Mitsubishi MDS Drive Alarm 33 (Overvoltage) — Causes, Diagnosis & Fixes

Alarm 33 on a Mitsubishi MDS drive means DC bus overvoltage — the voltage on the drive's internal PN bus has exceeded its limit (around 410–430 V DC on 200 V class units). It applies across the MDS family: MDS-A/B, MDS-C1, MDS-D/DH, MDS-D2/DH2 and MDS-E/EH servo and spindle units, shown on the drive's LED display and as ‘S01/S03 servo alarm 0033’ on the M60/M64/M70/M80 screen. The drive trips to protect its capacitors and IGBTs. Here is what actually causes it and how to find the culprit — in the order we check on the repair bench.

What is happening

When a servo or spindle motor decelerates, it acts as a generator and pumps energy back into the DC bus. That energy must go somewhere: into a regeneration resistor (smaller MDS units) or back into the mains through a power-regeneration converter / power supply unit (MDS-C1-CV, MDS-D-CV, larger systems). If the regeneration path cannot absorb the energy — or the incoming mains voltage is already too high — the bus voltage climbs until the drive trips alarm 33.

Cause 1 — Regeneration path faulty (most common)

• Regen resistor open or wrong value: power down, lock out, wait for the bus to discharge (5+ minutes), then measure the resistor — an open circuit or drifted value means every deceleration has nowhere to dump energy. Check the resistor cable and connector for burn marks.
• Regen transistor / control circuit failed inside the drive: if the resistor is good but never gets warm and the alarm comes on every deceleration, the drive's internal regen switching stage has failed — a board-level repair.
• Power supply unit (CV unit) fault: on power-regenerative systems, alarm 33 on the CV unit (or appearing on all axes at once) points to the converter itself — failed regen IGBT module, gate drive or bus sensing. This is the classic case we see on MDS-C1-CV and MDS-D-CV units.

Cause 2 — Incoming power problems

• Mains voltage too high: measure L1-L2-L3. Above ~253 V AC line-to-line (200 V class) leaves almost no regeneration headroom — common in Indian industrial estates at night when plant load drops. A step-down / buck transformer or stabilizer fixes it permanently.
• Phase imbalance or missing phase stresses the converter and can push the bus high during regen.
• Power factor correction capacitors on the same feeder switching in/out cause voltage spikes — check whether the alarm coincides with compressor or PFC panel switching.

Cause 3 — Application / parameter side

• Deceleration too aggressive: if the alarm appears only on rapid stops or spindle deceleration from top speed, the regen path may simply be undersized for the duty. Lengthening the spindle deceleration time or reducing rapid deceleration in parameters gives the bus time to bleed off.
• Vertical axis without proper counterbalance: a falling Z axis regenerates continuously — a failing counterbalance cylinder or brake shows up as alarm 33 on that axis's drive.
• Heavy new workpiece / fixture inertia after a job change raises regenerated energy beyond what the system was sized for.

Mitsubishi's official investigation sequence (from the maintenance manual)

For resistor-regeneration units, the manual's own six checks — useful as a strict order:

1. External or built-in regenerative resistor? Built-in → check the P–D short wire: not connected → connect it; connector disconnected or contact-faulted → reseat/replace. (A missing P–D jumper after a drive swap is a textbook cause.)
2. External → is the resistor/drive combination appropriate? Wrong combination → fit the correct regenerative resistor for this drive model.
3. Is the resistor/cable connection correct? Incorrect → rewire per the diagram.
4. Disconnect the resistor terminal and measure with a tester: resistor broken or resistance too high → replace the resistor; cable broken → replace the cable.
5. At accel/decel, does the speed overshoot reach the current limit? Yes → the time constant is too short: increase the acceleration/deceleration time constant.
6. All of the above normal → replace (repair) the drive unit.

Diagnosis checklist (in order)

1. Read which unit shows the alarm — the CV/power supply unit or an individual axis drive — and whether other axes alarm simultaneously.
2. Measure incoming mains at the drive terminals, all three phases, machine idle and running.
3. Power down, discharge, and measure the regeneration resistor and its wiring (P–D jumper on built-in configurations).
4. Inspect the CV unit and drive for bulged capacitors, burnt smell, dust bridging on the bus bars — conductive dust across the PN bus is a real cause of erratic overvoltage trips on shop-floor machines.
5. If mains and resistor are good: swap-test with a known-good drive if available, or send the drive and CV unit for bench testing — regen stage faults cannot be confirmed from outside.

Can you keep running?

Occasional alarm 33 on aggressive decel is a warning; frequent tripping is not safe to ignore — the failure mode of an unabsorbed regen event is a blown converter module or main-circuit capacitors, which turns a small repair into a large one.

Related guides

• Mitsubishi MDS alarm 30 (over-regeneration)
• Mitsubishi MDS alarm 31 (overspeed)
• Mitsubishi M70 / M80 alarm list
• MELDAS M64 backup & restore procedure