Problem Details
A used Staubli RX60 six-axis arm is being returned to service outside its original cell. The E-stop circuit has been repaired and a replacement interlock panel has been built to satisfy the safety inputs the controller expects. The controller powers up and communicates, but arm power (servo power) will not enable — the amplifiers never close their contactors, and the arm stays in a de-energized state with brakes applied.
This is the single most common wall in a robot recommissioning project, because "arm power on" is not one signal. It is the terminal result of a series chain that spans the safety relay stack, the operator-mode/enabling-device logic, the encoder/calibration state, the amplifier bus, and the application state machine. Any one open link produces the same symptom with no motion and often no obvious alarm at the teach pendant.
Root Cause Analysis
Treat "servos will not power up" as a chain, not a fault. Work the chain in the order the controller evaluates it, because a downstream test is meaningless while an upstream link is open.
| # | Chain link | Typical failure on a used/relocated robot | Evidence to collect |
|---|---|---|---|
| 1 | Incoming power & phase | Wrong supply voltage tap, missing phase, transformer tap left at prior site's voltage | Measure line-to-line at the main disconnect and at the transformer secondary; compare to nameplate |
| 2 | Control transformer / logic supplies | Blown control fuses, failed 24 VDC supply, sagging supply under contactor inrush | Measure 24 VDC under load, not open-circuit |
| 3 | Emergency stop loop | Dual-channel loop jumpered on one channel only; channel cross-fault detected | Continuity on both channels independently, with the loop under normal 24 V |
| 4 | Safety relay / arm-power relay stack | Relay not resetting because reset pulse is missing, or feedback (EDM) loop from the power contactors is open | Relay status LEDs, then the feedback loop terminals |
| 5 | Cell interlocks (gate, fence, light curtain) | The new interlock panel satisfies the wrong pair of terminals, or shorts a monitored dual-channel input in a way the controller diagnoses as a fault | Compare each landed wire against the schematic terminal-by-terminal |
| 6 | Mode select & enabling device | In manual/teach mode the pendant enabling switch (dead-man) must be held in the center position; a missing, damaged, or disconnected pendant blocks power | Test power-on in automatic/remote mode as well as manual |
| 7 | Arm cable / brake and encoder circuits | Damaged or mis-seated arm cables after transport; open brake circuit; encoder battery dead so joints are uncalibrated | Pin-to-pin continuity on the arm cables; check pendant for calibration/encoder alarms |
| 8 | Amplifier / DC bus | Failed amplifier module, blown bus fuse, failed precharge, thermal trip | Amplifier LEDs, bus voltage at the amplifier test points after the main contactor closes |
| 9 | Application / system state | Controller sits in a "system not ready" or error state that latches power-on out; a prior application is loaded and faulting | Read the pendant error history before clearing it |
Two root causes dominate on machines that were removed from a running cell:
- Incomplete safety-circuit emulation. The original cell provided several distinct monitored loops — E-stop, gate/guard, external power-request, and contactor feedback. A replacement panel that bridges only the E-stop leaves the remaining loops open. Worse, bridging both channels of a dual-channel input with a single wire can be diagnosed as a short-circuit fault and will block power on purpose.
- Uncalibrated or unbatteried encoders. If the encoder backup battery died in storage, joint position data is lost. The controller may allow limited jog under a special calibration procedure but will refuse normal arm power until the calibration state is restored using the manufacturer's procedure and the arm's recorded calibration values.
Solution Procedure
- Read the error history first. Capture every code and message on the pendant, with timestamps, before clearing anything. Codes are the fastest path to the correct chapter of the manual and are the first thing OEM support will ask for.
- Verify supply and taps. Confirm incoming voltage matches the controller nameplate and that the transformer primary tap matches your site voltage. Check every control fuse for continuity, not appearance.
- Map the safety terminals against the schematic. Print the safety page of the electrical drawings. Mark each terminal as: OEM-original, jumpered by you, or landed to the new panel. Any terminal you cannot account for is a suspect.
- Prove each loop electrically. With the controller off, ring out each channel of each loop end-to-end. With the controller on, measure the voltage across each input against the schematic expectation. Confirm the two channels of a dual-channel loop are truly independent conductors.
- Confirm the reset path. Many safety relays latch off after power-up and require a momentary reset input. Verify a reset button exists, is wired to the correct terminal, and that its pulse is seen (relay LED changes state on press/release).
- Check contactor feedback (EDM). The normally-closed auxiliary contacts of the power contactors typically feed back to the safety relay. If a contactor was replaced with a different model, or an auxiliary block was omitted, this loop is open and power will never be granted.
- Select the correct operating mode and hold the enabling device. In manual mode, arm power requires the pendant enabling switch held at the middle detent. Squeezing hard (third position) also drops power. If the pendant is missing, the controller will not grant power — source a matching pendant rather than fabricating a bypass.
- Inspect and re-seat arm cables. Check both the power/brake cable and the encoder cable at both ends. Look for bent pins, moisture, and crushed strain relief from transport. Ring out the brake conductors and confirm brake coil resistance is consistent joint-to-joint.
- Confirm the encoder battery and calibration state. Replace the backup batteries as a matter of course on any machine of unknown storage history. If calibration is lost, follow the manufacturer's calibration restoration procedure using the arm's documented calibration data — do not guess offsets.
- Watch the contactors during a power-on attempt. If they pull in and immediately drop out, the fault is downstream (amplifier, bus, feedback, brake). If they never pull in, the fault is upstream in the safety/permissive chain. This single observation splits the search space in half.
Verification
| Check | Expected result |
|---|---|
| All safety loop channels, independently | Continuity on each channel separately; no shared conductor |
| Safety relay status | Relay energizes and stays energized after reset, with no error LED |
| Arm power request, manual mode | Contactors close and hold while enabling device is at mid-position |
| Arm power request, automatic mode | Contactors close and hold with guards closed |
| Brakes | Audible release on power-on; arm holds position on power-off |
| Single-axis jog, low speed | Smooth motion, no following-error or overcurrent fault on any axis |
| E-stop test, each station | Immediate power removal and brake application; error logged; requires reset to restore |
| Calibration | Reported joint positions agree with a known mechanical reference; repeat point-to-point moves return to the same physical location |
When to Escalate to OEM Service
Incremental self-repair versus flying in a specialist is a cost-of-delay decision, not a pride decision. Escalate when you hit any of these:
- Amplifier or controller-board failure. Board-level diagnosis without the OEM service documentation and spares is not economical. Get a quote on an exchange module.
- Lost calibration data with no documentation. Recovering absolute calibration on a multi-axis arm requires the manufacturer's procedure and, for some joints, fixtures.
- Missing or corrupt system software. Controller system software and licensing generally come only from the manufacturer, and version compatibility with the arm and amplifiers matters.
- No electrical schematics. Reverse-engineering a safety architecture from the wire colors is slow and produces a cell you cannot defend in a safety review. Buy the documentation package.
- Sunk-time threshold reached. Set a limit up front — for example, 40 engineering hours or two weeks of downtime — and escalate when you cross it rather than renegotiating with yourself weekly.
Before booking a service call, assemble a package: controller model and serial number, arm serial number, full pendant error history, a photo of the safety terminal strip, your marked-up schematic, and a written list of what you have already proven good. This routinely converts a multi-day site visit into a phone or remote session, and it is the difference between paying for a specialist's diagnosis and paying for their time to rediscover what you already know.
Practical Notes on Used-Robot Projects
- Budget for consumables that expire in storage: encoder batteries, harmonic-drive/gearbox lubricant, cabinet filters, and cabinet cooling fans.
- Photograph and label every connector before the first disassembly; used machines arrive with prior owners' modifications already in place.
- Confirm the teach pendant, arm cables, and controller are a matched set. Mixed-generation combinations are a frequent cause of "communication" faults that masquerade as servo faults.
- Keep a build log with dated entries of every wire changed. When you eventually call support, the log is the deliverable.
- Independent used-robot dealers and third-party integrators are viable sources for spare arms, cables, and pendants when OEM lead times are long — verify the controller generation matches before purchase.
Why won't my Staubli RX60 arm power turn on after E-stops are fixed?
Arm power is the end of a chain: E-stop, guard interlocks, safety-relay reset, contactor feedback (EDM), operating mode, enabling device, encoder/calibration state, and amplifier health. Fixing E-stops alone leaves the other monitored loops open. Watch the power contactors during a power-on attempt — never pulling in points upstream to the permissive chain, pulling in then dropping out points downstream to the amplifier or feedback path.
Can I jumper the safety inputs to test the servos?
No. Dual-channel safety inputs are cross-monitored, so a single jumper across both channels is commonly diagnosed as a short-circuit fault and blocks power anyway. Restore each loop with real rated devices and validate the cell against your risk assessment before enabling automatic mode.
Do I need the original teach pendant to enable arm power?
In manual/teach mode, yes — the enabling (dead-man) switch must be held at its middle position, and both releasing and fully squeezing it remove power. Source a pendant that matches your controller generation rather than fabricating a bypass.
What happens if the encoder backup battery died in storage?
Absolute joint position data can be lost, and the controller will typically refuse normal arm power until calibration is restored. Replace the batteries as standard practice on any machine of unknown storage history and follow the manufacturer's calibration restoration procedure using the arm's documented calibration values.
When is it worth paying for OEM service instead of continuing in-house?
Escalate on amplifier or controller board failures, lost calibration with no documentation, missing system software, or when you have no electrical schematics. Set a sunk-time limit in advance — for example 40 engineering hours — and prepare an error-history and test-results package before the call to shorten the visit.
