated distance (1–15 mm). Immune to oil, dirt, vibration — the standard position feedback on hydraulic cylinders and slides.
A target on the moving part approaches the face → oscillator damps → Schmitt stage switches PNP (+24 V) output. Sensing range derates for non-steel materials.
PNP 24 V DC 3-wire is the shop standard; M12/M18 housings. Secure the air gap as specified (flush vs non-flush).
The 24 V muscle on every valve — voltage types, LEDs, manual overrides.
The 24 V muscle on every valve — voltage types, LEDs, manual overrides.
The wet-pin solenoid converts coil current to stroke force; the standardized plug (DIN A/B/C) carries power, the diagnostic LED and sometimes suppression diodes.
Energize: armature lifts, pin strokes the spool. De-energize: the spring returns. AC coils buzz when the armature can’t seat; DC coils are quiet and PLC-friendly — hence 24 V DC as the default.
Duty 100 %; power 3–30 W; check inrush on big AC coils. The LED in the connector is the fitter’s first diagnostic — use it before touching tools.
Water eats what fans cannot: compact heat export for dense halls.
Water eats what fans cannot: compact heat export for dense halls.
Brazed plate packs route oil and water in alternate channels — extreme heat transfer per volume. Choose when ambient air is hot, dirty, or noise-capped.
Heat duty = water flow × ΔT water ≈ oil loss to export; a water-side thermostat valve keeps oil from overcooling at winter startups.
Keep water pressure below oil pressure (leaks migrate to the safer side). Watch fouling on open towers; provide strainers both sides.
The Σζ budget: where pressure goes besides the straight pipe.
The Σζ budget: where pressure goes besides the straight pipe.
Every fitting, elbow, coupling and valve contributes local loss ζ·ρv²/2. Real circuits lose more in fittings than in straight pipe — count them.
Catalog ζ-values rough: elbow 0.3–0.9, T-piece 0.5–1.8, coupling 0.1–0.3, open valve full-bore ~0.2, filters 0.5–3 bar full-scale. Sum with line friction; compare against the 5–10 % budget.
Velocity-square again: doubling bore ÷32 the straight part and ~÷16 local parts. Sharp edges at manifold entries are silent watts.
Fire code, temperature band and seals decide the oil long before price.
Fire code, temperature band and seals decide the oil long before price.
HLP mineral oil is the default; HFC water-glycol and HFD-R/U synthetics serve fire-risk halls per ISO 12922. Each change touches seals, paints, filters and pump ratings.
HFC: ~45 % water — fire resistant, limited to ~50 °C, derate pump speeds; HFD phosphate ester: mineral-like performance, needs FKM/EPDM seals and special paint; HFD-U esters: biodegradable premium.
Never mix; full flush on conversions incl. dead legs. Check pump maker approvals (bearing lubrication ratings differ).
Load it against the relief, measure delivery, know your volumetric truth.
Load it against the relief, measure delivery, know your volumetric truth.
The field test for pump health: throttle the outlet against measured pressure, log flow at several points — the Q(p) droop exposes internal wear better than any listening.
With relief closed slowly via test rig: flow at 0 → system pressure points; drop from nominal = volumetric efficiency loss > 10 % → overhaul candidate. Temperature-logged for fairness.
Measure at rated viscosity window (40 ± 5 °C); record motor current too. Compare to commissioning baseline — trends beat absolutes.
Area ratios multiply pressure where you do not expect it — return lines become pressure lines.
Area ratios multiply pressure where you do not expect it — return lines become pressure lines.
Pressure anywhere times area ratio appears somewhere: capped annulus with load, meter-out on big φ, sealed thermal volumes. Designers must find these pockets BEFORE the hose does.
Example: cap side 100 bar on a φ=2 cylinder traps 200 bar on the rod side when pushing against a blocked rod. Locked fluid + sunshine = 10 bar/K — every °C of thermal expansion adds pressure.
Rate everything on intensifiable paths for the intensified value; provide relief/thermal valves on trapped volumes; paint-warning such lines.
ρ·c·Δv: the thousand-bar kiss of the snapping valve.
ρ·c·Δv: the thousand-bar kiss of the snapping valve.
Instantly stopped flow converts kinetic energy to a pressure wave traveling at c ≈ 1000–1400 m/s in oil. Long lines + fast valves + speed = spikes that flatten hoses.
Δp = ρ·c·Δv (Joukowsky): doubling speed doubles it; longer lines lengthen the wave duration until 2L/c passes. Slow the valve (proportional ramp!), soften with accumulators, shorten lines.
The countermove list: ramps, slower cartridge stages, shock accumulators near the valve, hose elasticity sometimes helps (steel pipe worst case).
The classic tractor circuit: at neutral the pump flow circulates at nearly zero pressure.
The classic tractor circuit: at neutral the pump flow circulates at nearly zero pressure.
An open-center directional valve connects P to T in neutral. The pump delivers full flow back to tank with only line losses, so almost no heat is generated while the machine idles. Simple, cheap, robust.
Move the spool and the P→T path throttles shut while A or B opens. Because P is unloaded in neutral, a fixed pump plus a relief valve is enough — no unloading valve needed. The penalty: only one function moves comfortably at a time, because the flow takes the path of least resistance.
Choose for simple machines with one movement at a time (log splitters, tail lifts). Avoid when several axes must run simultaneously: load disappears into the open center.
Pressure always ready: a pressure-compensated variable pump keeps P at the stand-by setting.
Pressure always ready: a pressure-compensated variable pump keeps P at the stand-by setting.
Closed-center valves block P in neutral, so the pump must not be a dead-headed fixed pump. A pressure-compensated piston pump holds the line at its compensator setting, delivering only leakage flow.
When a spool opens, pressure at the pump outlet drops below the compensator setting and destrokes the swash plate until flow matches demand. The compensator is effectively a pressure regulator acting on displacement.
Standby pressure 20-30 bar below working peaks drains energy; keep compensator differential and response in mind. Each function sees full system pressure — check weakest component rating.
Two pumps, two personalities: big flow for speed, small flow for force.
Two pumps, two personalities: big flow for speed, small flow for force.
A large and a small pump work together into P. At low pressure both deliver (fast approach); when pressure rises past the unloading valve setting, the big pump is dumped to tank at near zero pressure and only the small pump finishes the high-pressure stroke.
The unloading valve opens at the switch pressure and connects the large pump outlet to T through a check-protected branch. The check valve keeps system pressure on the actuator while the big pump idles. Total installed motor power stays small because at high pressure only the small flow remains.
Set the unloading valve above the fast-approach pressure plus margin, below the relief. Check valve on the high-pressure side is mandatory. Size small pump flow for the pressing speed.
Charge, unload, sleep: the accumulator holds the pressure while the pump rests.
Charge, unload, sleep: the accumulator holds the pressure while the pump rests.
The circuit charges an accumulator to the relief proximity, then an unloading valve dumps pump flow to tank at ≈ 2 bar. The actuator holds pressure from the accumulator until its pressure drops to the reload threshold, then the cycle repeats.
A piloted unloading valve senses accumulator pressure across a check valve and switches the pump between load and dump. Differential pressure (typically 15-25 %) defines reload frequency and pump duty cycle.
Size the accumulator from the usable volume between p2 (working) and p3 (load pressure), precharge p0 = 0.85-0.9 × p2. Pump duty cycle below 40 % saves energy but stress-cycles the relief hardware — verify with the machine duty profile.
Throttle the oil INTO the cylinder: accurate speed, but dangerous with pulling loads.
Throttle the oil INTO the cylinder: accurate speed, but dangerous with pulling loads.
A flow control valve sits between the directional valve and the cylinder inlet. The pump delivers constant flow; the throttle decides how much may enter the actuator, excess escapes over the relief valve.
Because the directional valve sees full pump delivery while the actuator receives less, the surplus oil crosses the relief at full system pressure — that is why throttling control is warm but wonderfully simple. Speed stays constant only while load stays constant.
Use on loads that always push against the cylinder (resistive loads). Never on overrunning loads: the rod side can cavitate or the load runs away. Add a check branch for free reverse flow.
Throttle the oil OUT: back-pressure holds any load, at the cost of pressure intensification.
Throttle the oil OUT: back-pressure holds any load, at the cost of pressure intensification.
The flow control restricts the return line from the cylinder. The load always works against a pressurized cushion, so even overrunning loads move at controlled speed.
The restriction creates back-pressure on the rod side. With a differential cylinder (area ratio φ ≈ 2) that back-pressure is multiplied on the rod side relative to the cap side — the rod seals and hose must survive up to φ × the inlet pressure.
Rule of thumb: meter-out as default for unknown or overrunning loads, then rate return-side components for the intensified pressure. Watch thermal growth: the throttled power becomes heat directly in the return line.
Divert the surplus instead of blocking it: cooler throttling, mushier speed.
Divert the surplus instead of blocking it: cooler throttling, mushier speed.
The flow control sits in a branch from P to T. Pump flow above the set value escapes through the throttle at the actual working pressure, not at relief pressure — substantial energy savings.
Speed equals pump flow minus bled flow, so any load change that shifts working pressure changes the bleed and therefore the speed: speed stiffness is poor. Only suitable where approximate speed is enough.
Combine with a fixed orifice + adjustable needle for coarse-fine control. Never use where exact speed under varying load matters.
Two valves in one: a throttling pin plus a compensator piston that keeps Δp constant.
Two valves in one: a throttling pin plus a compensator piston that keeps Δp constant.
Because throttle flow follows Δp, load changes would change speed. The pressure compensator senses the pressure drop across the metering orifice and opens or closes a second spool to hold that drop (typically 8-14 bar), making flow independent of both load and pump pressure.
Two-port (2-way PC) versions keep the excess on the pump (heat, like meter-in, but stiff); three-port versions dump surplus to tank at working pressure and act like bleed-off with load-insensitive speed.
Check minimum pressure window: compensation needs the compensator Δp plus throttle Δp, so below ≈ 20 bar supply the regulation dies. Prefer 3-port for energy, 2-port for simple plumbing.
Steal the return oil: rod-side flow joins the cap side and the cylinder flies.
Steal the return oil: rod-side flow joins the cap side and the cylinder flies.
During extension the oil expelled from the rod side is routed back into the cap side instead of tank. The cylinder extends with pump flow plus recycled rod flow at the cost of force (force works only on the annulus area).
With φ = 2 (cap area twice the rod area), regeneration doubles extension speed and halves extension force. At the end of approach a sequence or pressure switch kicks the circuit into normal mode for the working stroke.
Only useful with regenerative-capable valves (P→A, B→A paths) or dedicated regen check configurations. Verify buckling: the fast stroke still carries the full-load rod.
Pressure is the clock: cylinder 2 starts only after cylinder 1 has built force.
Pressure is the clock: cylinder 2 starts only after cylinder 1 has built force.
A sequence valve in the supply of cylinder 2 stays closed until the pressure from cylinder 1 (the clamp) reaches its setting. Only then does flow reach the work cylinder. Order is guaranteed hydraulically, no sensors needed.
Pressure builds when a cylinder finds its end stop or the clamped part. The sequence valve opens at e.g. 70 bar, the second actuator runs, and on return both retract together (the sequence valve is bypassed by its integral check valve).
Set the sequence value well above the travel pressure of cylinder 1, below the relief. Remember: a leak on cylinder 1 can drop pressure below the sequence setting mid-cycle — diagnose with the pressure gauge before blaming valves.
Two cylinders share one pump — the divider forces each drop to split 50/50.
Two cylinders share one pump — the divider forces each drop to split 50/50.
A rotary (gear) flow divider or a spool-type divider splits inlet flow into two nearly equal parts regardless of the individual loads, so two cylinders move in step with 2-5 % accuracy.
Two gear sections on a common shaft act as coupled motor-pump pairs: the faster (lighter loaded) side is throttled by the coupling, the slower side is boosted. Overrun conditions can even pass power from one branch to the other.
Accuracy 2-4 % typical, pressure intensification possible on dead-headed branches (relief per outlet!). Synchronizing errors accumulate — re-phasing at stroke end via bypass orifice is standard practice.
Rod-out oil of cylinder 1 feeds cylinder 2: geometric synchronism for zero cost.
Rod-out oil of cylinder 1 feeds cylinder 2: geometric synchronism for zero cost.
The rod-side outlet of the first cylinder connects to the cap side of the second. If Acap,2 = Arod,1 exactly, both pistons move at identical speed — the transferred volume forces it.
Because the first pressure must drive two loads, system pressure sees the sum of both loads referred to the cap area of cylinder 1. Manufacturing tolerances and rod creep accumulate error, so end-stroke re-phasing valves or make-up check circuits are common.
Perfect for two cylinders only, and only where exact area matching can be ordered. Make-up oil paths and bleed screws at the high points are mandatory.
One system, two pressure levels: delicate parts clamped at 40 bar from a 160 bar supply.
One system, two pressure levels: delicate parts clamped at 40 bar from a 160 bar supply.
A pressure-reducing valve in a branch line limits downstream pressure to its setting, independent of the main system pressure. Upstream may fluctuate — the clamped workpiece feels only the reduced value.
The reducing valve is normally open: downstream pressure acts on the spool against the spring; when set pressure is reached the spool throttles or closes. Relieving versions can even vent downstream overpressure to T (important with thermal expansion).
Always relieving-type for clamping circuits (sun on a fixture!). Take the pilot drain to tank separately if back-pressure exists. Reduced pressure costs energy: the dropped bar times flow becomes heat.
Machine safety principle: one valve that makes pressure disappear on demand.
Machine safety principle: one valve that makes pressure disappear on demand.
A normally-open solenoid valve piped from P to T (venting the relief spring chamber of the main relief or the compensator line) de-pressurizes the whole system when the safety chain breaks. Stored-energy accumulators need their own dump path.
Venting the relief drops system pressure to the vent-line back-pressure (few bar). Category 0/1 stops use dump + motor stop; monitored dump valves with feedback contacts satisfy performance levels in ISO 13849 circuits when dual-channel.
Dump rate must de-energize the machine within the risk-assessment time. Manual dump for maintenance (LOTO) belongs at eye height with a lock facility.
Both hands or nothing — and re-press within 0.5 s if released.
Both hands or nothing — and re-press within 0.5 s if released.
Two manual valves in series (hydraulic AND) supply the directional valve pilot stage. Both must be pressed within the anti-tie-down window; releasing either one interrupts pilot pressure and returns the main spool to safe state.
Anti-tie-down logic uses a small pilot accumulator/valve block that outputs only on simultaneous fresh press events within e.g. 500 ms. Holding one button with a clamp generates no output because the timing element never completes.
Buttons spaced ≥ 300 mm (inside edge) defeat one-hand-one-elbow operation. Prefer monitored pneumatic/hydraulic cartridges certified for the required PL over home-built spool logic on modern machines.
The four-stroke of every serious press — and why decompression saves the machine.
The four-stroke of every serious press — and why decompression saves the machine.
Fast approach via regeneration or hi-lo, controlled pressing speed via flow control, dwell under pressure holding (closed center + pump unload), then a controlled pressure release BEFORE the fast return.
During dwell the cylinder stores elastic energy (compressed oil + stretched frame + compressed tooling). Opening the return path abruptly releases it as a pressure wave — the notorious press bang. Decompression circuits bleed the cap volume through a small orifice or a proportional ramp until < 10 bar, then open fully.
Dwell time comes from a timer or a pressure-maintaining circuit (pilot-operated check holds, accumulator tops up). Decompression time typically 0.5-3 s scaled by stored volume.
The load can never run away: it lowers only when pilot pressure invites it.
The load can never run away: it lowers only when pilot pressure invites it.
A counterbalance (load-holding) valve in the rod-side line of a vertical cylinder closes leak-free under the load. To lower, pressure on the cap side pilots the valve open with a defined pilot ratio (e.g. 3:1 or 10:1).
Pilot ratio is a lever: with ratio 3:1 and spring setting 1.3 × load pressure, cap pressure only needs one third of the margin to start lowering. Too stiff a spring = jerky starts; too soft = the valve never truly holds under hose-burst assumptions.
Mount ON the cylinder port — a burst hose between valve and cylinder is then harmless. Oversize setting 1.3 × maximum load pressure. External drain version when the return line carries back-pressure.
Stop a spinning motor by making it pump against a valve that says how hard.
Stop a spinning motor by making it pump against a valve that says how hard.
Two relief valves mounted anti-parallel across the hydraulic motor ports limit pressure in both directions. When the directional valve closes, the spinning inertia drives the motor as a pump; the crossline relief absorbs the kinetic energy as heat.
Each relief covers one rotation direction; make-up check valves from tank prevent cavitation on the suction side while the motor coasts down. Braking torque follows the relief setting — the deceleration ramp is adjustable by bar.
Size make-up flow for the full coast volume. Set braking relief above working peaks, below motor case and shaft limits. For frequent braking watch heat: each stop converts the full kinetic energy.
Both cylinder ports to tank: the implement follows the ground by itself.
Both cylinder ports to tank: the implement follows the ground by itself.
A valve float position connects A and B to T while P stays blocked (or unloaded). The cylinder moves freely with external forces — this is how a plow or a road grader blade hugs the terrain.
In float, the valve center is an H- or Y-like configuration. Selecting float releases the hydraulics completely: the cylinder becomes a gas-free, oil-damped sliding link. Anti-cavitation checks often accompany the valve to prevent vacuum when the load drives the piston fast.
Never float over people-paths without mechanical support. In simulation, float is perfect for testing that a vertical load without counterbalance drops — a training moment.
The pump listens to the load: it delivers what is needed, at Δp constant.
The pump listens to the load: it delivers what is needed, at Δp constant.
A load-sensing pump compares its outlet pressure with the highest actuator load pressure reported through an LS shuttle network. Its controller holds a constant margin (stand-by Δp ≈ 15-25 bar), so flow follows valve opening, not pump ego.
Each valve includes a pressure compensator that keeps the local Δp across its metering notch equal; therefore each function receives the flow proportional to spool travel, independent of load. The LS line is the nervous system returning the highest pressure to the pump regulator.
Keep the LS line short, small-bore and damped; the compensator spring defines minimum Δp — below it, control dies. LS saves 30-50 % energy versus constant-pressure systems in mixed duty.
When oil runs out, LUDV shrinks all speeds proportionally instead of starving the light axis.
When oil runs out, LUDV shrinks all speeds proportionally instead of starving the light axis.
Conventional LS fails at saturation: the lightest function wins and the heavy function stops. LUDV compensators sense AFTER the metering notch against the highest load pressure, so a too-small pump simply scales every function down by the same percentage.
Because all compensators reference the same downstream pressure, the pressure ratios across notches stay identical — flows divide by notch area, not by load. The operator experiences graceful slow-down instead of sudden axis death.
Standard on modern multi-function excavators. Diagnosis differs from classic LS: check compensator spool stiction first when functions move lazily under saturation.
Ramp the spool, not the mechanics: 300 ms of electronics replaces liters of accumulator.
Ramp the spool, not the mechanics: 300 ms of electronics replaces liters of accumulator.
Proportional valves ramp spool travel electronically. Acceleration and deceleration times (separately for on and off edges) convert step-changing flow into smooth S-curves, killing water hammer and load sway.
The amplifier card multiplies the command by a time-limited slew rate. Load sway frequency (≈ 0.5-3 Hz for cranes) and cylinder natural frequency define the minimum ramp: faster than the system can follow only excites oscillation.
Ramp long enough that peak pressure stays below hose impulse rating; short enough that takt time survives. Separate ramps per direction help asymmetric cylinders.
Closed-loop positioning with oil: micron accuracy if you respect physics.
Closed-loop positioning with oil: micron accuracy if you respect physics.
A servo or high-response proportional valve, position feedback transducer and controller close a loop around the cylinder. Control quality is governed by the hydraulic natural frequency ωh — loop gain must stay a comfortable factor below it.
Stiffness comes from oil compressibility over trapped volume: short hoses, thick walls, no entrained air. Damping is naturally low (0.05-0.2), so controller filters and valve overlap design carry the stability burden.
Keep total trapped volume minimal (valve on the cylinder). Use accumulators? Never on axis lines. Expect valve sizing at 1/3 supply Δp across the metering edge at design point.
A small pump-filter circuit that polishes the tank 24/7, independent of the machine.
A small pump-filter circuit that polishes the tank 24/7, independent of the machine.
Off-line filtration circulates 5-10 % of tank volume per minute through a fine filter (e.g. β6 = 200) driven by its own small motor-pump. It runs while the main machine works, idles, or sleeps.
Because flow is constant and low, the filter works in its best efficiency window far from surges. Cooling can ride in the same loop (filter-cooler skid), cleaning while conditioning temperature.
Size for the contamination generation rate of the system, not for tank size alone. Take suction from the dirty zone (opposite the return diffuser), return via clean side.
Every wasted kW must leave as heat — pick the radiator before the summer does.
Every wasted kW must leave as heat — pick the radiator before the summer does.
Losses (relief flow, throttling, leakage) heat the oil; the cooler must dissipate the average loss power at the maximum ambient temperature while keeping oil below 60-65 °C.
Air-blast coolers sit in the return line with a bypass check (cold morning start protection!). Water-oil plate coolers give compact performance where water exists. A thermostat bypass protects viscosity during warm-up.
Place AFTER the filter? No: place the filter after the cooler (cooler dirties oil at high Δp). Return-line coolers see pressure spikes — bypass accordingly or use off-line cooling.
New pipes are dirty pipes: flush until the oil is cleaner than the machine needs.
New pipes are dirty pipes: flush until the oil is cleaner than the machine needs.
Before commissioning, oil is circulated at high velocity through the new or repaired circuit with components bypassed (hoses jumpered), until particle counts meet class -2 of the target cleanliness.
High Reynolds numbers (turbulence) scrub walls: velocity 2-4 × operating speed, pulsating if possible. Dead legs and cylinders are flushed with temporary loops. Sampling valves verify progress.
Never flush through servo valves — jumpers! Document patch-sample or online counts per line. Budget one full tank volume through the flush filter before judging.
Spongy actuators and banging pumps: air — the treatable disease.
Spongy actuators and banging pumps: air — the treatable disease.
Trapped air compresses, making cylinders spongy, noisy and position-drunk. Bleeding moves oil through every high point until bubble-free oil exits.
Procedure: low pressure first, stroke cylinders end to end 5-10 times with bleed screws open at the highest points, then the pump suction side, then the valve bodies with vent options. Never crack fittings under pressure: loosen at low pressure only.
Cylinders mounted rod-down trap air at the cap — their bleed screws matter most. Foam in the sight glass = suction leak; find it with the smoke or shaving-cream test.
Start-up is a procedure, not a key switch.
Start-up is a procedure, not a key switch.
Sequence: mechanical completion check, flushing class verified, oil and level correct, suction valve OPEN (the classic killer), case-drain lines first to tank, pump primed, relief backed out, inch the motor rotation, raise pressure in steps.
Rotation check with 1 s bumping; then pump pressure at 20 bar; listen for cavitation; then functions one by one at reduced speed; finally full pressure with relief calibration against a reference gauge of known class.
Document every setting (relief, compensators, switches) with values and technician initials — these baselines are gold for later troubleshooting.
Plan measurement BEFORE doubt: one coupling per decade of pressure.
Plan measurement BEFORE doubt: one coupling per decade of pressure.
Minimess-style test couplings at pump outlet, after the main valve, at both cylinder ports and on pilot lines turn fault-finding from days into minutes. They connect under full pressure with the appropriate fitting.
Position matters more than count: upstream and downstream of each suspected throttling element lets Δp be measured, not guessed. Permanent sensors feed the PLC trend; couplings serve the technician.
Cap them (dust kills). Route hoses away from moving parts. A single gauge with quick coupler covers the whole machine.
The pump feeds by atmosphere alone: give it a VIP corridor.
The pump feeds by atmosphere alone: give it a VIP corridor.
Atmospheric pressure pushes oil into the pump; any restriction consumes the tiny available head. Rules: large diameter (velocity 0.5-1.5 m/s), short, no sharp bends, no risk of air traps, strainer optional but generously sized.
Cavitation begins when local pressure falls below the vapor pressure of dissolved air — bubbles collapse at the pressure side with micro-jets that erode gears and vanes, sounding like gravel in the pump.
Cold starts are the test: viscosity 1000 cSt multiplies losses. Shutoff valves in suction lines need interlocks or position switches (forgetting = pump death in minutes).
One machine, three philosophies — the great filter debate settled by failure modes.
One machine, three philosophies — the great filter debate settled by failure modes.
Pressure filters protect expensive components directly (servo valves), return filters catch everything before the tank, off-line loops polish continuously. Most industrial machines combine return + component-protection pressure filters.
Pressure filters need high pressure rating and see full flow pulses; return filters see surge flows from differential cylinders (up to 2× pump flow!) and need bypass valves. Off-line filters never protect against sudden ingress.
Never put the only fine filter in suction: the pump dies of starvation before the filter clogs. Indicator (visual + electrical) on every filter, wired to maintenance-not-required-but-logged.
10-12 bar per degree Celsius: the sun pressurizes your parked machine.
10-12 bar per degree Celsius: the sun pressurizes your parked machine.
Oil trapped between two closed valves expands with temperature. With bulk modulus 14000 bar and thermal expansion 0.0007/K, every +1 K adds roughly 10 bar to a rigid locked volume.
Overnight cooling or noon sun can push a clamp circuit from 120 to 250 bar — or drop it to vacuum. Thermal relief valves set just above working pressure protect hoses and fixtures.
Any circuit that locks oil under load and sees weather or process heat needs a thermal valve or a relieving branch. Accumulator circuits already carry them if correctly designed.
600 bar clamping from a 200 bar machine — no second power pack.
600 bar clamping from a 200 bar machine — no second power pack.
Media separators and oscillating automatic intensifiers convert low-pressure flow into high-pressure micro-flow using area ratios. Clamping fixtures reach pressures the HPU never sees.
A large piston driven by system pressure pushes a small piston into the fixture volume; strokes reverse automatically through an internal pilot valve, pumping until the intensified pressure equals ratio × supply pressure.
High side must be dead-headed and leak-free: any leak drains the amplifier continuously and heats oil. Use only mineral-oil-rated seals on the HP side unless specially ordered.
Same flow through each shaft: equal speed, torque adds up.
Same flow through each shaft: equal speed, torque adds up.
Series-connected motors pass the same flow through each unit; each converts its share of the total pressure drop into torque. Speeds equalize naturally (like rollers on one belt of oil).
Because shaft seals on motor 1 see the outlet pressure feeding motor 2, case drains are mandatory and the first seal must survive intermediate pressure. Total available torque divides across stages.
Use for synchronized drives with modest shock tolerances; unequal loads do not desynchronize (flow forces the speed) but do shift the torque split. Case drain pressure limits chain length.
Same pressure, flow finds the easy path: speed wanders with load.
Same pressure, flow finds the easy path: speed wanders with load.
Parallel motors see the same pressure; each takes the flow its load allows. The lighter-loaded unit accelerates and steals flow — inherent differential behavior, sometimes useful, often annoying.
Useful when mechanical coupling exists anyway (both wheels of an axle): the differential action is welcome in curves. Without coupling, speeds drift until loads re-balance.
For forced equal speeds go to flow dividers or series connection. Check case drain and cross-line relief: a stalled motor in a parallel group is a relief-heating machine.
A small valve first, the big valve later: nobody hears the release.
A small valve first, the big valve later: nobody hears the release.
Before opening the main return path of a large charged cylinder, a small parallel path (needle valve, proportional throttle or special decompression spool) releases the stored volume slowly to tank.
Stored energy = p²V/(2β) in oil plus frame elasticity. Releasing through a big orifice in 50 ms converts it to a pressure wave; a 10 mm path over 2 s converts it to silence. Timers or pressure switches trigger the big spool only below a threshold (e.g. 10 bar).
Size decompress time from volume and pressure: rule of thumb 0.3-0.5 s per liter at 200 bar. Interlock the main valve to pressure feedback, not to time alone.
The pilot line is a remote control: switch between two pressure levels or go to unload.
The pilot line is a remote control: switch between two pressure levels or go to unload.
The pilot spring chamber of a pilot-operated relief valve can be piped away to small remote valves: a second relief setting for reduced pressure, or a solenoid valve to dump the chamber to tank for full unload.
Venting drops main pressure to the crack pressure of the main stage (≈ 5-10 bar): pump flow goes to tank at idle pressure. The lower of the two pilot settings always governs; switching is fast and electrically cheap.
Keep remote lines short (< 3 m) and small: long lines oscillate. Use damped poppets where switching shock hurts. Fail-safe: vent on power loss = unload.
The card is the machine behavior: dither, ramps, gains — tuned by screwdriver or software.
The card is the machine behavior: dither, ramps, gains — tuned by screwdriver or software.
Proportional amplifier cards translate command voltage into solenoid current (PWM), offering ramp times per direction, min/max (Imin/Imax) trims, and dither frequency/amplitude to fight spool hysteresis.
Dither (e.g. 100-300 Hz, 5-20 % of rated current) keeps the spool vibrating microscopically so static friction never wins. Ramps limit command slew rate; Imax scales maximum speed; Imin compensates deadband for smooth creep.
Tune in order: Imin with command at 1 %, then Imax at 100 %, then ramps against the mechanical reality, finally dither for hysteresis. Document trimmer positions — they wander.
Every permission condition gets a contact; all contacts vote in series.
Every permission condition gets a contact; all contacts vote in series.
Filter clogging, low level, over-temperature, over-pressure: each monitored by a switch whose contacts form a series chain that cuts the motor contactor or dumps pressure when any link opens.
NC (closed when healthy) wiring is the safe philosophy: a broken wire looks like a fault, not a permission. Monitored (positively-driven) contacts and two-channel redundancy raise the performance level.
Debounced PLC inputs or time relays prevent chatter from commissioning every pump start. Test each switch individually during maintenance — never assume the chain works.
Follow the kilowatts: relief, throttling, leakage — the three heating plants.
Follow the kilowatts: relief, throttling, leakage — the three heating plants.
Every pressure drop times flow becomes heat: Ploss = Δp × Q / 600. Walk the cycle with gauges and flow readings, compute kW per station, and attack the biggest loss first.
Classic findings: idle flow over the relief (open-center missing), meter-in during high-force dwell (throttle at 100+ bar), undersized lines (> 5 bar piping losses), leaking pump (ΔT across pump body).
A 3 kW saving at 6000 h/a and 0.15 currency/kWh pays most retrofits in a year. Bring numbers, not opinions — the temperature rise alone quantifies steady-state loss.
From force and speed data to a numbered schematic — the complete professional workflow.
From force and speed data to a numbered schematic — the complete professional workflow.
1) List actuators with forces, speeds, directions and duty. 2) Choose system pressure class. 3) Size cylinders/motors (with x1.5 force margin). 4) Compute flows and peak/average power. 5) Select pump type and drive power. 6) Choose control philosophy (valve types, throttle vs displacement control). 7) Size valves, lines, cooler, filter, tank. 8) Add safety: relief, load-holding, dump, interlocks. 9) Instrument and document (gauges, test points, settings list). 10) Simulate and verify edge cases (load loss, hose burst, blocked rod).
The order matters: pressure class before cylinder size, flow before pump, cycle profile before cooler. Skipping steps produces the classic over-heated, over-powered machine — every shortcut returns as degrees Celsius.
Simulate before building: an hour in the Studio finds sequence errors, intensification pockets and heat problems that would cost weeks on the floor.
Pressure times flow, divided by 600 — the one formula behind every kW decision.
Pressure times flow, divided by 600 — the one formula behind every kW decision.
Hydraulic power is flow against pressure. In field units: P (kW) = p (bar) × Q (l/min) / 600. Every design budget — motor size, cooler size, heat generation — converts through this single identity.
Double the pressure at equal force? Halve the cylinder and the flow halves; pump kW stays the same but the machine shrinks. This is why pressure classes exist: 70, 160, 250, 350 bar standards.
Always factor efficiency: a real power pack converts ~85 % of shaft power into useful hydraulic power; the rest is already heat at the source.
Leaking oil and rubbing metal: the two thieves described by two numbers.
Leaking oil and rubbing metal: the two thieves described by two numbers.
Volumetric efficiency ηvol measures internal leakage (delivered vs theoretical flow); hydromechanical efficiency ηhm measures friction losses. Overall η = ηvol × ηhm.
A worn pump typically loses ηvol first (more leakage) — but watch temperature: hot thin oil fakes wear by increasing leakage 2-3 %. Friction losses dominate at high pressure and low speed (stick-slip zone).
Compare like-for-like: state pressure, speed and viscosity with every efficiency figure. A datasheet η at 1450 rpm says little about your 600 rpm duty.
Oil is a spring: mass on that spring rings — and limits every control loop.
Oil is a spring: mass on that spring rings — and limits every control loop.
The compressible oil column between valve and piston plus the moved mass forms a spring-mass oscillator. Its natural frequency ωh caps the achievable speed of position and force control loops.
Short lines, rigid pipes, no trapped air, valve on the cylinder: every measure raises ωh. Mid-stroke with equal volumes is usually the worst case. Damping ratios of 0.05-0.15 are typical — the system loves to ring.
Check ωh before promising positioning accuracy. Below 5 Hz, expect visible settling oscillation on heavy tables.
Q follows root of Δp: the quadratic heart of every throttle.
Q follows root of Δp: the quadratic heart of every throttle.
Flow through a sharp-edged restriction follows Q = cd · A · √(2Δp/ρ). Doubling speed through the same orifice quadruples the pressure loss — the source of many surprised faces.
Blind (long) orifices behave more linearly (laminar) and are temperature-sensitive; sharp-edged ones are viscosity-robust but quadratic. Datasheet Δp-Q curves tell you which physics you bought.
Reading a valve datasheet: nominal flow defined at some Δp (often 5 or 10 bar per edge) — running it at higher Δp flows actually more than nominal.
Three numbers that decide whether servo valves live or die.
Three numbers that decide whether servo valves live or die.
ISO 4406 codes report particle counts per ml at > 4, > 6 and > 14 µm(c), coded on a logarithmic scale (each step = doubling). Example 18/16/13: typical clean industrial oil.
Servo/proportional systems want 16/14/11 or better; standard gear-pump systems are content with 19/17/14. One code step better roughly doubles component life in contaminated-sensitive hardware.
Sample from a turbulent return line, mid-stream, not from the tank bottom. Trend counts — a jump of two codes overnight means ingress, not dirtiness.
β<sub>x</sub> = 200 means 99.5 % capture — learn the one number that matters.
βx = 200 means 99.5 % capture — learn the one number that matters.
The β-ratio counts particles upstream vs downstream at a given size: β10µm(c) = 200 means only 1 in 200 of 10-µm particles passes (99.5 % efficiency).
Old "nominal 10 µm" claims mean almost nothing; absolute ratings with β ≥ 200 under ISO 16889 multipass testing are the professional language. Watch the (c): calibration changed particle sizing standards.
Choose by component sensitivity: servo valves 3-6 µm, proportional 6-10, standard 10-25. Finer filter = more frequent changes; balance with off-line polishing loops.
Bubbles born in low pressure, murdered in high pressure — and the damage they leave.
Bubbles born in low pressure, murdered in high pressure — and the damage they leave.
When local pressure falls below the oil vapor/air-release pressure, cavities form; transported into pressure zones they implode with micro-jets exceeding 1000 bar locally, eroding metal grain by grain.
Pump-side cavitation sounds like gravel; valve-side (sharp metering edges at high Δp) shows as erosion tracks downstream. Dissolved air worsens everything: aerated oil cavitates earlier.
Defenses: short fat suction lines, flooded inlet if possible, oil temperature in the 40-50 °C window, and keeping metering-edge Δp sane on gaseous duty.
16-36 cSt optimum, 1000 cSt start limit, 10 cSt danger — viscosity rules everything.
16-36 cSt optimum, 1000 cSt start limit, 10 cSt danger — viscosity rules everything.
Viscosity decides film thickness (lubrication) and leakage (efficiency). ISO VG grades name cSt at 40 °C; the viscosity index VI describes how flat the curve stays across temperature.
Cold morning oil at 800 cSt cavitates pumps; hot afternoon oil at 12 cSt weeps past every clearance and lubricates poorly. High-VI oils (HVLP) flatten the swing for outdoor machines.
Choose VG class by your working temperature window, not by habit: many plants run VG 46 where VG 32 would cut energy losses measurably.
Mineral, synthetic, water-based, bio: each fluid rewrites seal and fire rules.
Mineral, synthetic, water-based, bio: each fluid rewrites seal and fire rules.
HLP (mineral with additives) is default. HVLP adds viscosity-index improvers. HFC (water-glycol) and HFD (synthetic ester) serve fire-risk zones; HEES/HETG are rapidly biodegradable for environmental sites.
Fire-resistant fluids penalize lubrication (HFC: derate pumps 20-30 %) or attack standard seals and paints (HFD: FKM seals, epoxy paint). Bio fluids oxidize faster — temperature discipline and shorter change intervals.
Never mix fluid families: residues form gels. Conversion needs flushing procedures approved by both fluid and component makers.
Seals work by being squeezed into the gap — pressure feeds the seal or eats it.
Seals work by being squeezed into the gap — pressure feeds the seal or eats it.
Elastomer seals energize under pressure: lip seals open toward the pressure, O-rings deform into the extrusion gap. Above the gap tolerance for a given pressure, rubber extrudes and the seal is nibbled away.
Back-up rings close the extrusion gap for high pressure; PTFE step seals reduce friction and stick-slip on servo cylinders. Direction matters: a rod seal installed backwards becomes a pump.
Check surface finish (Ra 0.1-0.4 µm rods), hardness and gap against the pressure class in the seal catalog. Temperatures kill: NBR past 100 °C, FKM for heat, PU for abrasion.
Isothermal for storage, adiabatic for shocks: the exponent decides the volume.
Isothermal for storage, adiabatic for shocks: the exponent decides the volume.
Accumulator gas follows p · Vn = const. Slow charge/discharge (minutes): n ≈ 1, isothermal. Fast events (seconds): n ≈ 1.4 for nitrogen, adiabatic. Usable volume differs by 30 % between the two.
Precharge p0 = 0.85-0.9 × minimum working pressure keeps some oil reserve; above that the bladder hits the poppet and flow dies abruptly. Too low precharge slams the bladder against the shell — early failure.
Verify precharge annually with the machine depressurized (charging kit + N2 bottle, NEVER oxygen). Temperature swings change p0 — recheck seasonally on outdoor machines.
7000-14000 bar bulk modulus: stiff, but a spring nonetheless.
7000-14000 bar bulk modulus: stiff, but a spring nonetheless.
Mineral oil compresses roughly 0.7 % per 100 bar (bulk modulus β ≈ 14000 bar); with 1 % entrained air the effective value collapses to a third or less. Every trapped volume is a spring — useful for shock absorption, troublesome for control.
Effective stiffness of a cylinder spring: C = β · A² / V. Big volumes (long hoses! accumulator-side lines) soften the axis and lower the natural frequency — the main reason servos hate hoses.
Bleed air religiously for stiff axes; hard-pipe servo lines; keep valve-to-actuator volume minimal.
Push force is limited by a noodle: long thin rods bend before they reach nominal force.
Push force is limited by a noodle: long thin rods bend before they reach nominal force.
A pushing cylinder is a column. Buckling load falls with the square of free length; mounting (-guided vs free end) changes the effective length factor from 0.5 to 2.
Manufacturer tables give permissible push force vs stroke per mounting style. Stop tubes increase the effective rod diameter length inside the cylinder to reduce bearing span issues, extending allowable loads on long strokes.
Check buckling whenever strokes exceed ~10 × rod diameter in push duty. Horizontal mounting with sagging rods worsens it — mid-supports help.
The tank is a process device: settling, de-aeration, cooling — not a bucket.
The tank is a process device: settling, de-aeration, cooling — not a bucket.
Rules of thumb: stationary tank volume 2-4 × pump flow per minute; mobile 1-1.5 ×. Return oil enters diffused below level opposite the suction, separated by a baffle that forces a calm settling path.
Residence time releases air and drops particles and water. Suction zone sits lowest and calmest; return diffuser slows jets that would stir sediment. Breather sizing must cover maximum level-change air flow without pressurizing the tank.
Level switch + temperature sensor + breather filter are the minimum instrumentation. Sloped bottom with drain: water collects first, drain it weekly on humid sites.
Bend radius, torsion, length margin: the STAAMP and routing discipline.
Bend radius, torsion, length margin: the STAAMP and routing discipline.
Hoses fail from abuse more than age: bend below minimum radius, twist under pressure, rub against steel, stretch under shrinkage (-4 % at pressure!). STAAMP: Size, Temperature, Application, Media, Pressure.
Pressure shrinks and stiffens hoses: leave 4-7 % slack and machine-motion margin. Route so flexing happens in one plane only; clamp where rubbing is unavoidable; never mix bend with torsion (the deadliest combination).
Impulse rating matters: 4-spiral for shock duty, braided for gentle static. Date-code assemblies; replace by condition with a documented interval.
Steady-state temperature = losses vs dissipation; compute it before summer.
Steady-state temperature = losses vs dissipation; compute it before summer.
All hydraulic losses become heat. Steady state balances loss power against dissipation through tank walls (≈ 0.015 kW/m²K × surface × ΔT), components, lines and the cooler.
Estimate losses: standby unload 2-3 % of drive power continuously; throttling Δp × Q / 600; leakage flows at pressure; then check against cooler capacity at worst ambient (solar load for mobile!).
Oil ages per Arrhenius: every +10 K above 60 °C halves oxidation life. Alarm at 70, trip at 80 °C as common practice.
Gear, vane, piston, screw: duty, pressure, noise and budget in one comparison.
Gear, vane, piston, screw: duty, pressure, noise and budget in one comparison.
Gear pumps: cheap, robust, fixed, to 250-280 bar. Vane: quiet, medium pressure (≤ 210 bar), variable versions exist. Axial piston: 350-450 bar, variable everything, expensive, dirt-sensitive. Screw: silent, for lubrication duty and low pulsation.
Selection flow: required pressure class excludes types; variable vs fixed by duty cycle (intermittent high force = variable earns money); noise limits (labs, theaters) point to vane/screw; dirty environments forgive gear pumps most.
Do not oversize displacement: a 100 cm³ pump at 30 % duty runs out of its efficient window and wears unevenly. Match displacement to average demand with margin.
Scoring, pitting, spalling, cavitation mapping — autopsy of pumps and motors.
Scoring, pitting, spalling, cavitation mapping — autopsy of pumps and motors.
Failed components talk: circumferential scratches = abrasive particles; pitting clusters = cavitation; smeared metal = lubrication breakdown; fatigue spalling = overload cycles; polymer deposits = varnish from hot oxidized oil.
Cut open (or inspect ports of) the dead pump before ordering its replacement: the failure signature tells whether the new one will die the same way. Filter element inspection (pleat contents under bright light) completes the picture.
Photograph and log every autopsy; correlate with oil analysis trends. Warranty discussions become short when the evidence is a picture.
Weekly eyes, monthly senses, quarterly lab: oil tells the truth on a schedule.
Weekly eyes, monthly senses, quarterly lab: oil tells the truth on a schedule.
Weekly: level, leaks, temperature, noise. Monthly: breather, cooler fins, hose condition, accumulator precharge plausibility. Quarterly/500 h: oil sample (particles, water, viscosity, TAN), filter indicator review. Annual: precharge check, relief re-calibration, pump flow test.
Oil analysis trending beats reactive change: viscosity drift, water ppm, ISO code and TAN curves predict problems months out. Change by condition, not calendar — economics and uptime both win.
One responsible person owns the fluid; labels on filters with change dates; spare filter elements on site (a clogged bypass weekend is avoidable).
Which numbers belong in the FAT/SAT protocol — and their classes of gauge.
Which numbers belong in the FAT/SAT protocol — and their classes of gauge.
Record: motor current vs relief pressure curve, pump delivery at 3 pressures, cooler in/out ΔT over 30 min at worst-duty cycle, all valve settings against a class-1 reference gauge, response times of pressure switches.
Instrument classes matter: shop gauges (±2.5 %) cannot verify a relief set within ±5 bar at 250 bar — use a class 0.6 test gauge or digital sensor for witnessed points.
Freeze the protocol template: machine ID, oil type, temperature window, gauge IDs, signatures. Deviations get numbers, not adjectives.
4:1 for hoses, 4:1 for tubes by rule — know the numbers before signing.
4:1 for hoses, 4:1 for tubes by rule — know the numbers before signing.
Components carry rated (working) pressure and minimum burst pressure. Standards demand ratios: hoses and rigid lines commonly 4:1 against maximum working pressure including peaks; valves and cylinders per their ISO standards.
Pressure PEAKS count, not averages: recorded spikes of relief overshoot belong in the selection math. A 100-bar-rms system with 350 bar hammer peaks is a 350 bar selection exercise.
Never uprate a system by tightening the relief without rechecking every downstream rating. Document peak measurements with fast recorders — slow gauges lie.
A pinhole leak injects oil like a syringe — the injury you must fear most.
A pinhole leak injects oil like a syringe — the injury you must fear most.
Oil jets from pinhole leaks penetrate skin at glove-level pressures and cause tissue necrosis hours later — surgical emergency, not a plaster job. Never search leaks with hands; use cardboard.
Stored energy: accumulators and suspended loads keep machines dangerous with the motor off. LOTO procedure: isolate, dump accumulators via the manual valve, verify zero energy on gauges, mechanically support loads.
Personal protective: glasses always, gloves for oil work, faceshield for suspected leaks. Mark stored-energy zones on the machine with the standardized label.
When the expensive pump is the cheap machine.
When the expensive pump is the cheap machine.
Fixed pumps cost little but burn the same kW regardless of demand (heat at the relief). Variable pumps cost 3-5 × more but deliver on demand. The decision variable is the duty profile: hours per year at partial demand.
Payback math: saved kW = idle-ish pressure × surplus flow averaged over the year. Intermittent cycles (presses, clamps) repay variable displacement in 1-3 years; continuous full-load machines rarely.
Variable pumps add control complexity (compensators, load-sense) — budget commissioning skill and dirt sensitivity into the decision.
Before the hose bursts: enumerate every trapped volume and compute its worst pressure.
Before the hose bursts: enumerate every trapped volume and compute its worst pressure.
For every line section that can be closed at both ends, list the amplifiers connected: differential cylinder ratio, thermal expansion, external loads. Compute worst-case pressure per pocket and rate the segment accordingly.
Worksheet method: pocket, closing scenario, area ratio or ΔT, computed peak, relief provision, component ratings. Auditors increasingly ask for exactly this table on existing machines.
Do not forget pilot lines behind check valves — the classic unnoticed pocket. Solve with thermal valves small and near the pocket.
Read any schematic like a sentence: energy, control, actuation — in that order.
Read any schematic like a sentence: energy, control, actuation — in that order.
Method: find the energy source (circle with triangle), trace P to the consumers; identify each valve envelope count (positions) and port count (ways); identify actuation arrows (solenoid coil, pilot dash, spring); finish at the actuators and back through T.
Boxes = discrete positions; one box always drawn in rest state; proportional valves get parallel lines above/below; pilot lines dashed, drain dotted, enclosure boxes grouped by physical units. Flow direction arrows filled (hydraulic) or hollow (pneumatic).
Practice with five real schematics from different industries — symbol dialects exist (detailedd vs simplified pumps), method stays.
Clamp, drill, retract, eject — one circuit, every lesson applied.
Clamp, drill, retract, eject — one circuit, every lesson applied.
Specification: clamp at 60 bar, drill feed 12 mm/s at up to 140 bar, fast approach 60 mm/s, eject on return. You design: cylinders φ and margins, pump and motor, valve selection (sequence + reducing + meter-out), relief at 160, cooling check, and the schematic.
Expected moves: sequence valve for clamp-before-feed; reducing valve for clamp force limit; regeneration if takt demands; meter-out on feed for finish quality; test points at four stations; settings list with witnessed values.
Deliverables as a professional dossier: schematic, sizing calculations, settings list, bill of materials with ratings, and the sim-derived temperature estimate.
The workshop classic: cheap, tough, intermittent service to 8-12 bar.
The workshop classic: cheap, tough, intermittent service to 8-12 bar.
A crank-driven piston compresses air in one or two stages into a receiver. Single-stage reaches ≈ 8-10 bar, two-stage with intercooling 12-15 bar. Duty cycle is limited (typically 50-60 %) — it is an intermittent machine.
Intake stroke draws filtered air through the suction valve; compression stroke pushes it through the pressure valve into the aftercooler/receiver. The pressure switch cycles the motor between upper and lower setpoints.
Size by free air delivery (FAD) in l/min or Nl/min at working pressure, not by motor kW. Oil-lubricated units last longer; oil-free are mandatory for food/pharma air.
Continuous flow for plants: quiet, efficient, 100 % duty.
Continuous flow for plants: quiet, efficient, 100 % duty.
Two meshing rotors (an oil-flooded male/female pair) compress air progressively along their axis. No inlet/outlet valves, no pulsation: ideal for continuous baseload air.
Capacity control via inlet throttle, blow-off, or speed control (VSD). Oil is injected for sealing and cooling, then separated to < 3 ppm; dryers and filters downstream polish to the required class.
Choose VSD when the demand profile varies widely — fixed-speed idling wastes 20-30 % of rated power blowing off. Heat recovery (70-90 % of input power available as warm water) often pays for itself.
Buffer, cooler, water separator and pulsation damper in one steel bottle.
Buffer, cooler, water separator and pulsation damper in one steel bottle.
The receiver stores compressed air to bridge demand peaks, lets the compressor run longer efficient cycles, cools the air (condensing water out first, where it belongs) and damps pulsation from piston compressors.
Rule of thumb: volume (l) ≈ compressor delivery (l/min) for piston units; smaller for screw compressors. Peaks size it: V = Δp window math over the peak duration.
Drain daily (manual) or fit an automatic drain; corroding receivers are bombs — periodic inspection per local pressure-vessel rules with wall-thickness testing.
First defense: spin the air, throw the water at the wall, drain it away.
First defense: spin the air, throw the water at the wall, drain it away.
Wet compressed air enters tangentially; centrifugal force hurls droplets to the wall where they run down into the quiet zone of the bowl, removed by a float drain.
Removes bulk liquid (not vapor!) — typically 99 % of droplets > 10 µm. Sits directly after the aftercooler and before filters and dryers.
Works poorly at low velocity: oversized separators lose spin. Keep the drain functional — a blocked drain refloods the line.
Borosilicate fibers catch aerosols; 0.01 µm and 0.003 mg/m³ oil claims, verified by ISO 8573.
Borosilicate fibers catch aerosols; 0.01 µm and 0.003 mg/m³ oil claims, verified by ISO 8573.
Micro-glass fiber beds force sub-micron oil and water aerosols to collide, coalesce into drops and drain off the element. Grades: 1 µm / 0.1 mg/m³ (pre), 0.01 µm / 0.01 mg/m³ (micro), plus activated-carbon for vapor.
Depth filtration with Brownian motion capture: efficiency rises as fibers wet. Δp grows with loading — change elements when Δp passes 0.5 bar (energy cost of a dirty filter exceeds its price within weeks).
Always in series coarse to fine after the dryer. A carbon stage does not replace the micro-filter — vapor removal needs its own cartridge.
Cool the air to +3 °C dew point: the plant-standard dryer for indoor networks.
Cool the air to +3 °C dew point: the plant-standard dryer for indoor networks.
A refrigeration circuit chills compressed air to a pressure dew point of +3 °C; vapor condenses and is removed by the integrated separator/drain.
Remember: dew point applies AT LINE PRESSURE. +3 °C PDP prevents condensation in any indoor pipe that stays warmer than 3 °C — but outdoor winter lines still condense.
Size for worst summer: inlet temperature, pressure and ambient all reduce capacity. An inlet coalescing filter protects the heat exchanger from oil fouling.
-40 °C or -70 °C dew point for outdoor lines, instruments and processes.
-40 °C or -70 °C dew point for outdoor lines, instruments and processes.
Twin towers filled with desiccant (alumina/molecular sieve) alternate between drying and regeneration. Regeneration wastes 15-20 % purge air (heatless type) or uses heater/blower energy.
Water vapor adheres physically to the desiccant surface; when the tower saturates, valves swap the flow and a dry purge stream strips the moisture to atmosphere.
Oil destroys desiccant permanently — pre-filter to 0.01 mg/m³ always. Undersized purge = creeping dew point; monitor with an online dew-point meter.
Hollow fibers breathe water vapor out: compact, silent, no power — for small flows.
Hollow fibers breathe water vapor out: compact, silent, no power — for small flows.
Water vapor permeates selectively through hollow fiber membranes to the outside, where a dry purge stream (~15-20 % of flow) carries it away. Point-of-use drying to -20 °C PDP for small consumers.
Flow capacity is small (< ~1000 Nl/min) and purge loss is permanent — a continuous consumption in your compressed-air balance.
Requires an upstream 0.01 µm filter (oil coats the membrane irreversibly). Place directly at the consumer.
The classic triplet at every machine inlet: clean it, set it, (rarely) oil it.
The classic triplet at every machine inlet: clean it, set it, (rarely) oil it.
The air preparation unit conditions plant air for the machine: filter (5-40 µm with water drain), pressure regulator (downstream pressure kept constant), optional lubricator (oil fog for old-style cylinders and tools).
Order is fixed: filter first (protect everything), regulator second (sets working pressure), lubricator last (oil would be filtered out upstream). Modern practice: skip the lubricator and use lifetime-lubricated cylinders; if used, never starve it and never fertilize tools with stainless internals.
Bowl: polycarbonate needs guards in solvent atmospheres (metal bowl alternative). Set regulator under flow: static settings read 0.5-1 bar high (droop).
Filter + regulator in one body: the modern default service module.
Filter + regulator in one body: the modern default service module.
A compact combination: 5 µm filter with auto or semi-auto drain plus relieving regulator with gauge port. Saves space and a joint — every joint is a future leak.
Relieving regulators vent downstream overpressure to atmosphere — a safety property: blockage does not trap pressure behind the consumer.
Pick drain type by site: auto-drain where bowls fill unseen; semi-auto for maintenance-friendly boards. Gauge placement readable from the operator stance.
Repeatability to millibars: for tensioning, balancing and lab air.
Repeatability to millibars: for tensioning, balancing and lab air.
High-gain pilot-operated regulators hold set pressure within ±0.5 % with minimal droop, using a balanced poppet and sensitive diaphragm stacks.
Standard regulators droop 0.5-1 bar over the flow range; precision units hold their setpoint thanks to the amplified control loop and constant bleed design.
Air consumption note: constant-bleed types consume ~2-6 Nl/min continuously. Keep inlet air clean (5 µm) or the tiny nozzles foul.
Pressurize gently: cylinders glide home instead of slamming at power-up.
Pressurize gently: cylinders glide home instead of slamming at power-up.
At machine start the soft-start valve restricts filling until downstream pressure reaches ~50 % of supply; then it snaps fully open. On stop, the 3/2 quickly exhausts the machine.
Slow pressurization strokes all cylinders to their rest positions gently and predictably; the snap-through ensures full dynamics after the ramp. Adjustable needle sets fill time.
Combine with the emergency-stop dump function in safety circuits. Note: soft-start is NOT a safety component by itself — the exhaust function is.
Category 0 stop for pneumatics: remove the energy, monitored.
Category 0 stop for pneumatics: remove the energy, monitored.
A 3/2 valve (often redundant, monitored) on the machine inlet exhausts all downstream air when the safety chain opens. Monitored versions feed spool position back to the safety relay for fault detection (ISO 13849 channels).
Exhaust capacity must depressurize the hazardous volume within the risk-assessment time; check Cv of the exhaust path, not just the inlet path. Soft-start + dump in one block is the modern integrated solution.
Redundant + monitored for PL d/e; single valve acceptable only for lower categories. Verify exhaust time during commissioning — compute from volume and Cv if in doubt.
Legacy air tools love it; modern cylinders mostly hate needing it.
Legacy air tools love it; modern cylinders mostly hate needing it.
The lubricator meters oil drops into an air stream, creating a fog that coats downstream components. Essential for air motors, impact wrenches and older drill units.
Once used, always used: oil washes out factory grease, so removing the lubricator later starves components. Fog travels maybe 5-8 m and does not climb vertical lines — local lubricators for distant points.
Modern practice: lifetime-lubricated valves/cylinders, no lubricator, exhaust air free of oil mist (workplace hygiene!). For tools: precision micro-fog units at the drop line.
Water always wins unless someone drains it: design the low points.
Water always wins unless someone drains it: design the low points.
Condensate collects at every low point: receivers, dead legs, filter bowls, ring-main drops. Automatic float drains, timed solenoid drains or level-sensing zero-loss drains remove it without operator memory.
Timed drains blast expensive compressed air to atmosphere (2 s open every minute behaves like a continuous leak); level-controlled electronic drains open only when full — the small cost that pays back in compressor kWh.
Route condensate through an oil-water separator before the sewer: 95 % of condensate from lubricated compressors is legally waste oil mixture.
Particles, water, oil: three numbers name your air — e.g. class 2.4.2.
Particles, water, oil: three numbers name your air — e.g. class 2.4.2.
The standard grades solids (1-7 by particle concentration), humidity (1-7 by pressure dew point) and oil (1-5 by mg/m³). Spec sheets list machine requirements in exactly this code.
General machinery: 2.4.2 is typical. Paint/food: 1.2.1 or better. Instrument air outdoors: water class 3 (-20 °C PDP) at least. Class selection drives the entire treatment chain and its energy cost.
Specifying better than needed costs purge air and filters forever; specifying worse costs downtime. Write the class into the machine requirement document — suppliers cannot argue with standards.
Rings beat trees: balance pressure by giving air two roads to every machine.
Rings beat trees: balance pressure by giving air two roads to every machine.
A closed ring main equalizes pressure across the plant (vs tree layout where the last consumer starves). Drops to machines come from the TOP of the main (water runs down!), with a shutoff and a drain leg pointing down.
Pipe sizing by velocity: mains 6-9 m/s, drops 10-15 m/s acceptable. Slopes of 1-2 % toward drain points. Every dead leg collects condensate: eliminate or drain it.
Material choice: aluminum/stainless modular for clean air and easy change; galvanized steel economically; avoid plastic beyond point-of-use ratings. Pressure drop audit: ≤ 0.3 bar main, ≤ 0.6 bar total.
The basic switching element: at rest the consumer is vented, actuated it is pressurized.
The basic switching element: at rest the consumer is vented, actuated it is pressurized.
Two positions, three ports (1 supply, 2 output, 3 exhaust). At rest (NC) port 2 is connected to 3 — the cylinder is exhausted. Actuation connects 1 to 2 and blocks 3: the consumer receives air.
Spring return defines the rest state; actuation can be manual, mechanical, pneumatic pilot or solenoid. The exhaust port carries the silencer in control circuits.
Choose NC when the safe state = depressurized (default for single-acting cylinders). Flow rate (Nl/min) must support the cylinder size — small 3/2 valves starve big cylinders.
Rest = pressurized: air flows until you act — rare but decisive.
Rest = pressurized: air flows until you act — rare but decisive.
Same anatomy as NC but the rest position connects 1 to 2. Consumers stay pressurized until the valve is actuated.
Electrical analogy: a normally-open contact is a fail-closed wire — a broken spring or lost pilot leaves the consumer PRESSURIZED. That property is sometimes the hazard, sometimes the safety (clamps that stay clamped on air loss).
Use deliberately: breakaway detection, clamping circuits that must hold during air loss, purge lines default-on. Document the rest state on the schematic clearly.
The standard double-acting cylinder driver: one coil, one spring, two flow paths.
The standard double-acting cylinder driver: one coil, one spring, two flow paths.
Five ports (1 supply, 2/4 cylinder ports, 3/5 exhausts). At rest, one cylinder side is pressurized and the other exhausted; energizing the solenoid flips the paths. Removing the signal returns the spool via spring.
Power failure = defined return (spring side) — this is the key difference to the double-solenoid version and the basis of many safety philosophies.
Check power consumption and duty of the coil for long holding; for battery or heat-sensitive panels select low-power versions. Manual override (the small button) belongs in every commissioning ritual.
Two coils, no spring: the valve remembers the last pulse.
Two coils, no spring: the valve remembers the last pulse.
Detented spool: a short pulse on coil A switches and stays; a pulse on coil B switches back. Both coils energized simultaneously is an undefined fight (interlock in the PLC!).
Power loss = the spool stays: the machine resumes where it stopped (desired for clamping hold, undesired where rest position is safety). This memory property also makes purely pneumatic step chains possible.
PLC must interlock the two outputs and typically pulse (300-500 ms), not hold, to save coil heating. Commissioning: which coil extends? Mark cables — swapped coils are the classic start-up bug.
Three personalities of one valve family — the mid position is the message.
Three personalities of one valve family — the mid position is the message.
5/3 valves spring-center to a middle position. All-ports-closed stops the cylinder mid-stroke (approximately — air compresses, drift guaranteed!). Vented (center exhausted) relaxes both sides (free positioning, soft start). Pressurized (center to supply) balances both sides — holds a position roughly against external forces.
Center-exhausted is the favorite for commissioning and manual adjust modes. Center-pressurized gives pseudo-clamping: both piston areas pressurized, the differential area still creates net force on single-rod cylinders (they creep INTO the rod side — never forget area difference!).
Holding precision mid-stroke is a hydraulic job; pneumatics + 5/3 closed = seconds to minutes of drift. For real stops use rod locks / clamping units.
Voltage, inrush, holding, IP class: the small electrical print that kills machines.
Voltage, inrush, holding, IP class: the small electrical print that kills machines.
Coils come in 24 V DC (controls standard), 110/230 V AC (direct line), with DIN EN 175301 connectors (Form A/B/C) carrying LED + suppression options. AC coils draw heavy inrush until the armature seats — a stuck spool burns the coil.
DC coils need suppression diodes/varistors across switched contacts or the inductive kick eats PLC outputs and relays. LEDs in connectors speed diagnosis enormously (signal present? coil alive?).
100 % duty rating matters for held positions; check ambient temperature class with coil self-heating (T-classification in EX zones).
How the valve gets told: buttons and rollers in the air world.
How the valve gets told: buttons and rollers in the air world.
Manual actuators (pushbutton, mushroom, selector, foot) build human interfaces. Mechanical actuators (roller lever, idle-return roller, stem) sense machine positions. Pneumatic pilots amplify small signals into big spools; electric solenoids bridge control and power.
The idle-return roller actuates only in one travel direction — the classic element for end-position sensing in step chains (it must NOT trigger on the return pass). Three-way stem actuators sense precise positions.
Roller lever placement rule: cam approaches with ≈ 30 °, never head-on. Pilot valves get clean air upstream of the power valve.
Free flow one way, metered the other — the most-used fitting in pneumatics.
Free flow one way, metered the other — the most-used fitting in pneumatics.
An adjustable needle in parallel with a check valve: in one direction air passes freely (check open), in the other only through the needle. Mounted directly at the cylinder port it controls speed in exactly one direction.
Rule: meter the air LEAVING the cylinder (exhaust-air throttling). The piston then works against a compressed air cushion — stiff, repeatable speed. Supply-air throttling lets the load lurch until pressure builds: spongy.
Identify type by symbol orientation on the fitting; mark the adjusted needle paint after commissioning. Nearly closed needle + big piston = hours of confusion for the uninitiated.
Exhaust at the cylinder, not 5 m away: the spring-return speed booster.
Exhaust at the cylinder, not 5 m away: the spring-return speed booster.
Mounted at the cylinder port, it exhausts locally through a large orifice when supply pressure collapses, instead of pushing the exhaust air back through the long line and the valve.
Especially effective for single-acting cylinders where return speed depends on exhausting the full chamber, and for long line runs whose resistance would slow retraction.
Its exhaust port needs the silencer (it vents at the machine, loudly). Do not confuse with a shuttle valve — similar shape, different function.
Two inputs, one output, a ball that picks the stronger: pneumatic OR.
Two inputs, one output, a ball that picks the stronger: pneumatic OR.
A free ball or spool inside seals the weaker inlet; the pressurized inlet reaches the output. Two buttons (or two automatic signals) can trigger the same actuator independent of each other.
Logic layers cascade: shuttle trees implement multi-way OR chains in pure pneumatics. The piston-type versions switch more crisply at low pressures than ball versions.
Beware back-pressure tricks: a leaking second input path can hold the shuttle mid-way. Keep both input lines short and defined.
Only when BOTH inputs are live does the output breathe: pneumatic AND.
Only when BOTH inputs are live does the output breathe: pneumatic AND.
Two pistons oppose each other: pressure must reach BOTH sides so that the weaker one passes to the output while the stronger assists sealing. The building block of two-hand circuits and safety interlocks.
Output pressure equals the LOWER of the two inputs — mixing supply pressures creates surprises; feed both from the same regulator.
Response time depends on both signals arriving; long line differences skew simultaneity — route symmetrically in two-hand circuits.
A needle, a small reservoir, a 3/2: pneumatic timers from 0.2 to 100 seconds.
A needle, a small reservoir, a 3/2: pneumatic timers from 0.2 to 100 seconds.
Pilot air bleeds through an adjustable needle into a tiny accumulator; when its pressure crosses the 3/2 switching threshold, the output flips. Versions: ON-delay (signal appears after t), OFF-delay (signal persists t after removal).
Accuracy 5-10 %, temperature-dependent — plenty for process pauses (glue setting, blow-off duration), useless for precise takt. Reservoir volume times needle setting defines the range.
Place near the pneumatic logic board with clean air: the micro-needle clogs gladly. Mark set times next to the valve.
When force is reached pressure plateaus — and this valve converts that into the next step.
When force is reached pressure plateaus — and this valve converts that into the next step.
Pneumatic analog of the hydraulic sequence valve: it opens its output only when inlet pressure exceeds the set value. Classic use: clamp until full contact pressure, THEN advance the working cylinder.
Works because a stalled cylinder builds line pressure to regulator setting, whereas a moving one consumes it. The switching point setting sits between travel pressure and stall pressure.
Needs ≥ 1.5-2 bar margin between travel and sequence pressure; slow cylinders with high friction make this margin vanish — then use travel sensing instead.
N pulses in, one pulse out: batching pieces with pure air logic.
N pulses in, one pulse out: batching pieces with pure air logic.
A preset counter decrements with each input pulse; at zero it emits an output signal and resets (or blocks until manual reset, by type). Counting ejected parts, strokes, or cycles without electricity.
Combined with shuttle/AND valves, counters build complete batch machines: press n times, index, repeat. Reset behavior (auto vs manual) defines the machine start conditions.
Pulse length limits: signals shorter than ~8-15 ms may not register — debounce mechanical signal generators.
From 95 dB bang to 75 dB breath: sintered bronze, the cheap hearing protector.
From 95 dB bang to 75 dB breath: sintered bronze, the cheap hearing protector.
Porous elements (sintered bronze, PE, sintered metal felt) diffuse the exhaust jet, cutting noise 20-35 dB while keeping flow resistance modest. Every valve exhaust port deserves one.
Oversized silencers cost money; undersized ones throttle cycle time — the flow rating must match the valve. Oil mist from lubricated systems clogs silencers: rising back-pressure is the early warning.
In safety-relevant exhaust paths (dump valves!) verify silencer flow explicitly — a clogged silencer can double exhaust time.
Silencer and needle in one: set cylinder speed right at the valve exhaust.
Silencer and needle in one: set cylinder speed right at the valve exhaust.
A sintered silencer with integrated throttle screw mounts in the valve exhaust ports (3 and 5). Adjusting it meters the exhaust air for BOTH or one direction depending on port.
Fine for simple fixtures; for precise cylinder speed prefer port-mounted throttle-checks (control per direction at the cylinder). Heat: the throttling happens at atmospheric expansion — icing can creep into heavily used silencers.
Swapping needle settings between 3 and 5 swaps extension/retraction speeds — the classic commissioning mix-up. Label after tuning.
One-way discipline: protecting supply networks and holding pressure in reservoirs.
One-way discipline: protecting supply networks and holding pressure in reservoirs.
Simple poppet check valves stop backflow: between the machine and a drooping network (holding local reservoir pressure), on receiver inlets, and in logic circuits implementing path exclusivity.
Pilot-operated check versions trap cylinder air and release it on a pilot signal — the pneumatic relative of the hydraulic load-holding valve, used with 5/3 center-closed for longer (still imperfect) stops.
Spring cracking pressures (0.2-0.5 bar) matter in vacuum and low-pressure circuits — select soft springs there.
Spring-closed, air-released: the brake that makes pneumatic axes hold honest positions.
Spring-closed, air-released: the brake that makes pneumatic axes hold honest positions.
Clamping cartridges around the rod close by spring force and release at 4-8 bar pilot. On air loss the rod is mechanically locked — load stays, safety category satisfied with proper sizing.
Static holding: spec holding force 1.5-2 × the maximum axis load. Dynamic (emergency braking from speed) is a different, verified-rating business — few units are certified for frequent dynamic stops.
Mount concentric, rod surface spec (hard chrome, tolerance f7/h6), and pilot air from the SAFE side of the control chain. Orderly release after air-loss recovery prevents jumps.
16 valves, one cable: the manifold grew a brain.
16 valves, one cable: the manifold grew a brain.
Modular terminals group 4-64 valve slices on common supply/exhaust galleries with a single fieldbus/Ethernet node (PROFINET, EtherNet/IP, IO-Link). Wiring shrinks from cables to addresses.
Diagnostics per slice (coil open/short, voltage faults, cycle counters) flow into the PLC — predictive maintenance for free. Pressure zones within one terminal (separated galleries) implement multiple pressure levels elegantly.
Zoning and exhaust back-pressure sanity matter: a dump-safety slice should NOT share a gallery the E-stop cannot evacuate.
1300 Nl/min nominal, but at what Δp? Learn the four languages of air flow.
1300 Nl/min nominal, but at what Δp? Learn the four languages of air flow.
Pneumatic flow is compressible: above the critical pressure ratio the valve chokes (sonic flow), after which downstream pressure no longer increases flow. Manufacturers state flow as nominal Nl/min (@6 bar 1 bar drop), Cv/Kv (imperial/metric liquid-like), or ISO 6358 sonic conductance C with critical ratio b.
C (dm³/(s·bar)) describes choked-state throughput; b says at which downstream/supply ratio choking ends. These two fully describe the valve for any condition — the modern datasheet form.
Never compare Cv against C directly (convert!). For cylinders, valve C should support the peak consumption; undersizing shows as slow return strokes first.
Air pushes out, a spring pushes back: one hose, one job.
Air pushes out, a spring pushes back: one hose, one job.
One port, one power stroke; the return comes from an internal spring. Typical strokes to ~100 mm (longer = huge spring and housing).
Effective force = p × A minus spring force (which grows along the stroke — force at stroke end is notably lower than at start!). The spring-side chamber vents through a filtered breather.
Economics: cheap valve (3/2), half the air of double-acting, slower return, spring force lost as usable force. Push-out versions common; pull-types exist.
Air both ways: full force out, reduced force back (the rod occupies area).
Air both ways: full force out, reduced force back (the rod occupies area).
Two chambers, alternating supply. Extension force uses full piston area; retraction uses the annulus (piston minus rod), roughly 75-85 % of extension force.
Speed at equal flow is FASTER in retraction (smaller volume!). Air consumption per cycle: both chamber volumes at supply pressure per stroke — the sizing sheet entry for compressor load.
Size bore from required force with 25-50 % margin at the chosen pressure, then check speed/charts for the flow side. Long strokes: buckling rules for thin rods apply just like in hydraulics.
The last 10-25 mm decide whether the machine sounds like a toolbox or a church.
The last 10-25 mm decide whether the machine sounds like a toolbox or a church.
Near stroke end a spear closes the main exhaust passage; the trapped air escapes only through an adjustable needle — a decelerating air cushion absorbs the kinetic energy of piston, rod and load.
Cushioning capacity relates to kinetic energy (mass × v²/2) relative to cushion volume and pressure. Too soft: end bang. Too hard: bounce/stick before end — the needle finds the sweet spot per load.
For masses/speeds above the chart, order EXTERNAL shock absorbers — cylinders with overloaded cushions die young. Elastic seals (no cushion) only for tiny fast cycles.
When the load is not coaxial: guide rods or compact bodies share the job.
When the load is not coaxial: guide rods or compact bodies share the job.
Guided cylinders integrate guide rods with plain or ball bearings beside the piston rod (taking side loads and torques); compact (ISO 21287) cylinders offer short overall length with similar bore force.
Side load on a plain cylinder wears the rod bearing and seal oval — premature leak. Guided versions convert side forces into guide-bearing forces with documented ratings.
Ball bushings: precise, sensitive to side impact; plain (sintered) bushings: tough, more play. Eccentric load derating from the catalog graph, never from hope.
Stroke as long as the body: gantry axes without telescopic tricks.
Stroke as long as the body: gantry axes without telescopic tricks.
The piston couples to an external carriage either magnetically through the tube (clean, limited force) or mechanically via a slot sealed by a steel band (full force, fine dust vulnerable).
Footprint halves compared to a rod cylinder of equal stroke. Long strokes with external guides ride on linear rails; the cylinder only supplies the pull.
Magnetic coupling decouples on overload (protection AND position loss — re-home after!). Slot versions need band inspection intervals in dirty halls.
90 or 180 or 270 degrees of air-driven torque for swivels, flaps and grippers.
90 or 180 or 270 degrees of air-driven torque for swivels, flaps and grippers.
Vane actuators: pressure on a vane in a round chamber gives direct rotation (compact, limited angle ~270 °). Rack-pinion: a linear piston drives a gear rack over a pinion (higher torque, adjustable angle stops).
End cushioning via adjustable stops with elastomer or shock absorbers. Rack units allow mid-stroke multiple stops with accessory valves.
Torque derating at angle extremes for rack types (mechanical advantage varies). Check radial loads on the pinion shaft from attached tooling.
Stall it, drop it, soak it: the EX-proof continuous drive that simply refuses to die.
Stall it, drop it, soak it: the EX-proof continuous drive that simply refuses to die.
Air motors convert expansion into rotation (vane) or crank strokes (radial piston). Characteristic: max torque at stall, max power at ~50 % free speed — infinitely and safely stallable.
No electricity at the motor: intrinsically explosion-proof, sealed versions run underwater or in paint zones. Efficiency is poor (15-25 %) — use where robustness beats kWh.
Speed control by exhaust throttling keeps torque behavior; supply throttling reduces both. Mufflers mandatory; expansion cooling can ice vanes at high continuous duty (dry the air!).
Catch a moving pallet mid-air: the vertical fist in the roller conveyor.
Catch a moving pallet mid-air: the vertical fist in the roller conveyor.
Robust short-stroke cylinders (often with lever/roller heads) rise into the pallet path, absorb kinetic energy through damping versions, then retract to release the goods carrier.
Impact damping versions (air cushion plus shock absorber head) take the moving mass; simple versions suit gated (zero-pressure) accumulation. Sensor slots report the gate state.
Rate by moving mass × conveyor speed from the nomogram — never by static weight! Rebound protection prevents pallet bounce on fast lines.
Two fingers, three fingers, or angled: choose by part geometry and access.
Two fingers, three fingers, or angled: choose by part geometry and access.
Parallel grippers move jaws straight (precise gripping, needs jaw access); angular grippers swing jaws (clearance above the part); three-jaw concentric centers round parts automatically.
Gripping force at given pressure from the jaw-length derating curve (lever effect!). Internal gripping (from inside a hole) uses the same bodies with reversed fingers.
Safety: with air loss, parts drop — select spring-assist closing (fail-close) for held parts, or grip-rests in the station. Sensors (magnetic field) sense jaw position/proximity.
Contract like a muscle, lift like a bag: unconventional linear drives.
Contract like a muscle, lift like a bag: unconventional linear drives.
Fluid muscles: a rubber tube in a fiber braid contracts up to 25 % when pressurized — enormous initial force, no stick-slip, no seals. Bellows cylinders: convoluted rubber chambers lift tons at 2-8 bar with short strokes.
Muscles behave like springs (force falls along contraction) — beautiful for human-collaborative clamping; bellows isolate vibration while lifting (vehicle leveling!).
Muscles: watch braid damage (sharp edges), end fitting alignment only. Bellows: horizontal forces forbidden without guides.
Compressed air in, vacuum out: single-stage and multi-stage Venturi physics.
Compressed air in, vacuum out: single-stage and multi-stage Venturi physics.
A jet of compressed air through a nozzle entrains the connected volume (Bernoulli/ejector effect). Single-stage: simple, high vacuum flow. Multi-stage: higher efficiency at low vacuum levels — less air per suction liter.
Two specs matter: evacuation time for the connected volume (how fast the cup grips) and air consumption (running cost). Optimal working vacuum for porous parts vs dense parts differs (dense: 60-80 %, porous: low vacuum/high flow ejector).
Energy saving units switch the ejector off once holding vacuum is reached (with check valve) — payback in months on 24/7 lines.
The interface to the workpiece: flat, bellows, oval, foam — and the switch that verifies the grip.
The interface to the workpiece: flat, bellows, oval, foam — and the switch that verifies the grip.
Cup shapes fit surfaces: flat for rigid sheets, bellows for uneven/curved (compensates height), oval for narrow parts, foam for rough surfaces. Materials: NBR general, silicone food/hot, PU wear, mark-free NBR for glass.
Holding force = Δp × effective cup area; horizontal (shear) loads on smooth cups are friction-only — derate 4x and never trust shear at high acceleration. The vacuum switch confirms grip before motion (cycle permission!).
Cup count with redundancy: never design a crane where one torn cup loses the load. Check valves per cup isolate a single failure.
DN by inner diameter, PU for energy chains, PA for precision — and why push-in joints decide leaks.
DN by inner diameter, PU for energy chains, PA for precision — and why push-in joints decide leaks.
PA (nylon): rigid, precise, stable — panel builds. PU: flexible, kink-tolerant — moving applications, robots, energy chains. Push-in fittings grip the tube by an internal claw with an O-ring seal: fast assembly, demount cleanly by pressing the release ring.
Leak statistics: tubing cut with a dull knife at an angle is the top cause of micro-leaks. Tube cutters exist for a reason. Bend radius violations throttle flow permanently.
Size by inner diameter from the flow consumption chart — "6 mm tube" means OD 6, ID 4! Pressure drop budget: keep tube length < 3 m for fast cylinders or step up diameter.
One supply, eight consumers, three pressure levels: the plumbing that keeps panels sane.
One supply, eight consumers, three pressure levels: the plumbing that keeps panels sane.
Distribution blocks fan out supply air; sandwich regulators under each valve slice create per-function pressure levels off the same manifold (saving separate lines).
Exhaust galleries similarly common: back-pressure from a choked common exhaust affects ALL slices — size exhaust chokes and consider silenced manifolds.
Unused stations get blanking plates (and are documented spare capacity!). Line labels per consumer prevent the classic crossed-hose afternoon.
A magnet rides the piston; a switch in the slot tells the PLC where it went.
A magnet rides the piston; a switch in the slot tells the PLC where it went.
Reed switches: two ferromagnetic tongues close in the piston magnet field (cheap, needs bounce debounce). Solid-state (GMR/magnetoresistive): bounce-free, precise switching point, longer life at high cycle rates.
Hysteresis and repeatability define usable accuracy: end-position sensing with ±0.2 mm solid-state is routine; mid-position multi-sensing uses several switches or continuous analog position sensors.
Mind weld fields (migrated sensors lie!), strong external magnets, and correct slot/brand match. LED indicator = commissioning friend.
Mechanical snap switches gave birth to IO-Link transducers: pressure became data.
Mechanical snap switches gave birth to IO-Link transducers: pressure became data.
Mechanical pressure switches snap at a spring-set threshold (cheap, drift). Electronic switches offer two programmable thresholds with display; transducers output 4-20 mA / 0-10 V / IO-Link continuous values.
Modern practice: vacuum switches verify grip, pressure switches verify clamp force and regulator health, transducers trend leakage overnight (pressure decay = leak signature).
Hysteresis setup prevents relay chatter at the threshold; peak-hold functions catch spikes the PLC cycle would miss.
You cannot manage what you do not meter: machine-level air accounting.
You cannot manage what you do not meter: machine-level air accounting.
Thermal mass-flow or differential-pressure sensors measure Nl/min per machine, reporting to energy dashboards. Off-shift readings reveal base leakage with surgical honesty.
Placement at each production line feeder transforms leak hunting from annual campaigns to continuous management: consumption deltas after each fix quantify savings in currency.
Bidirectional (charge/exhaust) confusion: measure supply leg only. Pipe diameter straight runs required for the flow profile.
Button to valve to cylinder: the three-component hello-world of pneumatics.
Button to valve to cylinder: the three-component hello-world of pneumatics.
A 3/2 NC pushbutton valve feeds the SA cylinder directly. Press: extend. Release: the valve springs home, the cylinder is exhausted through port 3, the return spring pulls the rod back.
Direct control works while the valve flow rate satisfies the cylinder demand; that is usually true up to φ 50 cylinders with short tubes. Speed setting via throttle-check at the cylinder (exhaust side).
Valve close to the cylinder shortens exhaust path = faster spring return. Label the exhaust silencer: it IS the return path.
Small signal moves big spool: separating signal level from power level.
Small signal moves big spool: separating signal level from power level.
A small 3/2 signal valve pilots the large 5/2 power valve that feeds the cylinder. Long distances, big cylinders, weak sensors — the pilot stage solves all three.
Pilot pressure needs ≥ 2-3 bar reliably; long pilot lines delay switching (fill time). Internal vs external pilot supply: external pilot required for vacuum or < 3 bar supply duties.
Standardize pilot voltage and connector type plant-wide. The LED in the pilot connector is the first diagnostic reading on any dead station.
Press once, it goes; press another, it returns: start/stop logic without holding.
Press once, it goes; press another, it returns: start/stop logic without holding.
Start button pulses the extend side of a double-pilot 5/2 memory valve; stop (or end-position sensor) pulses the retract side. The valve holds its last state with no signal applied.
The detent holds position through power/air interruptions — sophisticated behavior, but the safety audit asks: with air loss and recovery, what moves? Answer deliberately, document it.
Signals must be pulses, not maintained levels: overlapping signals from both sides fight. Manage initial state at machine start (defined pulse or initialization routine).
Reach the end, roll the lever, come home: the fundamental automation step.
Reach the end, roll the lever, come home: the fundamental automation step.
A roller-lever valve (3/2) at the end position senses the extended cylinder and pulses the retract side of the memory valve — full cycles from a single start button.
Idle-return roller versions actuate in ONE direction only, so the returning cylinder passing by does not retrigger. Where to mount? End of stroke with overtravel margin of the adjustable lever.
Cam approach angle ≈ 30 ° for roller life: head-on hits destroy stems. Sensor alternative: reed switch + PLC equivalent.
Both palms within 0.5 s or nothing — the pneumatic safety circuit par excellence.
Both palms within 0.5 s or nothing — the pneumatic safety circuit par excellence.
Two 3/2 pushbuttons feed a dual-pressure (AND) valve; the AND output pilots the power valve. Anti-repeat and simultaneity supervision come from a purpose-built two-hand relay or a pneumatic timing block.
The 0.5 s simultaneity window and release-to-repeat requirements outlaw simple AND plumbing for press duties — certified two-hand devices implement them with monitored redundancy (type IIIC per EN 574).
Distance buttons ≥ 300 mm (inside) against one-hand-one-elbow defeat; shrouded against knee/elbow tricks; position outside the danger zone reach (calculate stopping time!).
Extend, wait three seconds while the glue sets, return — timed pneumatics.
Extend, wait three seconds while the glue sets, return — timed pneumatics.
The end-position signal starts a time-delay valve; its delayed output pulses retract. ON-delay topology: output appears after t; OFF-delay: output removed t after input falls.
Accuracy matters little for glue (10 % error tolerable), a lot for takt: balance needle setting against cycle time target, log actual dwell with the sensor data.
Dust ruins the micro-needle: filtered pilot air and an occasional blowing-through during maintenance. Note temperature drift across seasons.
Steal 40 % of the cycle time back by venting at the cylinder.
Steal 40 % of the cycle time back by venting at the cylinder.
A quick-exhaust valve at the cylinder inlet dumps the return chamber locally instead of through 5 m of tubing and the control valve. Retraction accelerates dramatically on single-acting setups.
Measure first: long lines, large bores and light springs benefit most. The effect stacks with port-size increases — but quick exhaust alone is the cheapest upgrade.
Silence its exhaust (it vents at the machine, suddenly loud). Seal versions matter in dusty shops.
No sensors: the clamp pressure itself authorizes the drill.
No sensors: the clamp pressure itself authorizes the drill.
A pressure sequence valve monitors the clamp line; while the clamp cylinder is still traveling (pressure low), the drill branch waits. Once the clamp stalls on the part, pressure peaks, the sequence valve opens, the drill advances.
Travel pressure vs stall pressure margin defines reliable switching (≥ 1.5 bar). Leaky clamps creep below the threshold mid-cycle — the drill aborts with cryptic irregularity (now you know why).
Prefer sensor-based sequencing where margins are tight; pressure logic shines in harsh environments where sensors die.
The design method that makes ANY multi-cylinder sequence buildable without overlap errors.
The design method that makes ANY multi-cylinder sequence buildable without overlap errors.
Group division: split the motion sequence into groups so that no cylinder switches twice inside one group. Each group gets its own supply line, activated by a step (memory) valve; signals from other groups cannot disturb because their supply is off.
Example A+ B+ B− A−: groups [A+] / [B+ B−] / [A−]. The start signal enters group 1; the end-position of the LAST motion in a group enables the next group supply via a shuttle tree. Result: at every moment only ONE group has signal pressure — signal overlap is structurally impossible.
This method scales to 6+ cylinders where intuition collapses. Modern translation: the PLC does the grouping in software, but understanding cascade makes you design correct PLC sequences faster.
Industrialized cascade: ready-made step-chain modules that snap together.
Industrialized cascade: ready-made step-chain modules that snap together.
Commercial step-chain modules (pneumatic relays in a cascade-tolerant block) implement step i: output to actuator i active, plus input interlock requiring previous step completion. Machine sequence = wired module order.
Advantages over home plumbing: defined initial step, manual step override, fewer assembly errors. The same thinking survives into PLC ladder step logic (SFC/GRAPH).
Documentation: the module map IS the control documentation. Spare modules on the shelf change MTTR dramatically in remote plants.
Extend gentle, retract strong — or clamp soft on one side only: per-direction pressure levels.
Extend gentle, retract strong — or clamp soft on one side only: per-direction pressure levels.
Pressure regulators in the valve slice (or sandwich plates under individual valves) throttle the supply per cylinder port direction. The throttle-check valves still meter speed; pressure meters force.
Classic misconfiguration: reducing supply pressure on a meter-out circuit collapses the cushion stiffness — speed stability dies. Reduce pressure on the SIDE THAT PUSHES, keep exhaust cushion strong.
Gauge per zone pays for itself in one troubleshooting session. Regulator droop under dynamic flow: verify at operating speed, not static.
Two sensors, one memory valve: endless reciprocation until told otherwise.
Two sensors, one memory valve: endless reciprocation until told otherwise.
End-position sensors pulse the two sides of a memory valve alternately: reach A, retract; reach B, extend. A start/stop selector gates the chain.
Frequency from flow settings and stroke; duty cycles wear valves: solid-state sensors and high-cycle valves pay off at > 10 cycles/min. Overrun of the sensor positions ends in travel-error — include margin or timeout supervision.
Add a cycle counter to log wear for maintenance intervals; the timeout interlock (no end-signal after 2× expected time) catches slipping cams early.
Ejector on, verify grip, move, blow-off: the four-beat rhythm of vacuum handling.
Ejector on, verify grip, move, blow-off: the four-beat rhythm of vacuum handling.
Cycle: ejector generates vacuum, cups approach the part, vacuum switch confirms grip (> threshold), axis moves, at destination the ejector stops and a short blow-off pulse frees the part positively.
Blow-off matters: elastic cups re-stick for 100-300 ms without it. Energy-saving ejectors hold vacuum with a check valve and cycle the jet on demand (70-90 % air saving).
Grip verification BEFORE vertical motion is non-negotiable: drop detection mid-air is damage. Size blow-off gentle (0.5-1.5 bar) to avoid part launch.
Dump now, restart only deliberately: safe-state pneumatics.
Dump now, restart only deliberately: safe-state pneumatics.
The safety chain energizes the dump/soft-start block: E-stop, guard door, pressure-OK, PLC-permission in series. Opening any link exhausts the machine; reclosing arms a monitored RESTART button — the machine restarts only on deliberate press.
After a dump, cylinders are relaxed and possibly mid-position: the restart philosophy must define behavior (hold position with locks? return home first?). Monitored restart prevents the classic accident: guard closed → machine resumes unexpectedly.
Soft-start ensures gentle repressurization after restart. Exhaust-time verification via risk assessment; rod locks where gravity holds hazards.
The machine keeps its own lunch: buffer tank + check valve against plant sag.
The machine keeps its own lunch: buffer tank + check valve against plant sag.
A small receiver at the machine inlet, fed via a check valve, bridges short network pressure drops and peak demands that exceed the drop-line capacity.
Sizing like the central receiver: V covers the peak volume times (pmin margin) — with the check valve isolating the buffer from a collapsing header.
Add a gauge and a relief per local codes. In leak math, remember the buffer hides sag during measurement — isolate it during audits.
Cleaning with air: timed, aimed, and quieter than you learned in the workshop.
Cleaning with air: timed, aimed, and quieter than you learned in the workshop.
Blow-off via 3/2 valve and nozzle array, triggered by part-present sensors, timed by PLC or pneumatic timers. Engineered nozzles (laval/air-amplifier) halve consumption at equal force compared to open pipes.
Air-amplifier nozzles entrain ambient air 10-25× — the compressors favorite upgrade. Noise follows velocity power 6-8: laminar slots beat drill holes.
OSHA-type rules: dead-end pressure < 2 bar on open jets (skin safety), chip guards mandatory. Timed bursts (200 ms) usually beat continuous blow massively.
Extend at 6 bar, retract at 2 bar: up to 30 % air saved on no-load returns.
Extend at 6 bar, retract at 2 bar: up to 30 % air saved on no-load returns.
A regulator or pressure-reduced branch feeds only the retract side; the full supply serves extension. Return force usually only overcomes friction — full pressure there is pure waste.
Savings = volume of retract chamber × Δp fed. At 24/7 takt machines this is one of the top energy retrofits with months payback.
Verify minimum retract force against friction and spring loads; reduced pressure also slows return unless the throttle-check is rebalanced.
Sensors to inputs, valves to outputs, steps to flags: the honest translation method.
Sensors to inputs, valves to outputs, steps to flags: the honest translation method.
Method: 1) write the motion sequence (A+ B+ A− B−); 2) list sensors (a1, a0, b 1, b0) as inputs and valve coils (A+, A−, ...) as outputs; 3) draw the step chain with transition conditions; 4) implement set/reset latches per step; 5) add start conditions (home positions, guards), 6) add fault timeouts.
The trap: feedthrough signals (sensor still true from previous step) fire steps early — the cascade insight returns as the software rule: transitions checked only in their step context.
Simulation BEFORE wiring: replay the sequence in the Studio PLC tab line by line; a dry-run catches 80 % of translation errors.
No PLC, no electricity at the machine: valves, sensors and logic elements run 30 years.
No PLC, no electricity at the machine: valves, sensors and logic elements run 30 years.
Car-wash gantries, ATEX conveyors and wet-area machines use valve logic exclusively: buttons = 3/2 valves, AND = dual-pressure, OR = shuttle, memory = 5/2 double, timer = delay valve, sequence = cascade.
Strengths: intrinsically safe, washdown-proof, EMP-indifferent, repair-with-wrench. Weaknesses: no easy logging, changes mean plumbing, diagnostics by ear and gauge.
Design discipline: signal levels documented (signal vs working air separated by color), panel layout mirroring the schematic, every element labeled.
Air gone, power gone, hose burst: rehearse the three funerals per axis.
Air gone, power gone, hose burst: rehearse the three funerals per axis.
For each actuator ask: air failure behavior (spring return? memory? lock?), power failure (solenoid drop), hose burst (rod lock? counter-balance pneumatic?), and document the resulting machine pose.
Gravity axes fail DOWN without locks — people-space defines need. Clamps fail OPEN with air loss unless NO valves or mechanical lock hold them (do you WANT clamp-hold on failure? sometimes yes!).
Risk assessment tables (hazard × failure mode) turn philosophy into checklists; simulate the failures — simulation makes the table honest.
Every step gets a timer: if the end sensor has not reported, stop and call maintenance.
Every step gets a timer: if the end sensor has not reported, stop and call maintenance.
Each motion step starts a watchdog timer (PLC or pneumatic). If the expected end-position sensor does not switch within 1.5-2× the normal step time, the machine stops with a step-specific fault message.
This converts vague machine hangs into exact diagnoses (which cylinder, which sensor, which valve). Pneumatic implementation: timing valves across step signals; PLC: TON per step with differentiated fault codes.
Timeout values from measured step times + margin, tightened during production optimization. Too tight = nuisance stops; too loose = long jam time.
A selector valve gives the commissioning engineer a legal back door.
A selector valve gives the commissioning engineer a legal back door.
A 3/2 (or PLC bit) mode switch routes signals either from the automatic chain or from manual pushbuttons, interlocked so that automatic start is dead in setup mode.
Pneumatic version: the setup line is a separate gallery pressurized only in setup mode (cascade thinking!). PLC version: mode bit gating the auto/routine outputs plus jog functions per axis.
Mode selection must be key-supervised (keys to authorized people) and the mode displayed. Guards may be muted only under specified setup conditions (risk assessment!).
Whatever the sequence thinks: one mushroom valve sends the tool home.
Whatever the sequence thinks: one mushroom valve sends the tool home.
A manually actuated valve (often pneumatic pilot on the power stage or an electrical override path) forces the process cylinder to retract immediately, bypassing the normal chain — distinct from E-stop (which dumps air).
Use when dumping air would drop a load or lose an in-process part: powered retract is safer than depressurized drift. Combine: E-stop dumps; emergency-retract returns the tool first, THEN dumps.
Override logic must survive PLC failure — hardwired pilot paths or dedicated safety relays. Drill the operators: retract first, dump second.
Speed is sprints and surgeons: 60 mm/s to the part, 4 mm/s through it.
Speed is sprints and surgeons: 60 mm/s to the part, 4 mm/s through it.
A second throttle path with a switching valve (or cam-actuated 3/2) changes the exhaust restriction mid-stroke: fast until the work point sensor, then precise slow feed for the process.
Classic drilling unit logic. Air-compressibility bump: the switching moment relaxes the air cushion — minimize with a properly pressurized meter-out design and short switching volumes.
Place the sensor/cam with overrun margin; the speed change position tolerance adds directly to depth accuracy.
Extend slow, retract fast: one cylinder, two speed personalities.
Extend slow, retract fast: one cylinder, two speed personalities.
Throttle-check valves at both cylinder ports, both oriented to meter OUT of the cylinder, give independent speed control per direction — the professional standard.
The common mistake of metering IN shows its usual symptoms: spongy starts and load-dependent speeds. Exhaust metering keeps the piston squeezed between two air cushions: stiff.
Retraction quicker than extension is natural (annulus area gain) — expect asymmetric speeds even at equal settings.
Air compresses: true pneumatic sync is a dream; useful sync is engineering.
Air compresses: true pneumatic sync is a dream; useful sync is engineering.
Two cylinders fed from the same valve with individually balanced throttle-checks move approximately together (±5-10 %). Mechanical coupling (shaft, bridge) or hydro-pneumatic feeds are needed for real synchronism.
Load differences translate directly into speed differences through the compressible medium — balance by throttle, verify by sensors, accept the spread or change technology.
Never promise < 5 % position spread without mechanical means. Guide the load so that skew jams are geometrically impossible even without sync.
Two pistons, one rod: twice the force in the same diameter.
Two pistons, one rod: twice the force in the same diameter.
A tandem cylinder stacks two pistons on a common rod with two pressure chambers for extension: force adds (nearly) double at stroke lengths of a single cylinder, in the tight envelopes where a bigger bore does not fit.
Air consumption doubles (both chambers feed) — the energy side of force doubling. Retraction works on the single annulus as usual.
Consider instead: pressure booster circuits if the plant allows higher local pressure, or a lever. Compare envelope, energy and cost for the specific case.
Press five times, index, repeat: a counter orchestrates a small series automatically.
Press five times, index, repeat: a counter orchestrates a small series automatically.
A preset counter receives end-of-stroke pulses; at zero it fires the index step and resets itself. The classic packaging rhythm without a single line of code.
Fault thinking: a missed pulse (sensor glitch) desynchronizes batch and counter — supervision via index sensor OR-ing the count ensures alignment, or the batch photoelectric check closes the loop.
Manual reset and batch-display are operator essentials. For batches changed daily, PLC costs less than operator confusion.
Down, grip, up, across, down, release, up, back: the eight-step machine that teaches sequencing.
Down, grip, up, across, down, release, up, back: the eight-step machine that teaches sequencing.
Three axes (Z-lift, X-transfer, gripper) run a complete handling cycle. Sequence design: Z− grip Z+ X+ Z− release Z+ X− — each step sensor-confirmed, grip verified before Z+.
Energy and speed: simultaneous axes where geometrically independent (Z+ and X+ can overlap when clear!) halves cycle time — step-chain groupings reflect this.
Grip verification, part-present at both stations, and vacuum-release blowoff timing are the three details that separate drawing board from line.
Parts vary; the fixture must forgive: compliance, travel margin, grip detect.
Parts vary; the fixture must forgive: compliance, travel margin, grip detect.
Design clamps with travel margin beyond nominal contact (spring-loaded jaws, bellows cylinders, telescoping stops) instead of dead-end positions, and verify clamp by pressure (sequence valve) or jaw-position sensor rather than by stroke end.
A clamp designed to touch exactly at its end-position jams on +1 mm parts and fails open on −1 mm parts. Float in the mechanism plus verification signal solves both.
Clamp-force window: sequence valve between minimum-safe-grip and part-deformation pressures — calibrate per part family, document on the setup sheet.
Cardboard, wood, fabric: when the part breathes, vacuum flows instead of building.
Cardboard, wood, fabric: when the part breathes, vacuum flows instead of building.
Porous parts leak air through their surface: the ejector must supply high FLOW at modest vacuum level (multi-stage ejectors, oversized cups with flow paths) rather than high vacuum at low flow.
Sealing cups with soft lips, larger gripping area (more holes paralleled), and continuous flow holding (no check-valve energy savers!) characterize porous handling. Grip verification uses flow-compensated vacuum switches.
Cup material hard-wearing (cardboard is sandpaper): PU lasts. Estimate leak flow empirically — datasheets help only after the first test.
The layer must stay held even when the plant hiccups: reserve volume and check valves.
The layer must stay held even when the plant hiccups: reserve volume and check valves.
A vacuum reserve tank with check valve per cup array keeps grip during ejector off-phases or supply dips; the vacuum switch supervises and refuses motion if reserve integrity is lost.
Sizing: reserve volume × allowable Δvacuum must cover the hold-time times the leak flow (sum of cup leaks + porosity). Measure, then double it.
Test reserve decay weekly (stop ejector, log decay seconds); the decay time IS your safety margin reading.
Compressed-air supply, hydraulic smoothness: the drilling feed without stick-slip.
Compressed-air supply, hydraulic smoothness: the drilling feed without stick-slip.
An air-over-oil unit: air pressure drives an oil-filled damper cylinder inline with the working cylinder. Oil incompressibility plus an adjustable hydraulic throttle yields feed smoothness air alone never reaches.
Rapid approach via air path bypass, work feed via closed oil circuit through a precision needle valve: mm/s stability at any load variation, with automatic re-coupling of chambers on return.
Oil level and bleed ritual keep it honest: air in the oil leg = spongy feed, diagnosed in seconds by the bubble in the sight glass.
Counterweight in a hose: the axis feels weightless at any position.
Counterweight in a hose: the axis feels weightless at any position.
A continuous pressure (precision-regulated, relieving!) on one cylinder chamber offsets the dead weight: motors/tools become feather-light on their guides (tool balancers, Z-axis assist).
Force must equal weight within the travel: pressure set = weight / area, verified at mid-stroke. Relieving regulator vents compression-backpressure on lowering — that is why precision-relieving types, not standard reducers.
Air failure = sudden weight: pair with rod lock in vertical safety cases. Hysteresis of cheap regulators makes the axis drift-hunting — precision regulator quality shows here.
Three truths before motion: door closed, part present, operator clear — AND them properly.
Three truths before motion: door closed, part present, operator clear — AND them properly.
Safety-rated door switch into the safety relay; part-present and operator presence as process sensors into the PLC; the cycle start requires all three plus the restart condition. Pneumatic equivalent: door valve + part valve + start valve through AND cascades.
Safety functions separate from process functions: guard = hardwired rated channel, process conditions = PLC. Mixing a convenience sensor into the safety channel is a design sin.
Tongue switches lie with worn doors: coded or RFID sensors with monitoring defeat operator bypassing fantasies.
Exhaust freezes at -5 °C ambient when expansion does the local cooling to -40.
Exhaust freezes at -5 °C ambient when expansion does the local cooling to -40.
Expansion in valves and cylinders chills locally below the line dew point: iced silencers, stuck spools, brittle seals. Defense: pressure dew point at least 10 K below the coldest point (adsorption dryer!), insulated panels, heated cabinets for outdoor valves.
Ice forms preferentially downstream of metering edges (fast exhaust!) and at silencers. The first frosty morning reveals every marginal drain in the plant.
Cold-weather kit: heated drain lines, antifreeze-rated lubricants where lubricators exist, winter-grade seals for mobile machines.
Every exhaust is particle emission: route it, filter it, or stop making it.
Every exhaust is particle emission: route it, filter it, or stop making it.
Cleanroom adaptations: exhaust air extraction lines from valve manifolds and cylinders to filters, grease-free (cleanroom lubricant) components, low-abrasion materials, laminar-avoidant layouts.
Cylinder rod retraction drags contaminated hall air into the tube: vacuum-drawn extraction collars around rod seals keep the class. ISO-class counting audits apply per class target.
Standard solenoid valves with silencers emit oils and particles — manifold extraction piping at the valve terminal is the cleanroom standard configuration.
Signal or power? Air or electric? The four-question method ends random part-swapping.
Signal or power? Air or electric? The four-question method ends random part-swapping.
Method: 1) does the command reach the valve (LED / pilot gauge)? 2) does the spool shift (override button test)? 3) does flow reach the cylinder (exhaust hiss / throttle position)? 4) does the cylinder move free (mechanics decoupled)? The first NO localizes the fault to one level.
LED on + override moves it = command path problem. Override does nothing = valve/air path. Cylinder decoupled moves = mechanics/jam. Each answer excludes half the machine.
Carry: one gauge, one plug connector LED, one manual override tool, one ears. Log findings — repeat faults love pattern blindness.
The valve costs €40, its air over ten years costs €900: buy data, not boxes.
The valve costs €40, its air over ten years costs €900: buy data, not boxes.
Compressed air costs roughly 1-2 currency cents per Nm³. Stations run millions of cycles: consumption per cycle × cycles maps directly to money. Seal friction class, dead volume in fittings, and oversized cylinders pay sleepless bills.
Oversizing cylinders one bore up wastes ~40 % air per cycle forever. Dead legs and long tubes between valve and cylinder add dead volume consumption without work.
TCO comparison in quotes: list price + 7-year air + expected maintenance stops. The cheap supplier often loses only on paper.
Displacement-step diagrams on the panel door: the operator manual engineers actually read.
Displacement-step diagrams on the panel door: the operator manual engineers actually read.
Document sequences as displacement-step diagrams (position vs step) plus the signal matrix (which end sensor + conditions fire which valve). One A4 glued inside the cabinet saves hours per fault for decades.
The diagram reveals design errors before steel: crossing conditions and unterminated loops jump out visually. Review it in the design meeting like code.
Update on every retrofit — out-of-date documentation is worse than none. Photograph the finished panel next to its diagram.
Nl/min, DN: the gas bill is counted at 1.013 bar, 20 °C — everything else converts.
Nl/min, DN: the gas bill is counted at 1.013 bar, 20 °C — everything else converts.
Compressed gas throughput uses normal liters: the volume the air would occupy at reference conditions (1.013 bar absolute, 20 °C, dry). At 6 bar(g) working pressure one operating liter equals ~7 normal liters of consumption.
Confusing Nl/min with l/min is the classic 7× sizing error in both directions. Compressor FAD, valve ratings and consumption calculations all speak normal units.
Always convert before comparing two datasheets; the humidity and temperature corrections matter for dryer loads in summer.
Volume × strokes × pressure ratio: the two-line formula behind compressor sizing.
Volume × strokes × pressure ratio: the two-line formula behind compressor sizing.
Consumption per cycle (Nl) = (cap volume + annulus volume) × (pabs/p0) plus the filling of dead spaces (tubes!). Rates: multiply by cycles/min for average demand; peaks handled by receivers.
Dead volumes dominate short-stroke, high-frequency machines: the tube from valve to cylinder refills every cycle without doing work. Cut it or pay for it permanently.
Typical sanity number: φ50 × 200 mm at 6 bar, 30 cycles/min ≈ 80 Nl/min average. Learn to estimate this in your head.
Every unnecessary bar of line pressure costs 6-8 % compressor energy.
Every unnecessary bar of line pressure costs 6-8 % compressor energy.
Compression work rises with pressure ratio: dropping line pressure from 7 to 6 bar saves ~6-8 % electrical energy plant-wide, before counting reduced air demand (leak flow scales with pressure!).
Leak rate scales roughly with absolute pressure; unregulated consumption devices (open jets, nozzles) consume more at every extra bar. Dual effect: a 1 bar reduction is often worth a small car payment annually.
Identify the real maximum need (usually one legacy consumer), solve it locally (booster! upsizing!), then drop the header.
A 1 mm hole at 7 bar ≈ 60 Nm³/h ≈ four-figure annual waste.
A 1 mm hole at 7 bar ≈ 60 Nm³/h ≈ four-figure annual waste.
Leaks are consumption without counter-force: typical uncontrolled plants lose 25-35 % of generated air. The physics: choked flow at every hole, proportional to area and absolute pressure.
Detection: ultrasonic gun during quiet hours, soapy water for confirmations. Priorities by sound pitch and measured flow — fixing the top five leaks usually halves the loss.
Off-shift flow measurement (production zero) shows base leakage honestly; tag leaks with repair tickets, re-measure after, bank the difference.
Set 6 bar static, get 5.2 bar flowing: the curve nobody reads until takt slips.
Set 6 bar static, get 5.2 bar flowing: the curve nobody reads until takt slips.
Downstream pressure of a spring regulator falls with flow (droop) and differs between rising and falling flow (hysteresis). Datasheet curves at a given inlet pressure tell the story: at rated flow the drop may exceed 1 bar.
Consequences: cylinder forces sag exactly during motion (when flow peaks), pressure switches set close to working pressure chatter, and regulators sized by pipe thread (not flow) underperform dramatically.
Size regulators by their flow curve at your consumption plus margin; pilot-operated models droop far less at high flow.
Why air that leaves the dryer dry still rains in your coldest pipe.
Why air that leaves the dryer dry still rains in your coldest pipe.
Air holds water vapor in proportion to TEMPERATURE, not pressure. Cooling compressed air (expansion, winter pipe runs) condenses the vapor; the pressure dew point (PDP) names the temperature at which condensation begins at actual line pressure.
Refrigeration dryer gives +3 °C PDP: safe indoors. Outdoor winter −10 °C needs adsorption. Vapor load itself scales with inlet temperature exponentially — summer sizing matters.
Monitor PDP online: a drifting dryer shows as rising PDP long before bowls fill. PDP alarm margin: 5 K below the coldest service point.
One cubic meter of summer air carries two spoonfuls of water into your compressor.
One cubic meter of summer air carries two spoonfuls of water into your compressor.
At 35 °C / 80 % RH, inlet air holds ~32 g/m³ water. Compression cannot destroy it: cooling the compressed stream drops most of it as condensate at aftercooler, receiver and first filters.
Plants learn to predict condensate: compressor FAD × humidity function × hours. The remaining vapor (after dryer) condenses only if local temperature falls below PDP.
Condensate from lubricated compressors is oil-contaminated waste: separator treatment before sewer is a legal must in most regions.
Above the critical ratio, downstream no longer matters: flow is frozen at Mach 1.
Above the critical ratio, downstream no longer matters: flow is frozen at Mach 1.
When downstream/upstream absolute pressure ratio falls below the critical value b (≈ 0.528 for ideal air nozzles), velocity at the narrowest point hits sonic speed and flow depends ONLY on upstream pressure. Extra pressure drop buys nothing downstream.
Practical rules: exhausting to atmosphere from > 0.9 bar(g) is always choked — every leak behaves like a choked nozzle; half-open valves under high Δp enter choked regime with strong noise.
For consumption estimates use C and p1 only in choked conditions; sonic conductance data belongs in every valve datasheet (ISO 6358).
6/4 tube feeds φ63 happily; φ125 needs 12/9: the ID chart that decides speed.
6/4 tube feeds φ63 happily; φ125 needs 12/9: the ID chart that decides speed.
Tube ID, length and fittings define the line resistance to the cylinder. Rules: keep total Δp (header + drops + tube) below ~0.5 bar at peak flow, velocity in mains 6-9 m/s, drops 10-15 m/s, direct lines to cylinders under 3 m wherever possible.
Equivalent-length method: each fitting counts as straight tube (elbow ≈ 0.5 m, tee ≈ 1 m) — a short line full of elbows can throttle like 5 extra meters.
When speed lacks, measure Δp DURING motion (instantaneous!): static pressure readings hide everything.
Force × 1.3-1.5, check the rod, then the speed diagram: three steps, no regrets.
Force × 1.3-1.5, check the rod, then the speed diagram: three steps, no regrets.
Bore from required force with load factor: static duty 1.3, dynamic 1.5, high dynamics 2. Rod diameter from buckling tables per mounting style. Speed capability from manufacturer nomograms relating bore/supply/valve size to achievable velocity.
Oversized cylinders waste air every stroke (your compressor bill) and slow the valve dynamics. Undersized ones stall at midday pressure sags. The load-factor margin absorbs seal friction and pressure variation.
End cushioning capacity check for mass × v²: above chart values specify external absorbers. Long-stroke push: buckling first!
The piston hesitates, then jumps: friction physics at low speed.
The piston hesitates, then jumps: friction physics at low speed.
Static seal friction exceeds dynamic: at creeping speeds the piston stores pressure until breakaway, then accelerates and re-sticks — jerk motion (stick-slip). Typical breakaway 0.5-1.5 bar depending on seal type.
Low-friction seals, hydro-pneumatic feeds, or higher minimum speeds avoid the phenomenon; lubricated legacy cylinders stick less but carry oil everywhere.
Select low-friction variants for slow feeds (< 20 mm/s); stick-slip in measurement fixtures ruins data quietly.
Air wins on power density and toughness; electricity wins on energy and flexibility: choose per axis.
Air wins on power density and toughness; electricity wins on energy and flexibility: choose per axis.
Pneumatics: unbeatable for end-to-end motion, clamping, harsh environments, overload stall safety, cost per axis. Electric: better for multi-position, precise profiling, energy per useful work, data capture (torque/position traces).
Energy truth: end-to-end compressed-air efficiency is maybe 10-20 % while electric axes exceed 80 % — at 24/7 duty the kWh decide. Occasional motion: pneumatics still wins by simplicity.
Hybrid decisions per axis: clamping pneumatic, positioning electric is the classic packaging-machine pattern.
Maximum vacuum and maximum suction flow are different knobs: read the characteristic map.
Maximum vacuum and maximum suction flow are different knobs: read the characteristic map.
Ejector curves plot achievable vacuum against free suction flow: at zero flow the highest vacuum (~85-92 %), at zero vacuum the highest flow. Workpieces with leaks (porous) operate down the curve — flow matters, level little.
Evacuation time for a known volume from atmosphere to target level derives from the integral of the curve (datasheet tables give it per liter). Multi-stage ejectors shift the cork: better suction at moderate levels per consumed watt of compressed air.
Never specify maximum-vacuum ejectors for porous duty; oversized supply pressure does not compensate a physics mismatch.
From wall socket to piston work: 10-20 % survives — see every station steal its share.
From wall socket to piston work: 10-20 % survives — see every station steal its share.
Chain: motor (95 %) × compression (50-65 %: heat!) × dryer/purge (85 %) × leaks (65-75 %) × distribution (90 %) × device end efficiency (throttling, friction). Multiplying honest numbers explains the infamous overall figure.
Heat recovery re-discounts compression losses (up to 70-90 % of drive power recoverable as warm water/space heat) — the best plants reclaim most of the loss back into the kettle.
Energy-efficient pneumatics: right pressure level, right bore, no leaks, idle shutoff, heat recovery. In that order.
Idling screws burn 30 % power for air they never deliver: control philosophy matters.
Idling screws burn 30 % power for air they never deliver: control philosophy matters.
Start-stop suits small piston units with receivers (rest is free). Load-unload keeps screw compressors spinning unloaded for fast response — at a shocking 25-35 % of rated power doing nothing. VSD tracks demand continuously and wins on variable profiles.
Unloaded hours × 30 % rated kW is your audit number; receivers smooth the demand so more hours fall into true stop.
Air-system controllers cascade multiple compressors and pick the efficient unit per load; retrofitting them on carrier fleets pays back yearly.
Blow before connect, bleed before bolt, slow before fast.
Blow before connect, bleed before bolt, slow before fast.
Order: visual/mechanical check (tube pulls, fitting torque, sensor targets), blow out lines BEFORE connecting valves, mist-free memory: connect, set service unit pressures low, jog every axis with manual overrides slow, then automatic at reduced speed, then full pressure/speed with documented settings.
Record: regulator readings, throttle positions (painted marks), sensor actuation points, cycle time, noise baseline. These baselines convert future troubleshooting from archaeology to reading.
The first 8 hours logbook (faults + fixes + settings) is the warranty bible. Signatures make it honest.
Weekly drains, monthly silencers, quarterly sensors: the small print of uptime.
Weekly drains, monthly silencers, quarterly sensors: the small print of uptime.
Weekly: drains/bowls, leaks by ear, gauge plausibility. Monthly: silencer back-pressure feel, tube rub points, regulator setpoints drift. Quarterly: sensor switching points, cylinder speed baseline vs record, filter Δp. Annual: dryer service, receiver inspection dates, pressure-switch calibration.
Trend small numbers rather than wait for failures: rising cycle time = silencer or seal; falling regulator reading = filter; rising night flow = new leaks.
Spare kit discipline: the five most-swapped parts (silencers, tubes, seals kit, connectors, one valve slice) live in the cabinet.
Triangles hollow, lines dashed for pilots: reading air schematics at speed.
Triangles hollow, lines dashed for pilots: reading air schematics at speed.
Same ISO 1219 grammar as hydraulics with pneumatic dialect: unfilled triangles (air), exhausted arrows to atmosphere, service-unit chained symbols (separator → filter → regulator → lubricator), and the classic 3/2 and 5/2 boxes with actuation stacks.
Reading method: find the service unit, trace supply to valves, identify each actuator control path, decode pilot signals, verify every exhaust has somewhere to go.
Practice reading the Studio example schematics until you can narrate each cycle: explaining aloud is the fluency test.
Feed, detect, lift, sort two ways, count: one complete machine designed by you.
Feed, detect, lift, sort two ways, count: one complete machine designed by you.
Specification: singulate parts from a magazine, detect metal/plastic, pusher sorts to channel A/B, counter batches by 10. Design: cylinder/valve selection with sizing numbers, sequence with sensor confirmations, E-stop with dump AND restart interlock, batch counter, and the complete documented schematic.
Expected moves: pusher sizing with load factor; 5/2 memory valves or PLC mapping; idle-return sensor placement; supervision timeouts; consumption calculation and service-unit chain specified per ISO 8573 class.
Deliverables: schematic, sensor/valve list with ratings, consumption math, sequence diagram, settings sheet. Build it in the Studio and defend it in review.
15 ms on, 40 ms off: at 30 cycles per minute, milliseconds are the budget.
15 ms on, 40 ms off: at 30 cycles per minute, milliseconds are the budget.
Solenoid valves publish ON and OFF response times (millisecond range). With cylinder fill/exhaust time, these define the maximum cycle rate of a function; exhaust throttles and long tubes degrade the far side of the budget.
AC coils switch slower than DC; large valves shift slower than small ones (spool mass!). For > 5 Hz cycling, select high-cycle valves with reduced spool travel and solid-state driver.
Compute the timing budget BEFORE promising takt: valve ON + fill + work + exhaust + OFF + sensor settle. The Studio chronometer view is your verification tool.
A φ40 gauge at the regulator answers 80 % of first-look questions.
A φ40 gauge at the regulator answers 80 % of first-look questions.
Bourdon-tube gauges display line and working pressure; at the service unit they verify regulator settings at a glance. Glycerine filling damps pointer vibration near pulsating consumers.
Gauge class (±1.6 % typical) matters for settings: a 0-16 bar gauge reads ±0.26 bar — set the 6.0 bar regulator with that in mind. Green/red arcs marked on the dial turn checks into glances.
Isolate gauges from permanent pulsation with a gauge cock or snubber. Marked arc rings: working range green, forbidden red.
Direct hand control of cylinders: rugged, immediate, honest.
Direct hand control of cylinders: rugged, immediate, honest.
5/2 or 5/3 hand-lever valves drive actuators directly on jigs, maintenance rigs and mobile equipment. Detented versions stay in position; spring-centered versions return to neutral.
Lever effort and grip placement define daily operator strain; panel layouts put the dangerous function under a guard flap. Detents carry the memory-valve risk discussion (unexpected motion on air return).
Choose spring-center for inching motions (release = stop), detent for hold-definite states. Label motion direction AT the lever.
Red mushroom, yellow ring, latching: hit it and the machine exhales.
Red mushroom, yellow ring, latching: hit it and the machine exhales.
A large latching 3/2 valve on the panel dumps downstream air when struck; twist or key release re-arms it. It is an air-logic E-stop that works with zero electricity.
Behavior equals the electrical category-0 stop: exhausting is the safe state. Restart requires deliberate manual reset — accidental bump restarts are impossible by design.
Size the exhaust path for the dump-time budget; downstream stored volumes (receivers!) need their own consideration. Position within reach but outside swing zones.
Store pressure, release violently: tons of force for 20 milliseconds.
Store pressure, release violently: tons of force for 20 milliseconds.
A large differential chamber is precharged; releasing the fast poppet fires the piston forward, converting stored gas energy into kinetic impact for stamping, marking, and punching.
Impact energy = stored pressure volume product over the reactable distance; repeated short-stroke hammering beats slow press force for brittle marking materials.
Mounting must survive reaction shock (rigid, guided); NEVER at hand-near stations without two-hand + guard logic. Seals see spectaculaire wear — stock the kit.
Air through the spinning joint: feeding actuators on rotating tables.
Air through the spinning joint: feeding actuators on rotating tables.
A sealed rotating interface passes one or more air channels across a continuously rotating joint (multi-channel coaxial designs for independent circuits). Leakage flows (controlled, small) are normal.
Channels keep pressure separation between circuits; the slip surfaces define life (speed × pressure derating curves). Air preparation upstream decides seal life more than the catalog does.
Mount with a flexible link preventing torque; never let hoses carry the union. Spare-leakage budget belongs in the compressor balance.
10 bar locally from 6 bar plant air: the amplifier that saves the header pressure debate.
10 bar locally from 6 bar plant air: the amplifier that saves the header pressure debate.
A double-piston amplifier compresses plant air to ≈ 2× line pressure for one dedicated consumer (test rigs, clamping, valve actuation). Driven by air itself — no electrics, intrinsically stall-safe.
Driving consumption runs continuously while pumping; output flow is small (high ratio = low flow). A small receiver downstream stabilizes bursts.
Perfect for one demanding consumer so the whole plant can drop one bar (the energy-accounting lesson in hardware). Outlet pressure regulation via pilot air pressure.
Machine asleep, air asleep: pressure switches that cut supply after inactivity.
Machine asleep, air asleep: pressure switches that cut supply after inactivity.
Battery/self-powered modules monitor flow or machine state and close the supply after X minutes of zero activity, restoring on demand signal. Standby leakage dies at the valve.
Plants with hundreds of branched stations harvest 10-20 % of total consumption by killing standby leaks at the branch level — payback months, logic trivial.
Choose versions without continuous bleed. Integrate restart soft-start so wake-up does not slam the machine.
Two coils govern grip elegance: generate, hold, release with positive puff.
Two coils govern grip elegance: generate, hold, release with positive puff.
Compact vacuum valve blocks integrate ejector supply valve, blow-off valve, vacuum switch and optional check hold in one slice on the terminal: grip, transit, release with a timed positive pulse.
Blow-off pulse lengths of 50-150 ms free the part without launching it; the integrated switch gives grip confirmation as a PLC input.
One slice per cup group (zone isolation!): one failed cup must not blind the rest. Flow restrictions per channel tame strong cups on small parts.
Dust rides the gripped surface into your ejector unless a filter says otherwise.
Dust rides the gripped surface into your ejector unless a filter says otherwise.
Inline vacuum filters between cups and ejector catch dust and debris before they clog the Venturi nozzle (whose bore is a millimeter-scale masterpiece of precision).
Filter resistance subtracts from suction capacity: select generously sized bodies, and check elements in dusty duty weekly. Transparent bowls show loading at a glance.
Wood, cardboard, powder handling: filters are not optional there. Position accessible — a hidden filter is a forgotten filter.
The panel needle: coarse speed control and deliberate bleeds.
The panel needle: coarse speed control and deliberate bleeds.
Simple needle valves throttle both directions (no check). Uses: approximate speed caps on non-critical motions, bleed-down rates for reservoirs, damping of gauge lines.
Temperature and oil-film shift their characteristic over weeks; for any speed that appears in a takt calculation, choose throttle-checks instead (repeatability + one-direction discipline).
Mount with direction arrow? There is none — inline throttles are direction-blind, unlike throttle-checks. Paint the set screw after commissioning.
The cabinet must breathe: silenced, downward, and never into the photocell.
The cabinet must breathe: silenced, downward, and never into the photocell.
Valve exhausts inside panels vent through silencers or routed manifolds; the cabinet itself needs filtered breather paths so pressure pulses do not pop doors or push dust.
Exhaust aimed at optics/sensors fogs them with oil mist (lubricated systems!) — route down and away. In cleanroom or food zones exhaust leaves the enclosure entirely via hose.
Sizing breathers: sum of worst-case exhaust flows must pass without +0.05 bar cabinet pressure (door seals!).
Five blocks that build any Boolean machine out of air.
Five blocks that build any Boolean machine out of air.
Beyond shuttle (OR) and dual-pressure (AND): NOT elements invert signals (output = supply when input absent), YES elements gate supply (output = supply when input present), and memory/flip-flop blocks latch states with set/reset ports.
With supply rails and these five blocks, any logic function is constructible: sequences, interlocks, alarms — the computational substrate for pure pneumatic controls.
Signal pressure losses accumulate across cascaded elements: after three OR/AND stages, re-amplify via a YES relay fed from the rail.
When 0.2 bar breakaway or 3 m/s matters, ordinary cylinders stay home.
When 0.2 bar breakaway or 3 m/s matters, ordinary cylinders stay home.
Low-friction cylinders (special seals, polished tubes, grease filling) reach breakaway < 0.2 bar for servo-pneumatic and regulation duty; high-speed variants (special cushioning, seals, guides) run to 3 m/s continuously.
These trade leakage and lifetime for performance: low-friction leaks more (regulation circuits accept it), high-speed seals wear to a plan (spares on the shelf).
Servo-pneumatic positioning loops specify low-friction always — the hysteresis of standard seals destabilizes any control effort.
Nominal flow at WHICH dp, force at WHICH pressure, life at WHICH side load — footnotes rule.
Nominal flow at WHICH dp, force at WHICH pressure, life at WHICH side load — footnotes rule.
Headline numbers hide conditions: nominal flow at 6→5 bar differs from Cv; theoretical force ignores friction (≈ 10 %); cushioning ratings assume specific mass and speed; sensor repeatability presumes clean supply.
Method: find the test condition behind every number (pressure, temperature, side load, mounting), convert to YOUR condition, apply margin honestly. Two brands until now have never meant the same thing by the same word.
When the footnote is missing, ASK. The answer (or silence) is supplier intelligence worth more than the number.
Dew point, particles, oil vapor: sensors where the ISO class must be proven.
Dew point, particles, oil vapor: sensors where the ISO class must be proven.
Continuous monitors (PDP sensor after dryer, particle counter at critical class-1 points, oil-vapor analyzer for food lines) document quality continuously instead of audit-eve sampling.
Alarm philosophy: PDP rising = dryer degrading; particle step = filter breakthrough; trend lines justify maintenance with evidence rather than schedule folklore.
Sampling points per ISO 8573 methodology: isokinetic, short lines, stainless where required — sloppy sampling certifies fiction.
Constant force over a moving roll: precision regulators keeping paper humble.
Constant force over a moving roll: precision regulators keeping paper humble.
A cylinder loads the dancer roll with constant force = half the desired web tension (dancer geometry); the precision relieving regulator holds pressure against load changes, converting air into spring-free constant force.
Relieving capability is the trick: dancer movement charges and discharges the chamber without hysteresis piles. Damping via throttle in the pilot path calms web flutter.
Low-friction cylinders hysteresis-free preferred; keep rod guidance clean (web dust!). Force-integrity scales with regulator quality.
Air plant, oil cylinder: 10,000 smooth Newtons without a hydraulic power pack.
Air plant, oil cylinder: 10,000 smooth Newtons without a hydraulic power pack.
Air-over-oil converters pressurize a closed oil volume from plant air: double-acting-ish hydraulic performance (rigidity, smoothness, high force via intensifier versions) at machines without hydraulic infrastructure.
Classic architectures: converter + oil cylinder for single pushes; hydro-check tandem unit for constant feed speeds; air-hydraulic intensifier + small cylinder for clamping kN ranges.
Bleed points, oil quality (ISO VG 32-46), and temperature effects inherit from hydraulics — the relevant lessons apply unchanged.
A found needle position is an asset: protect it from thumbs.
A found needle position is an asset: protect it from thumbs.
Every tuned throttle/regulator gets: paint mark or sticker on the set position, the value on the settings sheet in the cabinet door, and where abuse is likely a cover or administrator-locked knob.
Half of the operator-induced downtime comes from adjusted-never-restored knobs. Settings sheets turn Thursday-night faults into three-minute recoveries.
Photograph panels after commissioning: the photo IS version control. Setup sheets list only what changes between product variants.
Five boxes that save five nights: silencers, coil, sensor, seals, one valve slice.
Five boxes that save five nights: silencers, coil, sensor, seals, one valve slice.
Analysis of downtime on pneumatic machines points at a tiny set: clogged silencers, dead coils, drifted sensors, worn cylinder seals, broken connectors/tubes. Stocking exactly these covers most incidents at trivial cost.
Consumables with defined service lives belong in the maintenance plan with quantities (silencers per year!). Critical single-sourced valves deserve one backup each if lead time > 3 days.
Shelf discipline: labeled, quantities visible, withdrawal logged — a ransacked shelf is a false friend.
Mass versus speed at the end stop: two axes that decide between cushion and shock absorber.
Mass versus speed at the end stop: two axes that decide between cushion and shock absorber.
Every cylinder datasheet includes a cushioning diagram: allowed moving mass against impact speed for built-in cushions. Points above the curve demand external shock absorbers or speed reduction.
Chart reading ritual: compute your mass × v²/2, find the speed on the diagram (with load!), if above the line — walk to the absorber catalog. Pressure reductions shift curves down derating at lower supply.
Horizontal vs vertical mounting matters (gravity adds energy!). Vertical downward impact gets weight × stroke added to kinetic energy.
A tidy panel with labeled ports teaches more than the component list.
A tidy panel with labeled ports teaches more than the component list.
Training-rig lessons for life: tubes cut square with the cutter, push-in verified by pull test, ports labeled with destination, exhausts silenced, schematic pinned to the rig, tools returned.
Lab discipline mirrors plant discipline: the person who labels the test bench labels the machine; the one who blows lines before connecting saves seals in both worlds.
Rituals: 60-second end-of-session reset (valves to rest, pressure released, regulator backed out) means the next session starts at zero energy — safety as habit.
Slow retract, weekend pressure loss, jitter, midnight fault: recognize each by signature.
Slow retract, weekend pressure loss, jitter, midnight fault: recognize each by signature.
Field-proven signatures: slow retract = clogged silencer or crushed exhaust tube; weekend loss = leak; jitter at start = dried cylinder or dead volume; midnight fault = pressure sag when a neighbor machine starts; random stop = chattering pressure switch; water in bowls Monday = dryer duty undersized; cylinder drifts = 5/3 closed center reality; gripper drops = vacuum switch threshold; soft handles everywhere = meter-in throttling; banging ends = cushion needle open; coil death weekly = AC inrush on sticking spool; phantom sensor signals = weld field or crosstalk.
Each signature maps to ONE subsystem check first. Build your site list from these twelve and MTTR halves.
Write them on the maintenance-room wall; add your site-specific thirteenth when discovered.
Why some edges are wide, some narrow, and why hidden edges are dashed — the alphabet of technical drawings.
Why some edges are wide, some narrow, and why hidden edges are dashed — the alphabet of technical drawings.
Every technical drawing is a sentence written in lines. ISO 128 standardizes the types, widths and application of lines so that a drawing made in Hamburg reads the same in Tehran or Tokyo. Get these three right and 90% of sketch communication is solved.
Wide solid line type 01.2 — visible outlines and edges, the silhouette you can touch. Narrow solid 01.1 — dimension lines, extension lines, leaders, hatching. Narrow dashed 02.1 — hidden edges behind the part. Narrow chain (long dash–dot) 04.1 — center lines and hole axes. Widths follow the ratio narrow : wide = 1 : 2 (typ. 0.25 / 0.5 mm on A4).
Choose line type by meaning, never by look: an edge you could touch = wide solid; an edge behind material = narrow dashed; an axis of rotation or symmetry = chain thin. Do not dash a center line and do not chain a hidden edge.
Where dimension lines sit, how arrows and text are placed, and how limit deviations are written next to a size.
Where dimension lines sit, how arrows and text are placed, and how limit deviations are written next to a size.
A drawing without dimensions is decoration. ISO 129-1 defines the geometry of dimensioning; ISO 406 defines how tolerances are written with the dimension. Together they make a size unambiguous for the machinist.
Each dimension is given once — the shape stays dimension-free. Extension lines start with a small gap from the outline and pass slightly beyond the dimension line; arrows are filled and touch the dimension line ends; text sits above the line, readable from the bottom or right (aligned method). Tolerances: symmetric 50 ±0.1 or bilateral with the upper deviation first 50 +0.1/−0.0.
Dimension to functional surfaces first. Keep dimension lines ≈ 7–10 mm from the outline and 6–8 mm apart from each other. Never let a dimension line cross another line if avoidable, and never use an outline as a dimension line.
Every dimension without its own tolerance still has one — the table that quietly controls your whole sheet.
Every dimension without its own tolerance still has one — the table that quietly controls your whole sheet.
No workshop can produce an exact 50.000 mm, and writing ±0.05 on every dimension would bury the drawing in numbers. ISO 2768-1 defines general tolerances: four classes (fine, medium, coarse, very coarse) applied to every dimension that carries no explicit deviation — declared once in the title block.
For linear dimensions (mm), class m (medium): up to 3 mm → ±0.1; over 6–30 → ±0.2; over 30–120 → ±0.3; over 120–400 → ±0.5; over 400–1000 → ±0.8. Class f (fine) halves most of these; c and v relax them for flame-cut or rough parts. Angles and radii/chamfers have their own rows in the same standard.
Use ISO 2768-m as the house standard for machined parts; switch to f for gauge-fine work. Dimensions where function demands tighter control get their own ISO 406 deviation — the app calls this "tolerance on request".
Hole-basis fits in five minutes: what H7 for a hole and g6 for a shaft actually guarantee at assembly.
Hole-basis fits in five minutes: what H7 for a hole and g6 for a shaft actually guarantee at assembly.
ISO 286 defines the IT tolerance system: a letter for the position relative to nominal (holes capital, shafts lowercase) and a grade number for the width (IT number). In the hole-basis system the hole is always H (lower deviation 0) and the desired clearance is dialed in through the shaft letter.
Every +1 IT grade widens the tolerance ≈ ×1.6. Everyday pairs: H7/g6 sliding fit, H7/k6 slight transition (taps in by hand/hammer), H7/s6 press fit. Example Ø 10: H7 = +15/0 µm on the hole, g6 = −5/−14 µm on the shaft → guaranteed clearance 5–29 µm.
Put the hole tolerance on the hole dimension and the shaft tolerance on the shaft dimension — never both on one number. For one-off prototypes, H7 with reaming is the cheapest precise hole; below H7 you enter grinding territory.
When a size tolerance is not enough: flat faces, perpendicular holes, true position — the feature control frame.
When a size tolerance is not enough: flat faces, perpendicular holes, true position — the feature control frame.
Size tolerances control how big; geometrical tolerances control form, orientation, location and run-out. ISO 1101 defines the symbol frames (feature control frames) you see on professional drawings; ISO 5459 defines the datums they refer to. ASME Y14.5 is the US twin — the two are ~95% aligned.
A feature control frame reads left to right: symbol | tolerance zone | datum references. Flatness ▭: surface between two planes t apart. Perpendicularity ⊥ t|A: axis or face within a zone 90° to datum A. Position ◎ Øt|A|B|C: axis inside a cylinder Øt at the theoretically exact (boxed) location. Circular run-out ↗: surface variation per revolution.
Datum first: pick functional surfaces as A, B, C before tolerancing anything. GD&T controls only what function demands — every frame has a cost in manufacturing and inspection.
First vs third angle, the projection symbol, isometric vs axonometric views, and why your sheet is A4/A3.
First vs third angle, the projection symbol, isometric vs axonometric views, and why your sheet is A4/A3.
A flat drawing must say from which side each view was taken. Europe works mostly in first-angle projection, North America in third-angle. ISO 128-30 requires the projection symbol (truncated cone with circles) on every sheet where confusion is possible. Sheets themselves follow ISO 216: A4 = 210 × 297 mm, A3 = 420 × 297 mm.
Third angle = the view is placed on the side it is seen from; symbol: trapezoid left, two circles right. First angle mirrors that. An isometric view foreshortens all three axes equally at 120° apart: measuring along any axis reads true at scale 1:1 — exactly the (1,1,1) projection the app draws.
One sheet = one format, title block bottom right (ISO 7200). For reading: always check the projection symbol before assigning left/right views — it is the cheapest mistake on imported drawings.
The minimum data every workshop expects in the corner: title, scale, material, projection, tolerances, date.
The minimum data every workshop expects in the corner: title, scale, material, projection, tolerances, date.
ISO 7200 fixes the title-block content and its position (bottom right). Essential fields: title/identification, scale, units, projection method, general-tolerance reference, owner, date. ISO 5455 lists the preferred scales: 1:1, 1:2, 1:5, 1:10, 2:1, 5:1, 10:1 …
A drawing must scale within the preferred ladder — "1:1.37" is not a scale, it is a screenshot. Natural scale 1:1 whenever the part fits; enlarge small details on a detail view (II) instead of shrinking the sheet. Every revision bumps an index letter with a one-line change description.
Never scale a dimension off the paper: the number rules, the print is the packaging. Keep the general tolerance + projection symbol inside the block, standard notes outside it.
Current is charge in motion, voltage is the energy per charge package. Every circuit question is about these two numbers.
Current is charge in motion, voltage is the energy per charge package. Every circuit question is about these two numbers.
Electric charge Q (coulomb) is the conserved stuff inside wires. Current I (ampere) counts how much charge passes a point per second. Voltage V (volt) is how much energy each unit of charge carries between two points. Water analogy: voltage = pressure, current = flow.
I = Q / t: one ampere = one coulomb per second. V = W / Q: one volt = one joule per coulomb. A battery does not “fill the wire with electricity” — the charge is already in the copper; the source only pumps it around the loop, doing work in the loads.
Get fluent in prefixes early: mA (×0.001) rules small electronics, µA appears in sensors, kA in fault studies. The Circuit Lab reads engineering notation everywhere: 2k2, 10m, 4u7.
One equation ties voltage, current and resistance together: the most-used formula in all of electrical work.
One equation ties voltage, current and resistance together: the most-used formula in all of electrical work.
Ohm’s law states that the current through a resistor is proportional to the voltage across it: V = I · R. Given any two of V, I, R you get the third. It is empirical, but metallic conductors follow it astonishingly well at constant temperature.
The “VIR triangle” mnemonic: cover the one you want; V = I·R, I = V/R, R = V/I. Deeper reading: resistance is the ratio V/I at the operating point — for a diode the ratio changes with current, which is why we solve non-linear parts iteratively.
Sanity-check every hand calculation against magnitudes: 12 V on car-scale loads gives amps; 5 V on kilo-ohms gives milliamps; 230 V mains on anything small is a fire. The Lab’s units always display in engineering notation.
Power tells you how fast energy moves. In a resistor every watt becomes heat — that is why parts have ratings.
Power tells you how fast energy moves. In a resistor every watt becomes heat — that is why parts have ratings.
Power P (watt) = voltage × current for any two-terminal part. With Ohm’s law: P = I²R = V²/R. Energy is power × time — the kilowatt-hour your meter bills.
A resistor dissipates P = V²/R: doubling voltage quadruples the heat. Sources have power too: a battery delivering current out of its + terminal supplies power; the Lab shows positive P on sources when they deliver. Lamp brightness in the simulator is proportional to real dissipated power, so over-volting glows then burns.
Never run a part at its nameplate: derate resistors to 50–70 % of rated power for reliability. A “quarter-watt” resistor across 12 V must be at least 576 Ω.
The sum of currents into any junction is zero. Charge cannot pile up — so what flows in must flow out.
The sum of currents into any junction is zero. Charge cannot pile up — so what flows in must flow out.
KCL is charge conservation applied to a node: ΣIin = ΣIout. It holds at every instant, in every circuit, macroscopic to microscopic. It is half of what the simulator solves.
With two parallel resistors the supply current splits in inverse proportion to the resistances: the easier branch carries more. Current divider: I1 = Itot · R2/(R1+R2). The Lab’s node collector literally writes one KCL row per net into the matrix (modified nodal analysis).
Signs: decide “into the node = positive” and stick to it; negatives then mean “actually out”. Sloppy sign discipline, not Kirchhoff, is what ruins exam answers.
Around any closed loop the voltage rises and drops sum to zero. Energy per charge is a potential — like altitude.
Around any closed loop the voltage rises and drops sum to zero. Energy per charge is a potential — like altitude.
KVL is energy conservation: take a unit of charge around any closed path, and the energy gained in sources equals the energy spent in loads. ΣV = 0 around every loop, always.
In a series chain the supply voltage is divided among the resistors in direct proportion to their values: Vi = Vs · Ri/ΣR. Internal resistance hides inside the source symbol as an invisible series resistor — KVL still holds when you include it.
Pick a walk direction (clockwise), count rises as + and drops as −, and never switch midway. With two sources in a loop, KVL decides whether they help or fight.
Series adds resistances; parallel adds conductances. Recognising the topology is 90 % of simplifying a network.
Series adds resistances; parallel adds conductances. Recognising the topology is 90 % of simplifying a network.
Parts in series share one current; equivalent R = R1 + R2 + … Parts in parallel share one voltage; 1/Req = 1/R1 + 1/R2 + … Two equal resistors in parallel give exactly half.
The product-over-sum shortcut for two: Req = R1R2/(R1+R2). Networks that are neither pure series nor pure parallel (bridges!) need Kirchhoff directly — the simulator never simplifies, it solves everything numerically.
Parallel battery packs share current only when their voltages match — mismatched cells in parallel is a classic fire starter. Series strings add voltage; parallel strings add capacity.
Two resistors make any lower voltage you want — until you connect something. Then the bottom resistor is no longer alone.
Two resistors make any lower voltage you want — until you connect something. Then the bottom resistor is no longer alone.
Two series resistors across a supply present Vout = Vs · R2/(R1+R2) at their junction — the most common two-part circuit in electronics. But that formula assumes no current leaves the node.
Any load RL sits in parallel with R2: the lower leg becomes R2||RL, and the output sags. Design rule: make the divider current at least 10× the load current, so sag stays under ~10 %. That is exactly why dividers are for signals, not for powering things.
For sensor bias (LDR/NTC!) pick the fixed resistor near the sensor’s mid-range resistance for best sensitivity. Never drop 24 V to 5 V for a MCU board with a divider — use a regulator.
Every source hides a small series resistor. Under load the terminal voltage falls: V = E − I·r.
Every source hides a small series resistor. Under load the terminal voltage falls: V = E − I·r.
An ideal voltage source holds its voltage forever; a physical one cannot. The difference is modelled as an internal resistance r in series: terminal voltage V = E − I·r under load.
Measure open-circuit voltage (I = 0 → V = E), then a loaded voltage: r = (E − Vload)/I. A fresh AA alkaline ≈ 0.15 Ω; a tired one several ohms. The battery symbol in the Lab carries cells × 1.5 V and Rint per cell — the sag is computed honestly, not faked.
Cranking a car: 9 V on a “12 V” battery at 200 A means r ≈ 15 mΩ. LEDs and laser diodes hate stiff sources — a tiny r is why a coin cell limits current “for free”.
Direct current analysis assumes the last switch-dictated state has settled. Capacitors are open, inductors short — on purpose.
Direct current analysis assumes the last switch-dictated state has settled. Capacitors are open, inductors short — on purpose.
DC steady state = the circuit long after the last change. Then a capacitor conducts nothing (open) and an inductor drops nothing (short). AC analysis instead studies sine-wave behaviour with impedance Z.
During the first instants after switching, capacitors gulp current and inductors fight changes — the transient region. The Circuit Lab deliberately solves DC steady state: every answer there is exact. AC impedance (|ZC| = 1/(ωC), |ZL| = ωL) and RC/RL charge curves are a documented roadmap item, listed on the app’s help sheet.
Debugging practice: measure DC first (is the operating point right?) before blaming signals. 80 % of “analog ghost” faults are wrong DC bias.
A voltage is only ever <em>between two points</em>. One chosen node, called ground, is where the voltmeter’s black lead lives.
A voltage is only ever between two points. One chosen node, called ground, is where the voltmeter’s black lead lives.
Ground (reference node) is the circuit’s agreed 0 V. All node voltages are reported against it. Chassis, earth and signal ground are different ideas in power engineering; in a schematic solver there is exactly one: the reference.
All ground symbols in the Lab are one net (standard EDA behaviour). No ground = no reference: the matrix has no fixed zero and the solver politely refuses with a guidance message instead of a number salad. Parts completely disconnected from ground form a floating island — their voltage is physically undefined until tied through a DC path.
Real-work habit: connect meter grounds first, mind that cheap USB scopes share PC earth, and never clip one scope’s ground across a live bridge rectifier.
An ammeter adds resistance in series; a voltmeter steals current in parallel. Honourable instruments keep both effects tiny.
An ammeter adds resistance in series; a voltmeter steals current in parallel. Honourable instruments keep both effects tiny.
Measuring changes the thing measured. The ammeter inserts a small “shunt” resistance (≈10 mΩ) into the loop; the voltmeter puts a large resistance (≈10 MΩ) across the points it probes.
Insertion error scales with circuit impedance: across a 10 MΩ divider a 10 MΩ meter halves the reading (divider-load lesson again!). Ammeter burden voltage Vburden = I·Rshunt matters in millivolt circuits. The Lab’s meters carry their true internal resistances, so hovering a measured node exposes the effect instead of hiding it.
Measure current only where a break exists; never parallel an ammeter across a source — its 10 mΩ is a short circuit with a display.
C = Q/V. At switch-on a capacitor is a brief short; charged, it is an open door. That is why the DC solver treats it as open.
C = Q/V. At switch-on a capacitor is a brief short; charged, it is an open door. That is why the DC solver treats it as open.
A capacitor (farad) stores energy as separated charge on two plates: Q = C·V. Current flows only while voltage changes: I = C·dV/dt.
Connected to DC through a resistor, the voltage follows V(t) = Vs(1 − e−t/RC) — 63 % after one time constant τ = RC, ≈99 % after 5τ. In the Lab’s DC steady state, t → ∞: no current, full voltage across the plates.
Electrolytics know polarity — reverse them and they vent. Decoupling caps belong AT the IC pins. Never trust a big charged capacitor: discharge through a resistor, not a screwdriver (usually).
V = L · dI/dt. An inductor only cares about CHANGES of current; steady DC glides through with just wire resistance.
V = L · dI/dt. An inductor only cares about CHANGES of current; steady DC glides through with just wire resistance.
An inductor (henry) opposes current changes by building voltage: V = L·dI/dt. Once current is steady, dI/dt = 0 and the coil is just wire — a short at DC.
The current after switching follows I(t) = (V/R)(1 − e−tR/L) with τ = L/R. Kill the current abruptly and dI/dt explodes — the inductor answers with a voltage spike (flyback), which is why relay coils wear diodes.
Relays and solenoid valves in 24 V panels always get a flyback diode or an RC snubber — one 100 V spike retirement party per week otherwise.
The workhorse: a controlled obstacle that turns voltage into current — and current into heat, precisely.
The workhorse: a controlled obstacle that turns voltage into current — and current into heat, precisely.
The IEC symbol is a rectangle with two leads (the US zigzag is ANSI Y32 — the Lab follows IEC 60617). Value in ohms, tolerance band, power rating in watts.
E-series values repeat per decade: E12 gives 10, 12, 15, 18, 22, 27, 33, 39, 47, 56, 68, 82. A 5 % (gold) E24 part covers the shelf; 1 % metal film costs pennies more and is today’s default.
Pick by duty: signal = any film resistor; power = wirewound with margin; precision = 0.1 % metal film. Watch temperature coefficient only past ppm-level ambitions.
4k7 and 2M2 in its dialog to feel the notation.Three pins: two ends of a resistive track plus a wiper. One knob, any ratio between 0 and 100 % of its value.
Three pins: two ends of a resistive track plus a wiper. One knob, any ratio between 0 and 100 % of its value.
A potentiometer is a resistor with a movable tap. As a divider (all three pins) it = variable voltage source-ish reference; with wiper tied to one end it is a plain 2-pin adjustable resistor (rheostat connection).
Wiper at position p % splits R into p·R and (1−p)·R. As a divider across 10 V: Vw = p · 10 V — load-independent only while nothing draws from the wiper.
Logarithmic (audio) vs linear taper for volume vs control. Wire rheostats fail-safe: wiper tied to the unused end, so a worn wiper can’t open the loop.
One obeys light, the other temperature. Both are plain resistors with a mood — read them through a divider.
One obeys light, the other temperature. Both are plain resistors with a mood — read them through a divider.
LDR (light-dependent resistor): mega-ohms in the dark, hundreds of ohms in bright light. NTC thermistor: resistance falls exponentially as temperature rises, R = R25·eB(1/T − 1/298).
Neither outputs a voltage by itself — you always form a divider with a fixed resistor around mid-range. The Lab offers discrete LDR light presets (dark 100 kΩ → bright 500 Ω) and honest Beta-model NTCs at your chosen temperature.
Match the fixed resistor to the sensor at the condition you care about most. NTCs self-heat above ~1 mA — that error masquerades as +2 °C.
Two plates and an insulator. Real units: pF, nF, µF — one full farad is a physics demo, not a shelf part.
Two plates and an insulator. Real units: pF, nF, µF — one full farad is a physics demo, not a shelf part.
IEC symbol: two parallel plates (curved plate marks the negative of an electrolytic). Families: ceramic (pF–µF, tiny), film (stable), electrolytic (µF–mF, polarized), supercap (farads, slow).
C = ε0εrA/d: value grows with area and dielectric constant, shrinks with plate distance. Voltage rating is the dielectric’s breaking point — exceed it and the part becomes an ex-capacitor.
Electrolytic = one-way only (stripe = minus). Ceramic 100 nF at every IC supply pin. Class-2 ceramics (X7R) lose up to 80 % of label capacity at rated DC bias — design with 3× margin there.
Wire wound around (maybe) a core. Values: nH to H. IEC draws the half-circle humps; the Lab shorts it at DC.
Wire wound around (maybe) a core. Values: nH to H. IEC draws the half-circle humps; the Lab shorts it at DC.
IEC symbol: a series of semicircular humps (core = added parallel bar). Value in henry, plus maximum current (saturation) and DC resistance in the fine print.
L = μN²A/ℓ — turns count squared. Ferrite cores multiply μ but saturate: past Isat the inductance collapses and the part turns into a warm resistor.
Power inductors: choose Isat above peak current with 30 % margin. For EMI bead vs inductor vs common-mode choke — similar symbols, very different jobs.
A perfect short or a perfect open, changed by a finger. SPST, SPDT, momentary, maintained.
A perfect short or a perfect open, changed by a finger. SPST, SPDT, momentary, maintained.
SPST = one path open/closed. SPDT (changeover) = one input, two outputs. Pushbuttons are momentary (spring return), toggle switches are maintained. IEC shows the pivot dot and the lever.
Real contacts bounce for 1–10 ms — microseconds matter to a 16 MHz MCU, so firmware debounces or an RC does. Contact ratings: voltage AND current AND AC/DC (DC arcs don’t self-extinguish at zero-crossing).
For logic inputs use a pull resistor, let the switch short to ground (active-low): noise-immune and cheaper than active-high. In the Lab click a switch during simulation to toggle it live.
Electromagnet plus changeover contacts: galvanic isolation in its oldest, most repairable form.
Electromagnet plus changeover contacts: galvanic isolation in its oldest, most repairable form.
Coil (rated 5/12/24 V) pulls an armature that moves the common contact from NC (normally-closed) to NO (normally-open). Coil current ≈ tens of mA; contacts may switch amperes — control and load worlds stay electrically separate.
Pickup needs ≈75 % of nominal coil current; it releases only below ≈30 % — hysteresis you can observe. The Lab models exactly this: coil I = V/Rcoil, thresholds, then the contact legs change state and the solver re-balances the whole circuit.
Always fit the flyback diode across a DC coil (1N4148 across it, band to +). Contact life is specified at load — 250 VAC ratings collapse for DC motor loads.
Current flows anode→cathode once the toll is paid; reverse direction stays shut until breakdown.
Current flows anode→cathode once the toll is paid; reverse direction stays shut until breakdown.
IEC symbol: triangle touching a bar — the bar is the cathode. Silicon starts really conducting around 0.6–0.7 V; Schottky around 0.3 V; germanium ~0.2 V.
The Shockley equation I = Is(eV/nVT − 1) is exponential: 60 mV more voltage ≈ 10× more current. The simulator Newton-iterates exactly this law (1N4148: Is = 2.5 nA, n = 1.9), so your forward drop reads 0.6–0.75 V depending on current, honestly temperature-shaped at 300 K.
Check continuous current, peak reverse voltage (PIV) and speed. 1N4148 = 200 mA/100 V fast signal; 1N4007 = 1 A/1000 V slow mains.
Red pays ~2.0 V, blue ~3.0 V, white ~3.2 V. Photons out, but only while current stays below ~20–30 mA.
Red pays ~2.0 V, blue ~3.0 V, white ~3.2 V. Photons out, but only while current stays below ~20–30 mA.
A light-emitting diode follows the diode law with a bigger forward voltage set by the semiconductor bandgap: infrared < red < amber < green < blue < UV < white (blue + phosphor).
Because I(V) is exponential, LEDs are fed through a series resistor: R = (Vs − Vf)/Itarget. 5 V, red, 10 mA → (5−2)/0.01 = 300 Ω → shelf value 330 Ω. The Lab simulates photon output as I/20 mA brightness and kills the part over 30 mA — visibly.
Never parallel LEDs on one resistor — Vf mismatch makes one hog the current. One per resistor, or a series string on a higher voltage.
Run it backwards on purpose: past V<sub>z</sub> it holds its voltage and dumps the surplus current.
Run it backwards on purpose: past Vz it holds its voltage and dumps the surplus current.
A zener conducts like a normal diode forward, but reverse-biased it enters a controlled breakdown at Vz (2.4–200 V ecosystem). Cathode to + for regulation.
With source, series resistor and load, the zener takes whatever current the load leaves: Iz = (Vs−Vz)/Rs − Iload. Rs sized so Iz stays between knee (≈1 mA) and Pmax/Vz across all load corners. The Lab’s NR loop models both Shockley legs plus the breakdown knee — reads 5.22 V for a “5.1” part at real current.
Zener tolerance ≈5 % and the knee is soft — for anything a microcontroller’s Vref touches, use a shunt reference (TL431) instead. Zeners love protection duty and coarse rails.
A small base current I<sub>B</sub> licenses β·I<sub>B</sub> of collector current — like a relay without moving parts, but analogue.
A small base current IB licenses β·IB of collector current — like a relay without moving parts, but analogue.
NPN: current flows collector→emitter when base is pulled ~0.7 V above emitter. PNP is the mirror, switched by pulling base below its emitter. The Lab ships them as IEC reference symbols (documentation-only for now — amplifiers are on the roadmap, and the help sheet says so).
Three regions: cutoff (off), active (IC = β·IB, β ≈ 100–300), saturation (switch: VCE(sat) ≈ 0.1–0.2 V). Switching design uses forced beta of 10–20 to guarantee saturation.
For a 100 mA relay coil: IB = 10 mA, from 3.3 V logic → RB ≈ 260 Ω. Add the flyback diode anyway (the coil lesson refuses to die).
A calibrated weak link: I²R heat in a thin wire melts it when current exceeds the rating long enough.
A calibrated weak link: I²R heat in a thin wire melts it when current exceeds the rating long enough.
IEC symbol: rectangle with a line through it. Ratings: current, voltage, speed (FF/F/T/TT), breaking capacity (kA it can interrupt without exploding).
Melting integral I²t: brief 2× overload survives, sustained 1.2× kills slow-blow. A fuse protects against fire, not against electronics damage — semiconductors always sacrifice themselves first to save the fuse.
Panel rule: fuse at 125 % of max working current, wire sized above the fuse, breaking capacity above the available fault current. The Lab blows its fuses honestly after a moment of overload — replace via the part dialog.
Rating plate V and P define the hot resistance R = V²/P; the Lab lights it proportional to real watts.
Rating plate V and P define the hot resistance R = V²/P; the Lab lights it proportional to real watts.
The lamp symbol (circle with crossed filaments, IEC) is a non-linear resistor in life — cold filament ≈ Rhot/10, which explains switch-on inrush.
Run a 12 V / 6 W lamp on 6 V and it glows at ~1.5 W (P = V²/R with R = 24 Ω hot — quarter the voltage, quarter the power). Overvolt it and P climbs past rating until the filament loses: the Lab shows the warning, then burns it out.
Lamps make lovely teaching loads: cheap, honest, self-indicating. For panel use today an LED with the right resistor wins on life by 50×.
Spinning coils in a field generate a counter-voltage; that is why motors draw stall current only at standstill.
Spinning coils in a field generate a counter-voltage; that is why motors draw stall current only at standstill.
A permanent-magnet DC motor: V = I·Ra + E, where the back-EMF E grows with speed (E = kω). Stall: E = 0 and I = V/Ra — the huge start current.
Torque = k·I; speed stabilises where E ≈ V − IR. The Lab’s steady-state model takes V and armature R, reports the stall-point current and animates the rotor proportionally to V/Vrated — a window into the physics, not a full transient drive model.
Never run a small motor stalled: 8 Ω armature on 12 V = 1.5 A of pure heat. Drivers (L293D in the IC library!) exist because logic pins cannot feed amperes.
Rule 1: inputs draw nothing. Rule 2: with negative feedback, the output forces V<sub>+</sub> = V<sub>−</sub> — until it hits a rail.
Rule 1: inputs draw nothing. Rule 2: with negative feedback, the output forces V+ = V− — until it hits a rail.
The op-amp amplifies the difference of its inputs by open-loop gain A ≈ 200 000. With feedback the two golden rules let you compute any classic stage in two lines. The LM741 symbol in the Lab carries +/− inputs, output and both supply pins.
Voltage follower: out = in. Inverting: Vo = −Rf/Rin·Vi with the −IN node pinned at “virtual ground”. Comparator (no feedback): output slams to a rail depending on the input sign — the Lab’s fixed-point pass clips at rail−0.8 V, so saturation is simulated, not hand-waved.
741 is museum-grade: input common-mode range excludes both rails, output swings nowhere near them. Modern default: rail-to-rail MCP6021/TLV9062. Always tie unused inputs and always supply both rails symmetrically in lessons.
Two sources, three resistors, one shared node: write KCL once, KVL twice, and the matrix does the arithmetic.
Two sources, three resistors, one shared node: write KCL once, KVL twice, and the matrix does the arithmetic.
The classic exam circuit: 12 V through 4 Ω and 9 V through 6 Ω feeding a shared node loaded by 10 Ω. Everything hangs on the single unknown node voltage M.
KCL at M with Ohm’s law per branch: (12−M)/4 + (9−M)/6 = M/10. Solve: M = 180/20.67 ≈ 8.71 V; then the 12 V source delivers 823 mA while the 9 V one absorbs — the weaker source is being charged by the stronger. The Lab settles the identical numbers to five digits.
Method beats memory: label the unknown node, express every branch current as (Vfar − M)/R, sum to zero, solve. It scales to any node count — that is all nodal analysis is.
Four resistors in a diamond; the middle meter reads zero exactly when the ratios match: R<sub>1</sub>/R<sub>2</sub> = R<sub>3</sub>/R<sub>4</sub>.
Four resistors in a diamond; the middle meter reads zero exactly when the ratios match: R1/R2 = R3/R4.
A bridge turns one unknown resistor into a precision measurement: three known arms + the unknown, balance the ratio, and the galvanometer between the midpoints nulls.
Balanced: both midpoints sit at half supply and floating between them reads 0.000 V regardless of supply value or meter quality — the magic of ratiometric null. Unbalanced by 1 %: the center voltage moves ≈ Vs/4 × 1 % — linear, sensitive, and the whole strain-gauge industry lives on it.
Bridge sensors want an instrumentation amplifier (op-amp lesson) behind the center nodes and stable excitation. Four wires (sense pair) when the bridge is far from the electronics.
Five steps: know V<sub>s</sub>, pick the colour’s V<sub>f</sub>, choose brightness current, subtract, divide, take the next shelf value.
Five steps: know Vs, pick the colour’s Vf, choose brightness current, subtract, divide, take the next shelf value.
The everyday procedure: R = (Vs − Vf)/I. 12 V panel, green LED at 8 mA: (12−2.2)/0.008 = 1225 Ω → 1.2 kΩ on the shelf, power P = I²R ≈ 77 mW → a 0.25 W part with 3× margin.
The resistor exists because an LED on a stiff source is an exponential waiting to die: a 5 % voltage rise is +100 % current. Series strings: subtract ALL the Vf before dividing — three whites need >9.6 V of budget, so 12 V works, 9 V never will.
If the remaining “R-voltage budget” is under ~20 % of supply, Vf tolerances own your brightness — step up the supply or drop a diode from the string.
Charge a capacitor through a resistor and compare against a threshold: that single idea is half of all timers.
Charge a capacitor through a resistor and compare against a threshold: that single idea is half of all timers.
The NE555 (in the IC library: pinout, packages, the works) charges C through R until 2/3 VCC, then discharges to 1/3 VCC — astable period T ≈ 0.693·(RA+2RB)·C.
Thresholds at thirds come from a 3×5 kΩ divider inside the chip (hence “555”). The Circuit Lab solves DC steady state now, so the on-canvas 555 block is an honest IEC reference symbol with its true pinout — the timing curve itself is on the documented AC/transient roadmap. You can still verify the threshold voltages and bias currents with the op-amp samples: a comparator is what each half of a 555 internally is.
Astable duty must exceed 50 % by construction; add a diode across RB for symmetric blink. Supply decoupling 100 nF at pin 8→1 is mandatory — 555s crowbar rails during transitions.
Source, series resistor, zener to ground: the cheapest regulator there is — if you respect all four corners.
Source, series resistor, zener to ground: the cheapest regulator there is — if you respect all four corners.
Shunt regulator design: Rs must pass max load current at min input voltage while never exceeding zener power at max input and min load. Two inequalities, four corners, one resistor.
Rs ≤ (Vs,min−Vz)/(Iload,max+Iz,knee) and Pz = Vz·((Vs,max−Vz)/Rs) ≤ rating. Efficiency is terrible by construction (the zener eats the margin), which is why it feeds references and tens of milliamps, not power loads.
Beyond ~100 mA, shunt → LDO/7805 (both pinouts in the IC library) or a buck. Below that, a zener corner-checked like this runs for decades.
Two contactors, one motor: NC of each sits in the other’s coil circuit so FORWARD and REVERSE can never be true together.
Two contactors, one motor: NC of each sits in the other’s coil circuit so FORWARD and REVERSE can never be true together.
Classic motor control: direction relays K1/K2 swap armature polarity. Wire the NC contact of K2 in series with K1’s coil and vice versa — electrical interlock; add mechanical interlock bars for power contactors.
Why both contact & coil sides: the coils draw ~30 mA (logic-safe), the contacts carry armature amperes (arc country) — the relay lesson’s galvanic isolation exploited to full effect. A simultaneous press simply re-fights the interlock, not the fuse.
Add a few-ms dead time for DC motors (armature inertia keeps back-EMF alive after contact break) — snubbers/varistors across contacts double their life.
A structure holds still, a mechanism moves on purpose, a machine does work. Everything else is detail.
A structure holds still, a mechanism moves on purpose, a machine does work. Everything else is detail.
A structure is an assembly whose parts do not move relative to each other (a bridge, a bracket). A mechanism is an assembly of rigid links and joints that transmits a definite motion. A machine is a mechanism (or several) that also transmits energy and does useful work — a press, an engine, an excavator.
Links (bodies) are the rigid players; joints (kinematic pairs) are the rules between them. Add a motor and a load to a mechanism and it becomes a machine; lock one link as the frame (ground) and every motion is measured against it. The Mechanism Studio calls the frame ground and shows it as the hatched base.
Before drawing anything, say the sentence: «with one input motion, I want this output motion». If the output is a force, not a motion, you might want a structure or a press — not a linkage.
F = 3(n−1) − 2j₁ − j₂ — the three-line audit that tells you whether your sketch can move at all.
F = 3(n−1) − 2j₁ − j₂ — the three-line audit that tells you whether your sketch can move at all.
The mobility (degrees of freedom, DOF) is the number of independent inputs a mechanism needs before every pose is determined. One DOF: one motor pins everything down. Zero: a structure (with luck). Negative: over-constrained — it only assembles because tolerances bend.
In the plane, a free body has 3 DOF (x, y, angle). Ground one (−3). Every one-DOF joint (pin, slider) removes 2; every higher pair (cam, gear contact) removes 1: F = 3·(n−1) − 2·j₁ − j₂. A four-bar has n = 4, j₁ = 4 → F = 1. A five-bar has F = 2 — two motors or nothing.
Count F before you drive. A motor on a 2-DOF sketch animates a motion, not the motion. Spring-loaded gravity settles 1-DOF mechanisms at odd poses but does not reduce F — the extra freedom is just quiet.
Surface contact (pin, slider) or line/point contact (cam, gear)? The choice decides wear, friction and life.
Surface contact (pin, slider) or line/point contact (cam, gear)? The choice decides wear, friction and life.
A lower pair connects two links over a surface — revolute (pin), prismatic (slider), cylindrical, helical (screw). A higher pair connects along a line or at a point — cam-follower, gear teeth, ball bearings, wheels on rails.
Lower pairs spread load over an area: gentle pressures, easy lubrication, forgiving seals — the joints of excavators and linkages. Higher pairs concentrate load: tiny contact patch, huge Hertz stress, precise geometry required. That is why gear teeth are case-hardened and cams are polished.
Default to lower pairs for force paths; accept higher pairs where the motion law demands them (valve timing, indexing). Put the backlash-sensitive higher pair where precision matters least in the chain.
Ground, crank, coupler, rocker — four rigid links beat half of machine design.
Ground, crank, coupler, rocker — four rigid links beat half of machine design.
Four links in a closed loop pinned by four revolute joints. One link is fixed (the frame), the link hinged to it that can rotate fully is the crank, the one that can only swing is the rocker, and the floating link between them is the coupler.
Drive the crank and the rocker answers with a swinging motion whose shape you design through the four lengths. But the coupler is the quiet genius: every point on it traces a different coupler curve — ovals, figure-eights, near-straight runs. That curve, not the rocker, is what walks machines and steers wheels.
Name lengths r₁ (ground), r₂ (crank), r₃ (coupler), r₄ (rocker). Ratios decide character, not absolute size: scale the whole thing ×10 and the motion is geometrically identical — forces are not.
Shortest + longest vs the other two — one inequality decides the whole personality of a four-bar.
Shortest + longest vs the other two — one inequality decides the whole personality of a four-bar.
Sort the four link lengths: shortest s, longest l, others p and q. Grashof’s law: if s + l ≤ p + q, at least one link can make a full revolution; the class then depends on which link is grounded.
Ground the link next to s → crank-rocker (s spins, opposite side rocks). Ground s itself → double-crank (both sides spin, the drag-link). Ground the link opposite s → double-rocker. If s + l > p + q, no full revolution exists — every inversion is a double-rocker, and at s + l = p + q the chain hits change points where it can flip branch.
Keep a margin: s + l ≪ p + q assembles easily and tolerates wear. Designs at the edge go sloppy or jam in the field as pins wear the millimetre you had no margin for.
µ is the angle between coupler and rocker — near 90° the force flows, near 0° the linkage folds.
µ is the angle between coupler and rocker — near 90° the force flows, near 0° the linkage folds.
The transmission angle µ is the acute angle between the coupler and the output link. It is the force-quality gauge of a four-bar: at µ = 90° the coupler pushes the rocker squarely; at µ → 0° the push aims straight through the rocker’s own pivot and produces no torque.
Rule of thumb: keep 40° ≤ µ ≤ 140° in the working range. µ wanders sinusoidally as the crank turns, with worst values when crank and ground align. At the dead (toggle) points µ passes through 0° or 180° — great for clamps (infinite force multiplication), fatal as an operating point for rotating machines, which is why engines need a flywheel to coast through.
Check µ at both stroke extremes, not at mid-stroke. If µ dips under 40°, lengthen the coupler, shorten the crank, or move the output pivot — in that order of cheapest-first.
Push two nearly-straight links and the end force is F/(2·sin θ) — at θ → 0 it out-pushes any cylinder.
Push two nearly-straight links and the end force is F/(2·sin θ) — at θ → 0 it out-pushes any cylinder.
When two links are nearly collinear (the toggle position), a small force at their joint creates enormous forces at the far ends. Exactly: an elbow pushed with Q at the knee, arms at angle θ to straight, presses with Q / (2·sin θ) at each end.
At θ = 5° the multiplier is about 5.7×; at 1° it is 29×. In the limit the geometry demands infinity and reality answers with pin shear and link buckling instead. The same geometry in reverse is the over-centre lock: past the dead point, reaction forces re-seat the joint instead of opening it — clamps and deadbolts hold with zero spring pressure.
Use toggle force for clamping, not for travel: the last millimetres give the force but almost no motion. Check pin shear and link buckling at the computed end force — those are the real limits.
MA and VR are the same coin’s two faces: what you win in force you pay in travel — energy keeps the books.
MA and VR are the same coin’s two faces: what you win in force you pay in travel — energy keeps the books.
Velocity ratio VR = input speed / output speed (by geometry). Mechanical advantage MA = output force / input force. With no losses, P_in = P_out forces MA = VR exactly — every lever, gearbox, pulley and toggle obeys it, which is why none of them is free energy.
A 2:1 belt drive halves speed and (ideally) doubles torque. A block and tackle with 4 supporting segments quarters the rope force and quadruples the rope you must pull. With friction, MA = η·VR: the efficiency quietly taxes the force side, never the geometry side.
Decide which side you are buying: force (pay with travel — jacks, presses) or speed (pay with force — fans, spindles). Read the real ratio from the machine, not the brochure: block wears, belts slip, gears have backlash.
Every moving link is, for one instant, a pure rotation about some point — find that point and velocities read off like a lever.
Every moving link is, for one instant, a pure rotation about some point — find that point and velocities read off like a lever.
The instantaneous centre (IC) of a moving link is the point of zero velocity at this instant. A four-bar’s coupler does not translate or spin about its pins — it pivot-paints around its IC, which slides along as the pose changes.
Kennedy’s theorem: for any three links in relative motion, their three ICs lie on one straight line. For a four-bar this places the coupler’s IC on the ground-link line through the two fixed pivots AND on the line through the two moving pins — the intersection of the two side-link lines. Velocity ratio then falls out as a lever ratio about that point.
n links have n(n−1)/2 ICs, and (n−1) of them are instant pins or sliders — visible by inspection. Use ICs for sanity checks: if the velocity you computed implies a different IC than geometry does, the computation is wrong.
Every point of a floating link writes its own closed curve — ovals, eights, near-lines. Machine design libraries are atlases of them.
Every point of a floating link writes its own closed curve — ovals, eights, near-lines. Machine design libraries are atlases of them.
A point fixed to the coupler of a four-bar traces a coupler curve — an algebraic curve of up to sixth order, always closed. Change the point or any link ratio and the curve morphs continuously: new loops, new cusps, new flat runs.
Cusps appear where the curve’s point sits exactly on the coupler’s instant centre (the point circles a sharp corner); approximate straight segments come from points running on near-circular arcs with far-away centres. Classic design catalogues (Hrones & Nelson) are just atlases of these curves with the producing geometry printed beside each.
Harvest curves for motion tasks: flat segments for straight-line guides, a cusp-to-arc swing for a walking foot, a D-shaped loop with a flat bottom for a transport dwell. One chosen loop = one built linkage.
Same constant motor speed, slow cutting stroke, fast idle return — the asymmetry is pure crank geometry.
Same constant motor speed, slow cutting stroke, fast idle return — the asymmetry is pure crank geometry.
A quick-return drive turns uniform input rotation into a reciprocating stroke whose forward half takes longer than its return. Shapers and slotters cut only on the slow stroke and waste as little time as possible coming back.
The mechanism’s crank sweeps angle α during the working stroke and 360°−α during return, where α is set by the geometry (for the Whitworth: α is the arc of the crank pin on the slotted-lever side). The time ratio TR = α/(360°−α) with constant drive speed is also the stroke-time ratio. TR = 2 means α = 240°: two-thirds cutting, one-third flying home.
Higher TR = more violent return acceleration and inertia forces. Keep TR ≈ 1.5–2 for machine tools; the saved cycle time no longer pays for the bearing loads beyond that.
Before machine tools could scrape a true rail, linkages <em>drew</em> straight lines — Watt approximately, Peaucellier exactly.
Before machine tools could scrape a true rail, linkages drew straight lines — Watt approximately, Peaucellier exactly.
A straight-line mechanism guides a point along a (near-)straight path using only pin joints. Watt’s 1784 linkage gave the steam piston rod an honest line decades before planers existed; Peaucellier’s 1864 six-bar is exact, built on circle inversion.
Watt I: the coupler mid-region of a well-proportioned four-bar swaps its wander direction, cancelling curvature over a useful span. The Chebyshev lambda pushes the same idea into a longer, flatter run. Peaucellier: if OP·OQ = const for two points of a linkage, then when P circles, Q runs mathematically straight.
Approximate linkages tolerate wear and are compact; exact ones pack more links and small errors reappear as flex. For a long straight stroke today, compare honestly against a linear rail — linkages win where the guide itself must fold, walk or hang.
Cut the body free, replace every joint by its unknown reactions, put weight at the centre of mass — then count equations.
Cut the body free, replace every joint by its unknown reactions, put weight at the centre of mass — then count equations.
A free-body diagram (FBD) isolates one body and draws every force and moment the removed surroundings applied to it: contact forces, spring pulls, gravity at the centre of mass, nothing else. It is the single tool that converts “the machine pushes somewhere” into solvable equations.
In the plane each body yields three equations: ΣF_x = 0, ΣF_y = 0, ΣM = 0 (any pivot). A pin joint introduces two unknown reaction components; a slider one (normal to the slot); a two-force member (a rod loaded only at its ends) carries a single force along the rod — spotting those halves the algebra. Three-force bodies must be concurrent (or parallel): a gift for graphical checks.
Draw forces in their actual direction when known (rod along its axis, slot normal perpendicular), unknown ones as components and let the sign fall out. Never put a moment around a body the joint cannot transmit.
M = F·d and ΣM = 0 — two lines that size every lever, seesaw, brake pedal and crowbar you will ever meet.
M = F·d and ΣM = 0 — two lines that size every lever, seesaw, brake pedal and crowbar you will ever meet.
A moment is force × perpendicular lever arm: M = F·d, in N·mm or N·m. A body in equilibrium has the moments about any point cancel: ΣM = 0. Choose that point cleverly — through an unknown reaction — and it vanishes from the equation.
Worked example, the lab’s seesaw: a 30 kg kid at 400 mm left balances m₂ at +450 mm → m₂ = 30·400/450 ≈ 26.67 kg. The g cancels — moments of masses balance like moments of forces. Drag her 100 mm out of her mark (to 550 mm) and her 26.67 kg → 261.6 N of weight gains 0.1 m of lever: 261.6 × 0.1 ≈ 26 N·m of unbalanced moment, which the plank answers by slowly rotating.
Always balance moments about the pivot first; then the ΣF equations hand you the pivot’s own reaction for free. For hand-sized units: a 10 kg mass at 100 mm levers ≈ 9.8 N·mm of moment.
Statics is dynamics with the acceleration crossed out — and the first honest answer about any mechanism under load.
Statics is dynamics with the acceleration crossed out — and the first honest answer about any mechanism under load.
Static analysis asks: with all velocities zero and loads applied, what are the reactions? It is the foundation of structural sizing and of every fixture: if the statics is wrong the dynamics will only be more wrong.
Equilibrium needs ΣF = 0 and ΣM = 0 on every body. Too few constraints → the machine accelerates (a motor’s job). Too many → statically indeterminate: the reactions split according to stiffness, not geometry, which is why over-constrained doors sag and bind. The Studio’s ⚖ button settles the model by dynamic relaxation (a damped run to rest) and reports exactly the equilibrium it found.
Run statics at the worst pose, not the prettiest one: reactions peak where the geometry fights you (toggle-adjacent, full reach, heaviest lever arm). Compare two poses and you already have a fatigue story.
Coulomb resists with µ·N regardless of speed; viscous grows with speed; Real pins mix both plus a dry squeak of stiction.
Coulomb resists with µ·N regardless of speed; viscous grows with speed; Real pins mix both plus a dry squeak of stiction.
Coulomb (dry) friction resists sliding with a torque/force about µ·N, independent of speed, always opposing motion. Viscous friction resists like c·ω — proportional to speed, from oil films and dampers. Stiction is the same Coulomb story with a higher static µ before breakaway.
Consequences: Coulomb friction eats a fixed tax (pendulums die linearly) and can lock slow machinery entirely; viscous friction eats a percentage (pendulums die exponentially) and never quite stops anything. Designers choose: brakes and clutches are Coulomb devices on purpose, hinges want neither, dampers are bottled viscous friction sold as a product.
Choose lubrication to convert Coulomb into viscous wherever energy or precision matters. Keep dry friction where holding power is the product (clamps, torque limiters, parking brakes) — and never in a kinematic pair you intend to measure.
η = P_out/P_in. Drive from the wrong side and η ≤ 0 — the mechanism refuses to move at all. Sometimes that refusal is the product.
η = P_out/P_in. Drive from the wrong side and η ≤ 0 — the mechanism refuses to move at all. Sometimes that refusal is the product.
Efficiency η is power out over power in — the friction tax. A brilliant theorem hides inside: if reversing the drive (pushing from the output) makes η drop to ≤ 0, the transmission is self-locking: no load can back-drive it, however hard it pushes.
The screw thread is the showcase: forward efficiency ≈ tan λ / tan(λ+ρ), with lead angle λ and friction angle ρ = atan µ. Reversed: tan(λ−ρ)/tan λ — negative when λ < ρ. Hence a screw jack holds a car when your arms stop, and a worm gearbox holds its load when the motor dies. Ball screws roll ρ down and happily back-drive — that is why CNC axes need brakes when power fails.
Want holding without power? Size the friction side into self-lock (threads, worms, wedges). Want efficiency? Kill sliding friction (rolling, low helix), and then add a holding device for the failure you just removed.
I = Σ m·r² — mass far from the axis counts squared. That one exponent designs flywheels, wheels and propellers.
I = Σ m·r² — mass far from the axis counts squared. That one exponent designs flywheels, wheels and propellers.
The mass moment of inertia I measures resistance to angular acceleration: T = I·α. Because each mass element counts with its squared radius, material at the rim is worth far more than the same material at the hub.
Solid disc: I = ½·m·R². Thin rim: I ≈ m·R² — the same mass buys double. Scale law at constant material and thickness: m ∝ R², so I ∝ R⁴ — doubling the radius multiplies inertia sixteen-fold. The radius of gyration k with I = m·k² converts any shape into an equivalent rim for bookkeeping.
When you need rotational energy storage or smoothness, spend mass at the rim (flywheels, rings). When you need fast response, hollow the outside first (robot links, racing wheels): response time scales with I, and I answers mostly to the outer centimetres.
ΔE = I·ω²·Cs — a flywheel does not remove energy demand, it reschedules it inside one cycle.
ΔE = I·ω²·Cs — a flywheel does not remove energy demand, it reschedules it inside one cycle.
A flywheel stores rotational energy E = ½·I·ω² and returns it inside the cycle: it lends energy during the punch stroke and borrows it back during the idle return. The motor needs to supply only the average power — the peak is paid from the wheel.
The speed that breathing costs is set by the coefficient of speed fluctuation Cs = (ω_max − ω_min)/ω_mean. Energy bookkeeping: ΔE = I·ω²·Cs, so the required inertia is I = ΔE/(ω²·Cs). Punch press: a 0.25 s bite from a 25 rad/s wheel with Cs = 0.12 asks I ≈ ΔE/(625·0.12) — and if ΔE comes in spikes, the radius term (I ∝ R⁴) is your cheapest friend.
First balance the mean torque (the DC motor part); then size I for the ripple. Mount the flywheel on the fastest shaft: stored energy per kilogram grows with ω², and bearings prefer the smaller, faster wheel.
A mass on a spring obeys ẍ = −(k/m)·x and breathes at T = 2π·√(m/k) — the one formula of all vibration.
A mass on a spring obeys ẍ = −(k/m)·x and breathes at T = 2π·√(m/k) — the one formula of all vibration.
Attach a mass to a linear spring and displace it: the restoring force is proportional −k·x, the motion a pure sine. The oscillation period T = 2π·√(m/k) depends only on the ratio of mass to stiffness — not on amplitude (that is what isochronous means, and why pendulum clocks work).
Design uses: picking k for a target period is k = 4π²·m/T². Example, the lab shuttle: m = 2 kg, T = 2.00 s → k = 4π²·2/4 ≈ 19.7 N/m (= 0.0197 N/mm in the app’s units). Gravity in the spring direction shifts only the equilibrium, never the period — a linear system does not care where it oscillates about.
Isolate machines by making the mount’s T at least ≈ 3× longer than the forcing period (√2 is the theoretical border of isolation). Stiffen structures the other way to push resonance above operating speed.
ζ = c/(2√(k·m)) picks the character: ζ < 1 rings, ζ = 1 snaps, ζ > 1 crawls. Most machines order ζ ≈ 0.2–0.7.
ζ = c/(2√(k·m)) picks the character: ζ 1 crawls. Most machines order ζ ≈ 0.2–0.7.
The damping ratio ζ scales the physical damper c against the mass-spring stiffness: ζ = c/(2·√(k·m)). Under 1 the step response overshoots and rings; at 1 (critical) it returns in the shortest ring-free time; above 1 it never overshoots but gets lazy.
The overshoot percentage is a pure function of ζ: OS = 100·e^(−πζ/√(1−ζ²)). Read it backwards for design: 40% overshoot ↔ ζ ≈ 0.28; 10% ↔ ζ ≈ 0.59; 44% ↔ ζ ≈ 0.25 (the neighbourhood of the valve lab). Then c = 2·ζ·√(k·m). For the lab valve (k = 0.3 N/mm, m = 2 kg): ζ ≈ 0.265 needs c ≈ 13 N·s/m = 0.013 N·s/mm.
Never damp blind: too little and the machine hammers its stops; too much and response time dies in syrup — and the damper turns the difference into heat. Instrument the first run (peak and settle) before you touch c.
t = √(2h/g), v = √(2gh), and then the collision happens in milliseconds and nobody sees it — unless you design the catch.
t = √(2h/g), v = √(2gh), and then the collision happens in milliseconds and nobody sees it — unless you design the catch.
A dropped body accelerates at g ≈ 9.81 m/s²: after height h it has fallen for t = √(2h/g) and arrives at v = √(2·g·h). From 220 mm that is 0.212 s at 2.08 m/s — the lab drop, exactly.
The impact itself is an impulse: forces spike to many times the weight and vanish in milliseconds. The coefficient of restitution e (0 = plastic thud, 1 = perfect bounce) sets the rebound; real captures use compliant layers to stretch the stopping distance — peak force falls as the distance grows: F_avg·d = ½·m·v².
Catching means absorbing energy quietly: compliance at the contact, a damper behind it, and geometry that cannot bounce out. Check the peak force, not just the final rest — that is the number that dents trays.
Gears count teeth, belts count pitch radii: ω₂/ω₁ = z₁/z₂ = r₁/r₂. Torque follows inversely, power passes through.
Gears count teeth, belts count pitch radii: ω₂/ω₁ = z₁/z₂ = r₁/r₂. Torque follows inversely, power passes through.
A gear pair meshes teeth: ω₂ = −ω₁·z₁/z₂ (the minus is the external contact — it reverses direction). A belt or chain wraps pitch circles: ω₂ = +ω₁·r₁/r₂ (open belt keeps the sign; crossed flips it). Both are velocity constraints — ratios from geometry, not from force.
Torque trades inversely: T₂ = T₁·(z₂/z₁)·η — speed down, torque up. Efficiency: gears ≈ 0.98 per mesh, chains ≈ 0.95, V-belts 0.90–0.96 with slip of 1–2% making the true ratio a little soft. The pitch — not the outer diameter — is the geometry that counts.
Size the slow-shaft components at the multiplied torque; the little pinion is usually the fatigue item (more cycles per tooth). Keep belt reduction per stage ≤ 6 and gear stage ratios ≤ 8 or the diameters get silly.
P = T·ω = F·v. If a catalogue’s torque × speed beats its wattage, the catalogue is lying.
P = T·ω = F·v. If a catalogue’s torque × speed beats its wattage, the catalogue is lying.
Power is energy per second and the one budget that survives every transmission (minus the tax): P = F·v for linear, P = T·ω for rotational. 1 N·mm × 1 rad/s = 1 mW — in the app’s mm units, power reads straight in milliwatts.
Sanity example: a 13 N·mm steady torque at 25 rad/s carries 0.325 W. A gearbox cannot create power — 2:1 down means double torque at half speed at the same P (times η ≈ 0.98 per mesh). Motors are honest when you check them as P = T·ω across the catalogue curve; peaks above the advertised P are marketing or thermal games.
Size the thermal side at continuous P (motors hate rms current), the mechanical side at peak torque (shafts and pins hate peaks). When in doubt, measure ω live and multiply — the Studio’s chart gives you both factors.
Every real gear mesh has clearance — on reversal the driver crosses the gap and the load stands still. Precision dies in that gap.
Every real gear mesh has clearance — on reversal the driver crosses the gap and the load stands still. Precision dies in that gap.
Backlash is the operating clearance between mating gear teeth (and lead-screw flanks, splines, pins). Without it gears would bind and overheat; with it, every reversal of the driver sweeps a dead band of lost motion before the load feels anything.
Effects: positioning error on reversal (the classic dial test: hold the output, wiggle the input — the dial reads the backlash), impact noise at torque reversal, and limit-cycle hunting in stiff servo loops. Cures by design: anti-backlash split gears with springs, preloaded double nuts on ball screws, or putting the encoder on the load side so the loop sees the truth.
Specify backlash as a circumferential µm or an input angle; more teeth and smaller module reduce the angle at the output for the same physical clearance. In reversing duty, price the anti-backlash option before the first prototype — retrofitting is surgery.
Geneva, ratchet, escapement, cam indexer: four answers to “turn steadily, deliver in steps”.
Geneva, ratchet, escapement, cam indexer: four answers to “turn steadily, deliver in steps”.
Intermittent-motion mechanisms convert continuous input into start-stop output: a dwell, then a jump. Film advances, tool changers, counters and packaging wheels all live on this family.
Geneva wheel: a pin on the driver enters a radial slot and indexes the star by 2π/n, then a locking ring holds it dead until the next pass — smooth, self-locking at dwell. Ratchet & pawl: a saw-tooth wheel passes in one direction and bites in the other; step = tooth pitch, simple and brutal. Escapements mete one tooth per tick from a resonator. Cam indexers roll the whole law into a barrel cam: any dwell-rise profile, zero clearance, the expensive best.
Choose by step precision, speed and budget: ratchet for coarse and cheap (winders, jacks), Geneva for moderate counts with locked dwells, cam indexer where money buys silence. Watch the acceleration shock at engagement — it eats teeth at speed, which is why Geneva wheels grow slots (n ≥ 4) rather than shrink radii.
Analysis starts with a mechanism; synthesis starts with a wish. Three wishes (precision points) exactly price a four-bar.
Analysis starts with a mechanism; synthesis starts with a wish. Three wishes (precision points) exactly price a four-bar.
Synthesis inverts analysis: you specify what the machine must do — an input-output angle law (function generation), a path (path generation), or moving poses of a whole body (motion generation) — and solve for the link lengths that deliver it.
Precision points are the wishes you write down: Freudenstein’s equation fits a crank-rocker exactly through three angle pairs; three path points with free coupler shape (the Varedi-Koulaei & Rezagholizadeh method the Studio ships) fix the pivots as circumcentres; between the wishes the mechanism interpolates smoothly — which is polite language for approximately. Bracket the working range with the precision points and verify the rest by simulation.
After any synthesis, audit by re-simulation: Grashof class, transmission angle window, and replay error at ten intermediate poses. Never dimension from a synthesis print without that audit — the algebra guarantees the points, not the path between them.
Two links, one shared axis, one remaining freedom — the atom of planar machinery.
Two links, one shared axis, one remaining freedom — the atom of planar machinery.
The revolute joint (pin, hinge) constrains two points to coincide and leaves only relative rotation. It transmits any in-plane force through the pin centre but no moment.
Physical pins ride in bushings or bearings with radial clearance; the reaction always passes through the pin centre, so its direction is free — two unknown components for your FBD. Angle limits (hard stops) turn the hinge writable drawers, bonnets and calipers honest; a pin with dry friction adds the Coulomb tax from the friction lesson.
Bearing pressure p = F/(d·L_projected) is the check that kills pins; keep p within the bushing material’s limit and keep bending spans short. Wear budget: clearance grows roughly linearly with cycles under poor lubrication.
One axis, zero rotation — a straight-talking joint that pays for its honesty with side-load sensitivity.
One axis, zero rotation — a straight-talking joint that pays for its honesty with side-load sensitivity.
The prismatic joint lets two links translate along one shared axis and blocks relative rotation. It reacts with a force perpendicular to the axis plus a full moment — the only common lower pair that resists a couple.
Its enemies are cantilever loads: an off-axis force walks the slider into its guides at two contact patches (the guide-length lever sets the patch forces). Longer engagement = lower patch pressure = less binding. Uncheck rotLock in the Studio and the same slot becomes a cylindrical joint — slide plus spin, the drill-chuck connection.
Travel stops are part of the design, not an afterthought: end-of-stroke impact is the classic way sleds hammer their rails. Budget stop energy like a mini crash, damper included.
A weightless bar between two pins enforces exactly one thing: constant distance. It pushes only along itself.
A weightless bar between two pins enforces exactly one thing: constant distance. It pushes only along itself.
The rod link (dyad) is the ideal connecting rod: two pins, fixed length, negligible mass. Mechanically it is a distance constraint — and a two-force member, so its force always runs along the rod axis.
That direction certainty is analytically golden: the rod hands pure tension/compression between its ends, with no transverse shear and no moment. Add mass and the ideal fades — but as a modelling simplification it turns a painful 3-body problem into a lever problem, and the solver uses exactly this trick for coupler spans.
Use the rod link when the physical link’s inertia is small against the stroke forces (light pushrods, long tie-bars). Put physical rods in compression under column checks: Euler buckling waits at (π²EI)/L².
A pin riding in a straight slot removes one freedom and forgives a lot of geometric sin.
A pin riding in a straight slot removes one freedom and forgives a lot of geometric sin.
The pin-in-slot (point-on-line) joint pins one link’s point onto a line of the other: the pin may slide freely along the slot but never leave it. One constraint — between the revolute’s two and a cam pair’s one-sided contact.
Force-wise it transmits only the slot normal — no force along the slot. That single clean direction makes it the mechanism designer’s adapter: slotted levers turn rotation into oscillation with infinite side-play tolerance, and stops at the slot ends turn continuous travel into a polite bang. Jamming arrives when the transmitted force approaches the slot direction: efficiency slides to zero with sin.
Keep the working force at ≥ 25° to the slot; add end stops with an energy budget; pair with rollers or low-µ pads when cycle counts climb.
x = r·cos ωt exactly — the only common crank that writes pure harmonic motion without approximation.
x = r·cos ωt exactly — the only common crank that writes pure harmonic motion without approximation.
The Scotch yoke replaces the connecting rod with a pin riding a straight slot in the ram: crank rotation maps to exactly sinusoidal translation, x = r·cos(ωt). No rod angularity, no second-order distortion.
The price is side thrust: the slot carries the full reaction perpendicular to motion, so wear concentrates on the slot faces and the pin sees sliding at speed ω·r. Compared to a slider-crank, the yoke is shorter (no rod length) but thirstier for lubrication; its perfect sine is loved by vibration shakers, pumps and anyone allergic to harmonics.
Size the slot’s projected area for the transverse reaction force (it equals the crank force’s full radial component at mid-stroke), and choose pin/slot materials as a bearing pair, not as afterthoughts.
Two sliders at right angles, one bar: every point of the bar writes an ellipse; the mid-point writes a circle.
Two sliders at right angles, one bar: every point of the bar writes an ellipse; the mid-point writes a circle.
The elliptic trammel guides the two ends of a bar along perpendicular rails. A point at distances a and b from the ends traces an exact ellipse with semi-axes a and b; the bar’s midpoint (a = b) traces a circle.
The proof is one line of coordinates, but the mechanism is industrially real: it is the inversion basis of the Oldham coupling and the geometry behind ellipse-cutting attachments on lathes and CNC engravers. Its instant centre always sits at the corners of a rectangle on the fixed circle — a pretty way to see ICs at work.
Keep the rails square: angle error ε between the rails warps the ellipse by the same order. For instruments, harden the rail edges — the contact patches are the accuracy.
Want a swing from a motor? Make the crank shortest, ground a neighbour of it, and respect the length recipe.
Want a swing from a motor? Make the crank shortest, ground a neighbour of it, and respect the length recipe.
The crank-rocker is the workhorse inversion of Grashof: ground a link adjacent to the shortest and the shortest link spins while the far side rocks. It converts rotary drive into a designed swing — wipers, sieves, crushers.
A sizing recipe that rarely fails: choose the rocker swing ψ and stroke angle symmetry first; pick crank r₂ ≈ (0.2–0.35)×ground r₁ to keep the transmission angle in the 40–140° window; coupler r₃ ≈ rocker r₄ within ±30% for gentle curves. Then verify: Grashof margin, µ at both extremes, and the rocker’s quick-return asymmetry is set by the crank’s dead-centre poses.
If the µ-window fails, lengthen the coupler before you shorten the crank — the same swing, kinder forces. Mount the motor on the crank and keep the flywheel on the crank shaft too: the dead points need its inertia only there.
Ground opposite the shortest: both sides rock. Ground the shortest: both sides spin. Same four lengths, opposite machines.
Ground opposite the shortest: both sides rock. Ground the shortest: both sides spin. Same four lengths, opposite machines.
Grashof’s inequality gives two more inversions: the double-rocker (ground the link opposite s — both pivoted links only swing) and the double-crank / drag-link (ground s itself — both pivoted links rotate fully).
The double-rocker is the steering and level-luffing geometry: carefully oval lever arms, no full turn anywhere, coupler does the travelling. The drag-link is a rotation-to-rotation converter with a built-in nonuniformity: output speed surges and sags once per turn — the historical quick-return driver before the Whitworth. Non-Grashof chains (s + l > p + q) are always double-rockers and carry a forbidden branch you must not assemble into.
For double-rockers check the two toggle-adjacent boundary poses: that is where forces spike and where re-assembly after maintenance flips the branch if the stops are missing.
Opposite links equal: the coupler never rotates, it translates on a circle — platforms stay level by geometry.
Opposite links equal: the coupler never rotates, it translates on a circle — platforms stay level by geometry.
A four-bar with both pairs of opposite links equal is a parallelogram linkage: crank and rocker turn identically, and the coupler moves in curvilinear translation — every point circles, and the coupler’s angle never changes.
That angle-keeping is pure gold for carriers: wiper blades stay parallel, scissor-lift platforms stay level, and drawing pantographs copy instead of trace. One honest trap: at the flattened change-point pose the parallelogram can flip into the anti-parallelogram branch — counter-phase motion. Railway drivers solved the flip with a second rail; you solve it with a duplicate parallel coupler out of phase or with stops.
Duplicate the coupler (or the drive) when the pose must cross the flat pose; the second loop, pinned 180° out of phase, vetoes the branch flip for free.
A crank pin, a slotted lever, and a connecting rod: uniform rotation in, slow-cut / fast-return ram out.
A crank pin, a slotted lever, and a connecting rod: uniform rotation in, slow-cut / fast-return ram out.
The Whitworth quick-return (1900s shapers and slotters) drives a slotted lever with a crank pin; the lever’s swing drives the ram through a connecting rod. The crank pin circle vs lever pivot geometry sets the forward/return angle split — the time ratio of the quick-return lesson.
Because the crank pin passes on two unequal arcs (α and 360°−α), the ram strokes slow–forward and fast–home at perfectly constant motor speed. Moving the crank pin circle relative to the lever pivot re-tunes TR; stroke length follows the lever throw and rod ratio; the velocity profile stays smooth — no impacts inside the cycle, which is why the idea outlived the shaper.
Set TR from α directly: α/(360°−α). Check ram acceleration at the return reversal: it is (TR)²-ish against the cutting stroke, and it sizes the rod bearings.
Two levers and a short coupler: its mid-region walks a figure-eight whose middle is a very usable straight run.
Two levers and a short coupler: its mid-region walks a figure-eight whose middle is a very usable straight run.
Watt’s linkage (1784) suspends a coupler point between two rockers: the point traces a symmetric lemniscate-like curve — a figure-eight whose central crossing is, over a useful span, almost perfectly straight.
James Watt called it his proudest invention because planing machines for true guide rails did not yet exist: the linkage created straightness from pins. Proportions matter: equal arms, coupler short against the arms, trace at the midpoint — lengthen the arms for a longer straight percentage. Vehicle live-axle guides (Watt’s link vs Panhard rod) use exactly this cancellation today.
The longer the straight stroke you demand, the longer the arms grow for the same error — space decides feasibility. Check both stroke ends for departure from the line: the recallibrating move is coupler length, not arm length.
Chebyshev designed the curve’s flat run the way he designed polynomials — by spending the error budget where it buys the most line.
Chebyshev designed the curve’s flat run the way he designed polynomials — by spending the error budget where it buys the most line.
The Chebyshev lambda linkage is a double-rocker whose coupler point traces a closed loop with a long, nearly straight lower portion — the “foot” of a walking machine or the guide of a sorter.
Its proportions come from Chebyshev’s minimax thinking: instead of one exact line, distribute the departure error evenly (equal ripple) across the span — with a 2 : 2.5 : 1-ish ratio ladder the flat part reaches ≈ 1/3 of the loop with deviations you must measure to believe. The lambda shape converts continuous crank rotation into a stride with essentially constant foot speed on the ground half.
Use it when a foot-speed plateau matters (walkers, carrying fingers); use Watt when you need bilateral symmetry with simple parts. Both pay for flatness with arm length.
OP × OQ = const turns a circle into a line — six links and one inversion identity, zero approximation.
OP × OQ = const turns a circle into a line — six links and one inversion identity, zero approximation.
The Peaucellier–Lipkin linkage (1864) arranges a rhombus on two long arms so that a guide point Q is the geometric inverse of the driven point P: OP·OQ = k², an exact identity. Constrain P to a circle through the pivot O and Q must trace a mathematically perfect straight line.
Circle inversion is the entire trick: inversion maps circles through the centre to straight lines, and the rhombus cell is a machine that performs inversion continuously. It answered Lord Kelvin’s challenge and remains the canonical proof that pins alone can create exactness — the price is eight joints of slop stacked into one deterministic pose.
Joint clearance multiplies through the chain: micro-clearances grow to visible wander at Q; instruments keep pins lapped and pre-sprung. For long strokes the mechanism spreads wide — measure the envelope before romanticising exactness.
A crank-rocker coupler point with a long flat-and-fast ground stride: the genome of walking linkages.
A crank-rocker coupler point with a long flat-and-fast ground stride: the genome of walking linkages.
The Hoeckens linkage is a crank-rocker proportioned so its coupler point traces an almost-rectangular loop: a long, flat, constant-speed stride below and a quick return above — a robotic foot before robots existed.
The family tree is rich: Klann (Jansen’s Strandbeest uses another branch) and Chebyshev’s lambda chase the same footprint. The design knobs are coupler length ratios stretching the flat span, and crank speed blending stride frequency. Since feet must share the load, real walkers phase two or three loops per side, and ground clearance on the return half is the shape you tune, not the stride itself.
Judge a walker loop on three numbers: stride length / body height, foot-speed flatness through the stance, and clearance at mid-return. The Hoeckens is the yardstick because it delivers the plateau with only four links.
cotδ_outer − cotδ_inner = track/wheelbase — the one line that keeps tyres rolling instead of scrubbing.
cotδ_outer − cotδ_inner = track/wheelbase — the one line that keeps tyres rolling instead of scrubbing.
When a car turns, inner and outer wheels circle the same centre, so they must steer different angles. Ackermann geometry arranges the steering trapezium (a double-rocker) so that cot δ_o − cot δ_i = w/l holds approximately across the lock.
The exact ideal lives on one point per turn; the trapezium approximates it, and designers place the error budget strategically — near-true in the small-angle majority and growing only into parking angles, where scrub is slow and cheap anyway. The tie-rod length and the steering-arm angle are the two proportions that bend the error curve: the Studio sample carries a straightness-style self-check on this error, and low-speed geometry is honest — tyre slip angles bend the truth at speed.
Check the error at 20%, 60%, 100% lock, not just at full lock. Modern cars tune some anti-Ackermann for turn-in bite; load and toe change the live numbers more than the linkage ever will.
Two links snapped past dead centre lock; released with a finger. Holding force is sold separately from clamping force.
Two links snapped past dead centre lock; released with a finger. Holding force is sold separately from clamping force.
A toggle clamp is a four-bar driven over its dead point: at lock, the geometry multiplies hand force into tonnes of jaw load and then the geometry — not the spring — keeps it locked.
Two numbers define it: clamping force (what the jaw applies at snap) and holding force (what the joint can resist before the links shear). The first comes from the toggle multiplication F/(2·sin θ) with θ near zero; the second is plain pin/link strength. Release needs only a nudge back through the dead point — which is why clamps ship with ergonomic handles, not muscles.
Dimension from the holding side (the load that must never slip), then check clamping force is enough for the fixture’s squirm. Over-travel past dead centre is small — a workpiece thickness change eats it, so shim the spindle stop, not the handle.
v = rpm × lead. Self-locking by friction, back-drivable if you roll it — the oldest precision linear drive.
v = rpm × lead. Self-locking by friction, back-drivable if you roll it — the oldest precision linear drive.
A lead screw converts rotation to linear travel through thread helices: one turn advances the nut by the lead (pitch × starts). Torque amps: a small hand torque multiplies into tonnes of axial force.
Its secret is friction management: sliding-thread screws (trapezoidal/Acme) run lead angles of 3–6° against friction angles of 7–12°, which makes them self-locking — the load cannot spin the screw. Ball screws roll instead of slide: η ≈ 90%, no stick, no lock — and a mandatory brake or counterweight when the axis hangs vertical. Buckling caps long slender screws at Euler load (π²EI)/L² with the end-fixity multiplier.
Vertical axis without holding power? Self-locking Acme or a ball screw with brake. Duty cycle?: sliding screws heat fast — the PV limit (pressure × sliding speed) is the catalogue line that ends arguments.
One number — module m = d/z — sizes every tooth; the involute flank forgives centre-distance error nobody else forgives.
One number — module m = d/z — sizes every tooth; the involute flank forgives centre-distance error nobody else forgives.
Spur gears mesh parallel-axis teeth shaped as involutes. The module m (pitch diameter/teeth, in mm) sets tooth size; the ratio z₂/z₁ sets speed; backlash, the planned clearance, lets oil and error live in the mesh.
The involute’s superpower: exact conjugate motion even when the centre distance drifts — the contact normal and ratio stay put while the pitch circles redraw themselves. That tolerance gift is why gears won. Practical rest: pinion teeth do more cycles than wheel teeth (z₁ turns z₂/z₁ times per wheel rev), continuous contact needs contact ratio > 1.2, and below z ≈ 17 teeth (20° standard) undercut eats the root.
Pick module from bending strength (Lewis) and pitting (Hertz), not from a nice ratio. Force z combinations that hunt-tooth (coprime) when wear must spread; avoid them when timing matters.
One stage tops out near 8:1; compound stages multiply — and idlers flip only the sign, never the count.
One stage tops out near 8:1; compound stages multiply — and idlers flip only the sign, never the count.
A compound train stacks gears on shared shafts: each mesh contributes a z-driven/z-driver factor, and the total ratio is their product. An idler between two gears only reverses direction; it never changes the magnitude.
The product rule makes 100:1 manageable as 8×8×1.56 instead of one impossible wheel pair. Design craft: keep per-stage ratios ≤ 6–8 for geometry and bearing life, share shafts in matched pairs (compound wheels keyed together), and remember the total sign = (−1)^(#external meshes). Wheel counts and axle layout define the footprint; efficiency multiplies too — three meshes at 98% leave you 94.1%, so heat budgets grow with the ratio chain.
Split the total into near-equal stage factors (cube-root-ish) for compactness; put the precision stage at the slow end where backlash at the output is smallest and torque has outgrown the pinions it must shear.
ω₂ = ω₁·r₁/r₂ on the pitch circles, slip subtracts a percent, and the elastic in the loop forgives what gears never would.
ω₂ = ω₁·r₁/r₂ on the pitch circles, slip subtracts a percent, and the elastic in the loop forgives what gears never would.
Belt drives wrap a flexible loop around pulleys: open keeps the sign, crossed reverses it (like an external gear), and the ratio follows pitch radii. Chains are the toothed-belt answer to slip: positive engagement, no creep.
Friction belts slip 1–2% — the true ratio is a hair softer than r₁/r₂, and torque capacity = (T_tight − T_slack)·r with the centrifugal and wrap terms hiding inside. Synchronous (timing) belts and chains remove slip at the cost of noise and polygonal ripple. Elasticity is the charm: belts absorb shock and overload slips before teeth break — a safety valve built into physics.
Watch the wrap angle on the small pulley (bite falls below ~120°); tension belts to ring, not to hurt, and re-tension after run-in. The tension side does the work: put the slack side up, the tight side down.
A pulley changes direction for free; count the segments pulling the moving block and you hold the mechanical advantage.
A pulley changes direction for free; count the segments pulling the moving block and you hold the mechanical advantage.
An ideal pulley redirects rope tension unchanged. A block and tackle doubles rope around moving pulleys so the load rides on n supporting segments: the hand pulls with W/n — and pays by hauling n metres of rope per metre of lift.
Real pulleys tax: journal friction and rope stiffness eat ≈ 4–6% per sheave, so a treasury-style “6:1” rig delivers more like 5:1 in work terms. The Atwood machine (one fixed pulley, two masses) is the physics-lab limit of the same family: acceleration = g·(m₂−m₁)/(m₁+m₂), the slow-motion gravity instrument.
Count segments only at the MOVING block, and count the hauling end only if it lands there. For rigging life, keep sheave diameter ≥ 16–20× rope diameter and re-tension after the first settling stretch.
v = ω·r — the pinion spins, the rack walks (or vice versa); steering racks are the proof under your hands.
v = ω·r — the pinion spins, the rack walks (or vice versa); steering racks are the proof under your hands.
A rack and pinion is a gear mesh with one wheel of infinite radius: the pinion’s rotation maps to the rack’s translation at the pitch line, v = ω·r_pitch, and the rack can equally drive the pinion — a steering wheel turned into a tie-rod’s walk.
Geometry inherits spur-gear truth: module sets the teeth, backlash sets the reversal slop, and the same involute forgiveness applies around the pitch line. Z-motion details decide life: lubricate or the dry rack-polish polishes itself; support long racks against sag (a rack is a beam, not only a gear); end stops before the pinion jumps the last tooth — a jumped rack bends teeth and dogs.
Choose the pinion teeth count against undercut (z ≥ 17 standard) and the rack module against bending; adjustable-pinion “zero backlash” housings re-seat wear without new parts.
The cam’s contour <em>is</em> the program: lift, dwell, return, timed to the shaft — then the pressure angle bills the cheque.
The cam’s contour is the program: lift, dwell, return, timed to the shaft — then the pressure angle bills the cheque.
A cam converts shaft rotation to a designed follower displacement s(θ): dwell–rise–dwell–return on any schedule the profile carries. Followers ride it by roller, flat face or knife edge.
Design backwards: pick the motion LAW first (cycloidal and modified-sine keep acceleration kind; constant velocity snaps at corners), then the prime circle from the pressure angle φ — the contact normal’s tilt against the motion: keep φ ≤ 30° in the thrust stroke or the follower jams and wears. Flat followers limit lift to profile convexity; rollers forgive everything but their own bearing RPM.
Check in order: pressure angle, contact stress, follower spring float (never lose contact — a spring above inertia force at max rpm), and only then the contour’s manufacturing tolerance class.
tanφ_out = tanφ_in · cosα: one cross through an angle — output speed surges and sags twice per revolution.
tanφ_out = tanφ_in · cosα: one cross through an angle — output speed surges and sags twice per revolution.
A single Cardan / Hooke joint transmits rotation through a shaft angle α. The output obeys tanφ₂ = tanφ₁·cosα — rotationally one-to-one over a turn, but its instantaneous speed oscillates between ω·cosα and ω/cosα twice per revolution.
That wobble is vibration incarnate: at α = 30° the speed ripple is ±15°, and inertia turns the ripple into shaking torque ~ ω². The classic cure is the double Cardan: two joints phased so the second’s error cancels the first (intermediate shaft with equal angles both sides) — and the CV joint family packages the cancellation as rolling balls. Heavy trucks accept some ripple and pay the price in driveline harmonics.
Keep single-joint angles ≤ 15–25° at speed; phase yokes properly on double joints (ears aligned or the error doubles instead of cancelling); re-check angles under suspension travel.
Two tongues, one floating disc with grooves at 90°: fixed parallel offset between shafts, absorbed completely.
Two tongues, one floating disc with grooves at 90°: fixed parallel offset between shafts, absorbed completely.
The Oldham coupling joins collinear shafts whose axes drifted apart in parallel by a fixed offset. A centre disc with perpendicular grooves lets each half slide its own direction: the disc orbits at the offset radius, the shafts run 1:1 at every instant.
Kinematically it is the elliptic-trammel inversion (each slide point circles the other axis exactly): constant angular velocity 1:1, no wobble — the cardan’s honest cousin for the offset (not the angled) case. The disc carries the offset as a continuous orbital slide at speed e·ω: that sliding is the wear budget and the size limit, so offsets are small (a few mm) and speeds modest.
Select on: peak radial offset (disc travel), disc/groove pressure (that is where it dies), and floating mass if rpm is real (the orbit is unbalance). Never use it to absorb angular misalignment — that is the cardan’s day job.
Pin enters slot, wheel indexes 2π/n; pin leaves, ring locks it dead: motion duty = 1/2 − 1/n of the cycle.
Pin enters slot, wheel indexes 2π/n; pin leaves, ring locks it dead: motion duty = 1/2 − 1/n of the cycle.
The external Geneva converts constant driver rotation into an indexed star-wheel: with n slots the wheel advances 360°/n per driver revolution, then a locking ring on the driver holds it perfectly still — dwell you can trust, motion you can schedule.
Timing is geometric: the driver is engaged over angle π − 2π/n — for n = 4 that is 90° motion vs 270° dwell (a quarter-duty index). The wheel’s speed ratio peaks at λ/(1−λ) with λ = sin(π/n): n = 4 → 2.41× driver speed at mid-index with zero dwell — and the price at entry is a finite acceleration step that rings the slots at speed. Size n ≥ 4, keep drive speed down as torque grows, and respect the lock: the dwell is where machines print, fill and seal.
Pick n from the required index angle; n = 4–8 for sanity. If the entry shock eats teeth, step to a cam indexer (you buy the whole law) or an internal Geneva (same rotation sense, gentler).
A pawl lets sawteeth pass one way and bites the other: the cheapest irreversible joint ever mass-produced.
A pawl lets sawteeth pass one way and bites the other: the cheapest irreversible joint ever mass-produced.
A ratchet is a saw-toothed wheel plus a spring-loaded pawl: the pawl rides up the tooth flanks one way and drops into the tooth face the other — rotation granted one way, hard stop the other. Advance is quantised to the tooth pitch.
The honest numbers: angle step = 360°/z per tooth; reversal slop ≈ one pitch before the pawl seats (fine feed? cut finer teeth or add a second staggered pawl); pawl tip and tooth face are a sliding-impact pair — point harden both. A winder oscillating 97° per stroke on a 15° pitch nets 6 teeth = 90° per stroke (the lattice floors the fraction) — measured, not poetry.
Design the pawl spring for seat assurance, not force; check pawl shear and tooth crushing at stall torque; silence with fine pitch or a friction member when the click is the complaint.
Friction plates capped at T_cap: below it, one body; above it, energy becomes heat and the machine survives the jam.
Friction plates capped at T_cap: below it, one body; above it, energy becomes heat and the machine survives the jam.
A slip clutch couples two shafts through friction faces pre-loaded to a torque cap: torque under the cap → rigid shaft; over → the faces slide at the capped torque and the excess power becomes heat. A brake is the same friction pair with one plate bolted to the frame — a clutch to ground.
Two duties: overload protection (a jam slips instead of shearing — the torque fuse) and controlled engagement (a launch lets speed grow through slip). Energy bookkeeping: slip power = T_cap·Δω is all heat; temperature fades µ, so big slip energies need big discs or oil baths. Brakes size by energy per stop plus frequency, not just torque capacity.
Set the cap ABOVE worst honest load, BELOW weakest-component torque. Re-set after break-in (N settles); ventilate or oil-bath beyond ~10 W continuous slip per palm of disc.
F = k·x, energy = ½kx² — every suspension, latch and scale is this equation wearing iron.
F = k·x, energy = ½kx² — every suspension, latch and scale is this equation wearing iron.
A spring obeys F = k·x within its range: a linear energy bank (E = ½·k·x²) that returns force for displacement and hysteresis for nothing.
Variants are spring-dimension craft: helical compression (k = Gd⁴/(8D³n): fourth power of wire!), extension with pre-load, leaf and torsion bars (M = k_θ·θ). Springs in series soften (1/k adds), in parallel stiffen (k adds) — the resistor rules upside-down. Pre-load shifts force level without touching k; solid length and buckle are the travel ends, not suggestions.
Design k from the deflection budget first (statics), then check stress and solid length; dynamic duty adds surge (natural frequency of the coil) above ~1/10 of impact speed. Stainless for chemistry, music wire for fatigue life.
F = c·v — a damper stores nothing, returns nothing; it quietly converts motion into warm oil.
F = c·v — a damper stores nothing, returns nothing; it quietly converts motion into warm oil.
A damper (dashpot, shock absorber) resists velocity, not position: F = c·v. Paired with a spring it decides whether machinery rings or settles — the damping-ratio lesson is about exactly this pair.
Real dampers are asymmetric (bump vs rebound valving), multi-stage (shims open progressively), and thermal (oil fades when hot — long downhill runs make brakes of dampers). Unlike a spring the damper returns no energy: all of F·v leaves as heat, which is why dampers are placed in the motion path that must die — doors closing, landings, valve returns.
Set c from the ζ lesson (target 0.2–0.7 for machinery), then valving asymmetry for comfort (soft rebound, firmer bump). Pitfalls: damper friction (stiction) — a stiction-y damper is a spring with amnesia.
A motor holds its speed only while torque fits under the cap — above it, the machine slows and tells you so.
A motor holds its speed only while torque fits under the cap — above it, the machine slows and tells you so.
A drive motor is a speed source with a torque limit: it holds target speed as long as required torque stays under T_max; at the cap it stalls into whatever the load gives. That cap — not nameplate speed — is where designs fail.
Real motor curves pair speed and torque (DC: roughly linear from stall to no-load; servos: flat under rated speed). Two disciplines follow: size rms/continuous torque from the duty histogram (heating), and re-check peak events against the instantaneous cap (starts, jams, stroke extremes). The Studio’s motor element is exactly this contract: target speed until T_max, then honesty.
Gear for torque first (ratio η×i), then verify speed; when a machine “loses power under load” the number to raise is the ratio or the motor size, in that order of cost.
Add one dyad to a four-bar and six links bloom: one ternary link decides the whole family’s character.
Add one dyad to a four-bar and six links bloom: one ternary link decides the whole family’s character.
Six-bar linkages start from a six-link chain with one DOF and ground different links to get the Watt and Stephenson families — mechanisms whose output can dwell, reverse mid-swing, or run a stroke the four-bar cannot approximate.
The catalogue uses: dwell mechanisms (coupler curve with a near-circular arc feeding a second dyad: while the pilot point rides the arc, the output sleeps), function generators with two inflections, and luffs/level links in cranes. Analysis does not scale: solve them as a four-bar plus an appended dyad-driver, iterate, and verify against closure every degree — exactly what your solver does in hardware speed.
Before reaching for six links, exhaust four: transmission window, stroke, special curve. The added dyad multiplies joint clearances — six-bar dwell quality is slop-sensitive, which is why cams took that market.
Every challenge is a broken machine plus live solver checks: fix the physics, the checks turn green — no scripts, no mercy.
Every challenge is a broken machine plus live solver checks: fix the physics, the checks turn green — no scripts, no mercy.
The Mechanism Studio’s Guided lab (🎓 Labs) drops a real apparatus on your bench with one thing deliberately wrong: a pulley set to the wrong ratio, a seesaw with a kid off her mark, a shuttle with the wrong spring.
The checks you see are the very engine that proofs the sample library: the solver sweeps angles, periods, statics and ratios from your live model — not from a stored answer key. Edit any Inspector number (or drag a body), re-run, and the checks re-grade instantly. When every check passes, the fixed machine is yours to keep editing.
Method that works locally: read the brief, name the ONE wrong thing, compute the target on paper, type the exact value, run, read the measured line (the checks show what the solver got — compare, don’t guess).
Sweep both shafts for a full turn, divide the travels — the measured ratio, not the hoped one.
Sweep both shafts for a full turn, divide the travels — the measured ratio, not the hoped one.
A velocity ratio is measured, not assumed: run the drive through one full driver turn and read how far the driven side swept. sweep_B / sweep_A is the truth of the machine as wired, belts, slip and all.
In the lab the checks compute exactly this: the solver tracks each pulley’s unwrapped angle and divides the sweeps. Your job is smaller — type the fix and watch the measured line land inside the tolerance band. Belt drives remember: ratio = r₁/r₂ with sign kept; gears negate. The lab’s band is ±5% and it measures, so 1.95:1 passes and “about half” does not.
Whenever a drive “feels wrong”, measure the ratio before touching anything: a typed 120 instead of 80 mm reads instantly as 1/3 instead of 1/2.
ω_B = ω_A/2 asks r_B = 2·r_A — one line of pitch-circle arithmetic, then proof by solver.
ω_B = ω_A/2 asks r_B = 2·r_A — one line of pitch-circle arithmetic, then proof by solver.
For an open belt the pitch circumferences eat the same belt speed: ω_A·r_A = ω_B·r_B. Halving speed doubles the radius; the mixer’s r_A = 40 mm driver asks an r_B = 80 mm follower.
The belt joint carries a second set of numbers: anchor phases (a0, b0) place where the belt leaves each pulley — mirror them when you mirror the ratio side, or the strands cross. These are layout numbers, not physics; the physics is only the two pitch radii. After your fix, the solver’s sweep check must read 0.500 within its band.
Check direction first (crossed vs open), then radii, then anchors. If the ratio is right but the machine looks wrong, it is the anchors — not the physics.
Ratio right, strands closed, wrap alive, slip budgeted: the four glances that catch every belt mistake.
Ratio right, strands closed, wrap alive, slip budgeted: the four glances that catch every belt mistake.
Belts fail quietly: a mirrored anchor crosses the strands, a big ratio starves the small pulley’s wrap, tension sags into slip. Four glances catch it all before commissioning.
One: ratio — r₁/r₂ with the right sign (open vs crossed). Two: strands — the belt must leave and arrive on clean tangents (the lab’s b0 exactly). Three: wrap — ≥ 120° on the small pulley or the torque capacity vanishes. Four: slip — friction belts give up 1–2%, so a measured 0.495 against a nominal 0.5 is the belt being a belt, not a bug.
If the machine demands an exact count (printing, indexing), a friction belt is the wrong technology — timing belt or chain, and stop reading ratios to two decimals.
m₁·x₁ = m₂·x₂, computed in one line and verified by a solver that cannot be sweet-talked.
m₁·x₁ = m₂·x₂, computed in one line and verified by a solver that cannot be sweet-talked.
The seesaw of the lab is moments incarnate: two kids as point masses, one pivot. Balance needs the moments equal and opposite — 40·x₁ = m₂·x₂ — the same line your textbooks make you prove on paper.
Compute the lab directly: left kid 30 kg at 400 mm, right kid m₂ = 26.67 kg. Her balancing seat: x₂ = (30·400)/26.67 = 450 mm. The lab ships with her at 550 mm — an extra 100 mm of lever means 26.67 g · 0.1 m ≈ 26 N·mm of unbalanced moment, visible as the slow tip the check measures while it waits for level.
Press Reset after the edit: equilibrium is judged on the settled pose, and a plank mid-swing from a disturbed start lies to you for seconds.
x_com = Σ(m·x)/Σm — put it over the pivot and the world holds still; the seesaw is the two-body case.
x_com = Σ(m·x)/Σm — put it over the pivot and the world holds still; the seesaw is the two-body case.
The centre of mass of an assembly is the mass-weighted average position: for a set of point masses, x_com = Σ(m_i·x_i)/Σm_i. Gravity acts on the assembly as if all mass lived there — so balance about a pivot needs the COM above it (with zero offset).
This is the seesaw’s second reading: pivot at 0, kids as two masses — x_com = 0 is exactly the moment equation. The idea generalises far: a machine’s static tippiness tracks COM height vs base width; cars hate tall load; cranes counterweight their jibs so the slewing COM sits near the mast axis. Move any part in the Studio and the COM follows — visible as the plank’s bias.
Compute COM of every add-on before mounting: 2 kg at 400 mm overhang costs the same moment as 8 kg at 100 mm — price it before it tips something.
The lab grades the settled pose, not the journey: damping drains the swing and the LAST reading is the verdict.
The lab grades the settled pose, not the journey: damping drains the swing and the LAST reading is the verdict.
A dynamic truth with a static verdict: the lab’s plank check reads the state after settling. That is why the brief nags you to press Reset after edits — the solver needs a rest-start to reach a trustworthy equilibrium.
With damping present (the lab’s pin friction), any unbalanced start decays to the pose where ΣM = 0 — or keeps drifting if the moments never balance. The last value of the angle is therefore an equilibrium gauge: for a balanced plank within 3.5° the check passes; an unbalanced one keeps trending and fails on the same number, honestly.
Order of operations: edit → Reset → Run → read. Measuring a moving system is the #1 self-inflicted lab bug, and Reset costs one click.
k = 4π²m/T² — for the 2 kg shuttle and T = 2.00 s that is 19.7 N/m, one crisp number.
k = 4π²m/T² — for the 2 kg shuttle and T = 2.00 s that is 19.7 N/m, one crisp number.
The lab’s shuttle is a pure mass-spring: its period obeys T = 2π√(m/k) and only that. Solved for the spring: k = 4π²·m/T² — and with m = 2 kg and the demanded T = 2 s, k = 2π² ≈ 19.739 N/m.
Units are the only art: the Studio counts springs in N/mm, so 19.739 N/m = 0.019739 N/mm — the value the spring Inspector wants. Amplitude drops OUT of the formula (isochronism): stretch the start 10 mm or 100 mm, the period is identical — verify that on the chart twice, it is worth seeing.
If the measured period lands short, k is too big by the SQUARE of the period error: 10% fast = k 21% too stiff — square rules, not linear ones.
Count upward zero-crossings of displacement and divide time by count: the robust way the lab itself measures T.
Count upward zero-crossings of displacement and divide time by count: the robust way the lab itself measures T.
Peaks smear and damp; zero-crossings are sharp. The lab’s period check timestamps every upward crossing of the displacement mean and takes T from their spacing — the same method you should use on any strip chart.
Why upward only: it fixes the phase slot, so no half-periods sneak in. Why the mean, not zero: gravity or pre-load shifts the trace, and a raw-zero rule misses cycles. Three crossings yield two intervals — average them. With damping decaying amplitude late in the run, early crossings are the reliable ones, which is why the lab’s tolerance of ±0.05 s is achievable.
On your own charts: mark crossings FIRST, then compute; on noisy traces low-pass before crossing-hunting — a counting error of one cycle is a T error of 33% at three crossings.
x₀ = m·g/k shifts the centre of oscillation; T = 2π√(m/k) is blind to it — a linear system’s superpower.
x₀ = m·g/k shifts the centre of oscillation; T = 2π√(m/k) is blind to it — a linear system’s superpower.
Hang a spring vertically and it settles stretched by x₀ = m·g/k before oscillating. That offset is the gravity bias: the motion centres on the stretched equilibrium, not the spring’s free length.
The linear spring equation absorbs g entirely into the constant term — the oscillatory part is unchanged, so the period is the same horizontal, vertical, or on the Moon (with a different bias). The lab’s sideways-rail shuttle demonstrates it by construction: perpendicular gravity cannot bias the rail, so the measured period is pure k and m. Any vertical spring check you invent must subtract x₀ before crossing-counting.
Diagnose springs VERTICALLY in two steps: first statics for x₀ (does the sag match k?), then dynamics for T; mixing the two numbers is the classic lab-report error.
A step input writes the system’s signature: overshoot, peak time, settle — three numbers that identify (k, m, c).
A step input writes the system’s signature: overshoot, peak time, settle — three numbers that identify (k, m, c).
Apply a constant force at t = 0 to a spring-mass-damper and watch: the trace leaps, overshoots, rings down, settles at F/k. That curve — the step response — carries every second-order parameter visibly.
The three measurements: final value = F/k (stiffness), overshoot % → ζ via the log formula (damping), peak time t_p = π/ω_d with ω_d = ω_n√(1−ζ²) (natural frequency). For the lab valve: F = 10 N, k = 0.3 N/mm → settle at 33.3 mm; at ζ ≈ 0.265 the peak grazes ≈ 33.3·1.44 ≈ 48 mm inside the graded 43–51 band.
Instruments first, edits after: read the three numbers off the chart before honouring any knob. A step that settles at 30 mm when 33.3 was demanded says the spring — not the damper — is wrong.
ζ = |ln OS|/√(π²+ln²OS), then c = 2ζ√(km): the two-line factory that converts “about 40%” into hardware.
ζ = |ln OS|/√(π²+ln²OS), then c = 2ζ√(km): the two-line factory that converts “about 40%” into hardware.
Overshoot is a pure function of damping ratio ζ: OS = e^(−πζ/√(1−ζ²)). Invert it (safe for any OS between 0 and 1): ζ = |ln OS|/√(π² + ln²OS). Forty percent? ζ = 0.916/√(9.87+0.84) ≈ 0.28.
Then the physical damper: c = 2·ζ·√(k·m). With the lab’s k = 0.3 N/mm (300 N/m) and m = 2 kg: √(k·m) = √600 ≈ 24.5 (SI), so c = 2·0.28·24.5 ≈ 13.7 N·s/m ≈ 0.0137 N·s/mm — the ballpark the lab grades. Too little c and the stop hammers; too much and the valve is late: the lab’s band is exactly this trade made explicit.
Overshoot targets are system specifications (product feel and hammer-wear), not taste: 5–10% for precision axes, 30–45% for snappy valves that must arrive NOW.
x_∞ = F/k is statics inside a dynamic run; the 2% band says when the ringing stops mattering.
x_∞ = F/k is statics inside a dynamic run; the 2% band says when the ringing stops mattering.
A step response ends where statics said it must: the final displacement is F/k, exactly the statics equation. The settling time t_s is how long the trace needs to stay within ±2% (or ±5%) of that value — t_s ≈ 4/(ζω_n) for the 2% rule.
The two checks grade independent things: the band catches damping too small (never quiet) or too big (still creeping); the final value catches stiffness errors regardless of ringing. For the lab: 10 N / 0.3 N/mm = 33.3 mm, and a correct damper (ζ ≈ 0.265, ω_n = √(300/2) ≈ 12.2 rad/s) settles in ≈ 4/(0.265·12.2) ≈ 1.2 s.
Final-value errors are NEVER damping: rings that fade to the wrong number say k (or the spring’s anchor) is wrong. Diagnose by averaging the last second of trace, not by the peaks.
t = √(2h/g) = 0.212 s and v = √(2gh) = 2.08 m/s: the two numbers that remove myth from “it just fell”.
t = √(2h/g) = 0.212 s and v = √(2gh) = 2.08 m/s: the two numbers that remove myth from “it just fell”.
Constant acceleration g ≈ 9.81 m/s²: distance h = g·t²/2 and speed v = g·t combine into t = √(2h/g) and v = √(2·g·h). From the lab’s 220 mm the ball lands after 0.212 s at 2.08 m/s — nothing mystical, and yet too fast to eyeball.
Uses of the two numbers: the time tells you where the tray must be when the ball arrives (the lab’s task), and the speed prices the catch: stopping in 10 mm of compliance asks an average force m·v²/(2·d) ≈ several times the ball’s weight — peak force scales inversely with stopping distance, always.
Never eyeball sub-second events: chart them. Air drag matters only beyond ~10 s of fall at these sizes — drop tests and catch tests are pure g arithmetic.
Rest = speed decayed to zero AND contact maintained: energy out through bounce and friction, geometry making sure it stays there.
Rest = speed decayed to zero AND contact maintained: energy out through bounce and friction, geometry making sure it stays there.
A ball on a tray is “at rest” when its speed has fully decayed and it sits at the supported height — tray top plus ball radius. Two facts, and the lab checks both: a decay-to-zero speed AND a final y in the support window.
The decay has three doors out: restitution less than 1 (bounce losses), friction scrubbing (the wood/steel pair does this well), and the stop itself (the tray is ground — infinite mass). The little permanent “slop” between nominal contact height and settled y is penetration & compliance — not a bug, but the price of soft contact models. That is why the check’s window is tray + r − slop, honestly worded.
If your catch stands up but keeps micro-bouncing, friction is too low or the material pair too lively — real catches add felt or rubber for exactly this reason.
A catch is geometry plus energy: under the drop line, wide enough for scatter, soft enough to hold.
A catch is geometry plus energy: under the drop line, wide enough for scatter, soft enough to hold.
A successful catch answers three questions: where does the object arrive (aim), how much error can the receiver swallow (area), and how does the energy leave (compliance and friction). The lab’s empty-space drop tests exactly these.
Aim: the drop line is set by the release pose — x must bracket it within half the tray width. Area: real drops scatter (release jitter, air currents) — design catch width ≥ 6σ of the scatter if the receiver is expensive. Slop: the settled height sag is the compliance tax; measure it once and design the checks of your own machines the honest way, with the slop written into the window.
Drag is not positioning: for repeatable catches, type the tray x with the exact-entry box — the lab brief teaches that habit on purpose.
Cs = (ω_max − ω_min)/ω_mean — the one number that grades every uneven machine and every flywheel cure.
Cs = (ω_max − ω_min)/ω_mean — the one number that grades every uneven machine and every flywheel cure.
Machines with pulsing loads do not spin steady: the speed breathes between ω_max and ω_min each cycle. The coefficient of speed fluctuation Cs normalises the ripple to the mean, and every flywheel spec is written in it.
The lab demands ω inside 25 ± 1.5 rad/s: that is Cs = 3/25 = 0.12 or 12% — a tight press spec. Measure Cs from the chart (max and min over steady cycles, not the spin-up transient); if the ripple is asymmetric about the mean, the motor’s mean torque is wrong before the flywheel is small. Cs trades against flywheel size through ΔE = I·ω²·Cs — half the Cs, double the iron.
Typical Cs: crushers 0.2, machine tools 0.03–0.05, looms 0.1, spinning machinery 0.005 — the catalog number is the duty you accept, written as a fraction.
At constant thickness and material, I = ½mR² with m ∝ R² collapses to I ∝ R⁴ — the radius exponent that ends all arguments.
At constant thickness and material, I = ½mR² with m ∝ R² collapses to I ∝ R⁴ — the radius exponent that ends all arguments.
Double a solid disc’s radius (same thickness, same steel): mass grows ×4 (area), inertia ×16 (area × R²). I ∝ R⁴ is the cheat code of flywheel sizing — and the lab brief’s hint, printed in the challenge itself.
Do the arithmetic for the lab: the Ripple lesson gives ΔE from the pulse; I_need = ΔE/(ω²·Cs). At ω 25 rad/s and Cs 0.12 the denominator is 75 — a modest I suffices. A 45 mm solid steel disc, 30 mm thick, carries m ≈ π·45²·30·7.85e-6 ≈ 1.50 kg → I = ½·1500·45² ≈ 1.52e6 kg·mm² — the size class that tames the punch, and the lab measures the result live.
Spend inertia at the rim (rings beat discs per kilogram); put the flywheel on the FAST shaft (ω² discount); stress-check rim speed v = ωR against material (grey iron ≲ 35 m/s).
Pulses average to a number, and the steady torque must equal it — flywheels handle the ripple, not the average.
Pulses average to a number, and the steady torque must equal it — flywheels handle the ripple, not the average.
A periodic load like the punch pulse does not average to zero: the mean torque is duty × pulse. The motor’s steady torque must carry exactly that mean, or the wheel coasts down cycle by cycle no matter its size.
For the lab: −26 N·mm at 50% duty costs a mean of −13 N·mm, hence the +13 steady torque that ships with the apparatus — mean-balanced by construction, as the sample’s label states. The unbalanced residual is only the ripple, and ripple is the flywheel’s department (ΔE = ∫(T − T_mean)dt). Forget the mean step and you will tune radius forever on a wheel that is also sinking.
In any pulsing-drive design: first compute T_mean × time per cycle as the energy the prime mover owes; only then assign ΔE_a.c. to inertia. Two budgets, two owners.
Choose n from the index angle, read engaged arc = π − 2π/n, price the 2.41× style peak, and let the dwell be the promise.
Choose n from the index angle, read engaged arc = π − 2π/n, price the 2.41× style peak, and let the dwell be the promise.
A Geneva design starts from the index angle 360°/n you need per step: n = 4 gives 90° (the lab’s indexer), n = 6 gives 60°. Everything else follows: engaged driver arc = π − 2π/n per revolution, dwell the remainder.
Then the dynamic price: with r (pin radius from driver centre) and centre distance CD, λ = r/CD = sin(π/n) exactly; peak wheel speed λ/(1−λ) times driver speed — 2.41× at n = 4 — and an acceleration step at entry. Design checks: CD and r consistent with sin(π/n); pin diameter for the slot pressure angle at mid-pass; and the lock ring arc covering the wheel rim at every dwell pose.
More slots = smaller peak speed but longer relative dwell; n below 4 is entry-shock country. If torque at entry matters (heavy tables), size the pin for it or move to a cam indexer.
Pitch 15° means 24 teeth; an oscillation of 97° nets the floor — 6 teeth, 90° per stroke, whatever the extra swing.
Pitch 15° means 24 teeth; an oscillation of 97° nets the floor — 6 teeth, 90° per stroke, whatever the extra swing.
A ratchet quantises: tooth pitch p = 360°/z, and the pawl’s floor function grants whole teeth only. The winding arithmetic: net advance per stroke = p·⌊swing/p⌋ — the fractional remainder is returned, not banked.
The lab’s winder swings ≈ 97° on a 15° pitch: ⌊97/15⌋ = 6 teeth = 90° net per stroke (6.9° given back to the return). Over the 6 s of the sample that compounds to the ≈ 467° of wound run that its check reads — a staircase, not a line: dwell, jump, dwell. Fine feeds trade: finer pitch = less return loss, weaker teeth; staggered dual pawls halve the effective pitch without touching the wheel.
Count backwards from required resolution: needed steps per rev = z; then check tooth root shear at stall torque and the pawl spring’s seat force at max return speed.
Cap above honest load, below weakest link: the two fences, then the slip energy bill if it works for a living.
Cap above honest load, below weakest link: the two fences, then the slip energy bill if it works for a living.
A slip clutch has one number that matters: the cap. Set it above the worst legitimate transmitted torque (start surges included) and below the torque that breaks the weakest downstream part. Between the fences it is a rigid shaft; outside, a fuse.
The lab’s 4 N·mm cap: with the brake engaged the speeds split by conductance — driver at 5 rad/s, clutch gain 1.2, brake gain 0.06 settle where clutch drag equals brake drag: 1.2·(5−ω_B) = 0.06·ω_B → ω_B ≈ 4.76 rad/s, while the torque check parks exactly at the 4 N·mm cap — read both on the checks. Working slip burns P = T_cap·Δω as heat at the faces: a clutch that slips for a living needs the thermal budget of a brake, not just a torque setting.
Document the cap on the drawing and in maintenance (springs relax); after any slip event, re-check faces and re-set. If slip is frequent, the protection philosophy wants a shear pin or a clutch with thermal design — pick deliberately.