E-stop active on a used robot controller: trace the circuit

Short answer: a released button does not prove a closed chain. On a second-hand robot cabinet, an e-stop message with the buttons up almost always points at the external safety loop — no 24 V DC reaching it, or the links and devices the last integrator left out. Measure every safety rail to the documented negative reference with the circuit connected, expect the full loop voltage at each input pair, and treat the contact showing that full voltage as your open. Fit only the links the installation manual documents for a cabinet without external safety devices; any other bridge across e-stop contacts is a defeated safety function, not a repair. The chain also runs through the cabinet door, its interlock switch and the gland loom, which is where enclosure work and safety work meet on a used machine.

Why does the pendant say e-stop active when both buttons are released?

Your robot refuses to jog because its controller still sees an open circuit somewhere in the e-stop chain, and on a used cabinet that circuit is usually the external safety loop, not the two buttons you can see and press.

A released actuator closes two contacts; the permission to jog also needs the loop around them, its 24 V DC supply, and the reset the safety design expects. The controller latches the stop command, so it keeps reporting e-stop long after the button springs back. A second-hand cabinet carries the last integrator's choices: safety devices removed, cables cut at the gland, links left in or taken out. None of that shows on the pendant, which is why the message survives a button swap.

What else is wired into the robot's e-stop chain?

The loop the controller watches runs from the safety supply through every external safety contact in the cell — gate switch, cabinet door interlock, floor e-stop stations, a guard relay — before it lands on the external e-stop input terminals.

Many controllers run two channels, one per stop category: category 0 removes power at once and category 1 brings the axis to a controlled stop before removing power (IEC 60204-1 names both). Both channels have to close before jogging returns. Sharing one loop across several stations is normal practice on a robot cell, but every contact added is one more open to find later. The cabinet door puts contacts in the same chain: the interlock switch, and the loom that crosses the hinge through a gland, so door work is safety work. The door side has its own design rules, covered in what each enclosure door interlock needs; the entry itself is covered in cable entry into a controller cabinet.

Which fault comes first on a second-hand cabinet?

Rank the suspects by what changes hands at resale: a loop left unpowered, then the devices and links the last integrator removed, then an open contact in a door interlock or its gland loom, then your own meter reference.

Each pattern points somewhere specific:

• Fault present at power-up, before anyone touches the cell — the loop is unpowered: fuse blown or pulled, supply switch open, return conductor cut. Check the loop fuse from both ends first.
• Fault appears only when someone cycles the cabinet door — the interlock contact or the hinge loom. Measure across the interlock with the door closed.
• One channel reports open while the other stays healthy — a single contact open in one channel of the dual loop. Compare the two channels at the input terminals.
• Stickers or paperwork mention light curtains, gates or mats — those inputs are probably open with no links fitted. Read the external input terminals against the manual.
• You read nothing but zero volts everywhere — often the reference, not the circuit. Move the black lead to the safety supply's return terminal and start again.

How do you measure the safety circuit and find the open contact?

Keep the circuit connected, set the meter to DC volts, and read every rail against the documented negative reference: a healthy rail holds the loop voltage at its load end, and the contact that shows the full loop voltage is the open one.

Work from the source outward. At the loop supply, read line side then load side: nothing on the load side with the supply voltage present on the line side is a blown or removed fuse. Then measure each input pair at the controller, rail by rail, and write down actual volts — a recorded 23.8 V says more than a tick marked present. The physics that makes the search short: in a series loop with one open contact, no current flows, so nothing else drops voltage and the whole loop voltage sits across the open contact.

Say the loop supply holds 24.1 V DC and the controller's external e-stop input is a safety relay coil drawing 50 mA — about 1.2 W at that voltage; replace both figures with the ones in your controller's drawing. With every contact closed, the loop current is 50 mA, so a contact and its wiring totalling 0.2 Ω drop 50 mA × 0.2 Ω = 0.01 V, which no meter hunt will notice. Now the door interlock hangs open. Current stops everywhere, every other element drops nothing, and the full 24.1 V sits across the open contact. Four stabs across four contacts finish the job: three read near the reference and the door interlock reads the full 24.1 V. Measured to the reference rail instead, the input side of that contact shows 24.1 V and the far side shows ground potential — the same conclusion, one stab per point.

Can you jumper the external e-stop terminals to clear the message?

Fit only the links the controller manufacturer's installation manual documents for a cabinet running without external safety devices, and treat any other bridge across e-stop contacts as a defeat of a safety function rather than a repair.

Robot builders publish the terminal pairs — sometimes with polarity — for exactly this configuration, and the controller's own cabinet and pendant e-stops stay in circuit when you use it. Get the manual for your exact cabinet and software revision from the manufacturer or its distributor; a photograph of another cabinet's terminal strip is not a drawing. The stop-side rules still bite. ISO 13850 requires the stop command to latch, with reset as a separate deliberate action taken once the hazard is cleared. NFPA 79 and IEC 60204-1 require direct opening contacts: the actuator mechanically forces the contacts apart, so a button taped closed or looped around cannot be part of any answer. Decision: if the manual documents the links, fit them, confirm the input rail at the loop's rated voltage to the documented reference, and record them; if it does not, the fault is still in the circuit, so keep tracing.

Why do dim LEDs and a clicking relay mislead you?

An indicator glows on a couple of milliamps of leakage or backfeed, and a relay clicks when its own coil alone is fed, so neither proves the rail you care about is healthy under load.

Judge the circuit with the load connected. A rail that reads the loop voltage open-circuit and collapses when the coil pulls in has a high-resistance joint upstream, and the drop tells you where: a joint losing 3 V at a 50 mA loop current is dissipating 0.15 W and hides about 60 Ω (R = V ÷ I = 3 ÷ 0.05). A relay that picks up on one permissive — air pressure, for example — proves that permissive only, and the rest of the chain stays unmeasured. Write actual volts at each named point instead of present or absent, and the pattern shows up on paper.

What do you fix, and how do you keep the fault from coming back?

Restore the loop to the installation drawing — supply fused, rails at the rated voltage under load, devices on the terminals the drawing names, documented links recorded — then leave the drawing and your test record inside the cabinet door.

Work the fix in the same order as the causes:

1. Refeed the loop. Replace a removed fuse at the rating the drawing shows; at 24 V DC, a 0.5 A fuse carries any coil load up to 12 W with margin (P = V × I).
2. Refit documented links. Use the pairs the manual names for the no-external-devices configuration, then label the terminals so the next person knows the links are factory parts, not field ones.
3. Replace the open contact. A worn door-interlock block or a seized cell e-stop contact gets a new contact block, and the loop returns to the drawing rather than to a wire looped around it.
4. Remake every joint that dropped volts under load: terminal, gland conductor or relay socket.

Keep it fixed with labels and one number. Mark the external e-stop terminals "links fitted per drawing, rev 2" so nobody wires a device in parallel with them, and re-measure the loop end to end after every washdown or cabinet move: supply off, a closed loop reads under about 2 Ω, and a rising value is a corroding contact or a wicking loom. The layout side — fuse sizing, e-stop wiring and power routing — is covered in fuses, e-stop and power layout for a machine cabinet, and the monitoring choices in door switch monitoring design checks. Where a documented link runs across terminal blocks, fit a bridge made for that terminal series so it seats fully; terminal bridging and jumper compatibility explains why a mismatched comb fails to make.

When does the enclosure itself become the e-stop problem?

The chain's most exposed contacts sit on the cabinet door and out in the cell, where washdown corrosion, a wicking hinge loom or a cut cable from a removed safety device opens the loop months after the sale.

Rate the parts for the environment, starting at the door. A washdown interlock earns IP67 at least — dust-tight, and sealed against 1 m of immersion for 30 minutes under the IEC 60529 test — and every gland gets a drip loop below it, so water runs off the loom instead of into it.

Out in the cell, the e-stop station is its own small enclosure, and it carries the same duty as the robot cabinet: our stainless wall-mount line runs from 6 x 4 x 4 in to 60 x 36 x 16 in outside (H x W x D), rated NEMA 4X / IP66 by the manufacturer and not UL listed, in 304 for indoor washdown or 316L for salt, brine and chemical areas. You set every gland hole yourself; cutouts, glands and HMI windows covers the hole sizes, and custom enclosures take a one-off station from drawing approval to delivery in 10 business days.

Start from the stainless enclosure size finder, size the box with an enclosure for remote robot I/O, and for the controller's own housing, a robot controller enclosure in a dusty cell covers the environment side. Decision: most used-cabinet e-stop messages die in one session with the meter — loop powered, links per the manual, one contact replaced. Re-specify the enclosure when the fault keeps returning at the door or the cell; that is an environment problem, and a gasketed 316L station with sealed glands and a drip loop ends it.

Catalog dimensions to compare

Compare these catalog dimensions with your component layout. Open each enclosure for its rating, material and usable space, then check those specifications against your application.

Questions