Variateur bloqué après arrêt d'urgence — nécessite une coupure secteur

Most modern drives require STO to be re-asserted (24 V present on both STO channels) before the fault reset pulse. If your E-stop circuit cuts STO and the PLC issues the reset while STO is still low, the drive ignores the reset — or worse, latches a secondary STO fault on top of the original one.

Check: Measure STO terminals (typically X40/X41 on Siemens, SX1/SX2 on B&R) with a multimeter while pressing reset. Both channels must read 24 VDC. If they're low, the E-stop relay hasn't re-closed — fix the reset sequence in your safety PLC or relay logic first.

Fix in PLC: Add a 50–200 ms delay between "STO confirmed active" and the fault reset rising edge:

// Wait for STO channel OK before issuing fault reset
tonSTODelay(IN := xSTO_OK AND xResetRequest, PT := T#150ms);
xDriveFaultReset := tonSTODelay.Q;  // Rising edge → drive control word bit 7

Many drives require the fault reset input to be held for a minimum time (50–500 ms depending on make). A PLC output that goes high for one scan (1–4 ms) is often too short. This is especially common after copy-pasting reset logic from a different project with a different PLC scan time.

Check: Look at the drive's digital input in the drive scope/oscilloscope. If the reset pulse doesn't appear at all, or is a thin spike, that's your problem.

Fix: Use a timer to hold the reset output for at least 300 ms:

// Extend reset pulse to 300 ms regardless of PLC scan time
tonResetPulse(IN := xResetRequest AND NOT xDriveFaultReset, PT := T#300ms);
xDriveFaultReset := tonResetPulse.Q OR xResetRequest AND (tonResetPulse.ET < T#300ms);

Alternatively use a dedicated R_TRIG + TON monoflop pattern to generate a clean 300 ms pulse on each rising edge of the reset button.

Almost every modern servo drive (Siemens, B&R, Beckhoff, SEW, Lenze…) uses the CiA 402 / DS402 state machine over EtherCAT or PROFINET. A fault reset is not simply setting a bit — you must transition through specific states in the correct order. Skipping states causes the drive to stay faulted.

Required control word sequence after E-stop:

// Step 1 — Issue Fault Reset (bit 7 rising edge)
ControlWord := 16#0080;   // only bit 7 set
WAIT 100ms;

// Step 2 — Return to Switch On Disabled  
ControlWord := 16#0000;
WAIT 50ms;

// Step 3 — Shutdown (transition to Ready To Switch On)
ControlWord := 16#0006;
WAIT 50ms;

// Step 4 — Switch On
ControlWord := 16#0007;
WAIT 50ms;

// Step 5 — Enable Operation
ControlWord := 16#000F;

// Check StatusWord bit 9 (Remote) and bit 2 (Operation Enabled) = 1
// before commanding motion.

Many beginner implementations jump straight to 0x000F after fault, skipping the intermediate states — the drive ignores it and appears "stuck".

If the E-stop caused a hard mechanical stop (motor forced to stop while spinning), the regenerated energy dumps into the DC bus. If the bus voltage is above the drive's reset threshold, the fault cannot clear until the bus discharges — which takes 3–30 seconds depending on capacitor size. A power cycle drains the bus through the bleed resistor, which is why it "works".

Check: After an E-stop, wait 30 full seconds before pressing reset. If the fault clears, this is your problem. Monitor DC bus voltage in the drive diagnostics.

Fix:

• Add a reset delay of 10–30 s after E-stop in your PLC logic before attempting reset.
• If the delay is unacceptable, add an external braking resistor sized for the worst-case regenerated energy (E = ½ × J × ω²).
• Check the internal bleed resistor hasn't failed (measure resistance between DC+ and DC- with AC off and caps discharged — should match drive spec).

When E-stop is configured as Category 0 (coast — immediate removal of power), the motor freewheels. If the load has high inertia and overshoots the encoder position significantly, a following error or encoder overspeed secondary fault stacks on top — and this secondary fault has a higher severity level that cannot be reset without power cycle.

If E-stop is Category 1 (controlled decel before power removal), the drive needs to stay enabled long enough to decel — but if STO fires before the motor stops, it still coasts and stacks the secondary fault.

Fix: In the drive parameters, set the E-stop / STO reaction to "controlled stop" (Category 1) if your safety category allows it. In the PLC, add a speed-zero interlock before cutting STO:

// Don't cut STO until motor is actually at speed zero
// Prevents secondary following-error fault during coast
xSTO_Safe_To_Cut := (ABS(rActualVelocity) < 5.0)  // rpm or mm/s
                    OR tonSTOTimeout.Q;              // safety fallback max 3 s
tonSTOTimeout(IN := xEStopActive, PT := T#3s);

Absolute multiturn encoders (Heidenhain, Sick Stegmann, Tamagawa) keep counting via a battery while the drive is off. If the encoder battery is low, the drive raises a non-resettable battery fault — not the E-stop fault — but it appears at the same time, causing confusion. A power cycle clears it temporarily because the drive re-reads the encoder with a fresh supply.

Check: Read the full fault code, not just the fault LED. Battery faults are usually coded separately (e.g. F07.0035 on Siemens, 29056 on B&R). The battery voltage is also typically visible in the drive diagnostic parameters.

Fix: Replace the encoder battery (CR2032 or specific pack). After replacement, perform a reference run / homing cycle before re-enabling.

When the E-stop cuts 24 V to the panel, if the PLC or IPC also loses power (even for 50 ms), the fieldbus drops. The drive raises a bus communication fault (typically higher severity than a normal E-stop fault) which requires power cycle. The operator sees "E-stop fault" but the real fault is the comms loss.

Check: Look at the drive's fault history — is there a comms/bus fault timestamped a few milliseconds after the E-stop fault? That sequence confirms this cause.

Fix:

• Give the PLC/IPC and the fieldbus switch an uninterruptible 24 V supply (separate rail from the E-stop circuit, or UPS).
• In the drive parameters, set the bus-fault reaction to a lower severity (e.g. "controlled stop" instead of "immediate fault + power cycle required").
• On Siemens: p2040 (bus fault reaction) — set to 0 (coast) or 1 (ramp) instead of 3 (fault latch).

Several firmware versions have documented bugs where a specific fault combination cannot be cleared via software reset. This is more common on drives that have been in service for years without firmware updates.

Known affected versions (non-exhaustive):

Fix: Check the manufacturer's release notes for your firmware version. Always update drives in pairs (drive + motor firmware must match).

Some installations wire the drive's hardware reset input (DI for "Fault Reset") to a Normally-Open contact on the E-stop relay. When E-stop fires, the relay drops — the NO contact opens — and the reset input goes low, which is fine. But when the operator releases the E-stop mushroom, if the relay is wired in an unusual way and the reset DI is still seen as "active" (input = 0 V where the drive expects 24 V = reset), the reset never happens.

Check: Measure the drive's DI reset input terminal when pressing the HMI reset button. Should momentarily go to 24 V. If it never changes, trace the wiring back.

Fix: Rewire to a separate momentary NO pushbutton or PLC digital output dedicated to fault reset — not shared with any E-stop circuit element.