A servo motor rebuilt for a Vickers/Selema 70-DBC-III-F18 drive on an OSAI-controlled CMS router has lost its resolver-to-rotor phasing. The drive commutates from the angle the resolver reports. If the resolver stator or rotor was removed, re-seated, or replaced during the rebuild, that angle no longer matches the rotor magnet poles. Selema documented the drive-assisted phasing procedure through the F1PC DOS utility over the drive's RS-422 port. That software is hard to find. The physical relationship it sets can still be restored with a scope and a reference motor.
Commutation angle error and torque per amp
One quantity decides this case: the electrical angle error between where the drive thinks the rotor flux is and where it actually is. The drive places its current vector 90 electrical degrees ahead of the resolver-reported flux position. With an offset error θ, only the cos θ component of phase current produces torque. The rest heats the windings and does no work. Electrical angle equals mechanical angle times the motor's pole pairs. On an 8-pole motor, a 5° mechanical slip of the resolver housing is a 20° electrical error.
| Electrical error θ | Torque per amp (cos θ) | Current for same torque | Copper loss for same torque | Field behavior |
|---|---|---|---|---|
| 0° | 1.00 | 1.00× | 1.00× | Normal |
| 15° | 0.97 | 1.04× | 1.07× | Hard to detect by feel |
| 30° | 0.87 | 1.15× | 1.33× | Runs warm, earlier I²t trips |
| 45° | 0.71 | 1.41× | 2.0× | Overheats, current-limit hits on accel |
| 60° | 0.50 | 2.0× | 4.0× | Weak axis, following-error faults |
| 90° | 0 | — | — | Full current, no torque; shaft locks or buzzes |
| >90° | negative | — | — | Positive feedback: runaway or violent oscillation on enable |
The values are cos θ, 1/cos θ, and 1/cos² θ, assuming constant winding resistance. Small errors appear as heat. Large errors appear as control instability.
Heat faults versus logic faults after the rebuild
Before touching the resolver, confirm the symptom is commutation and not wiring or feedback loss. A phasing error produces the same current regardless of direction problems. A feedback or wiring fault usually trips at enable before any load is applied.
| Observation | Class | Likely cause | Where to read / measure |
|---|---|---|---|
| Fault at enable, no motion | Logic | Resolver cable, excitation loss, swapped sin/cos pair | Drive fault display; resolver winding resistance at connector |
| Shaft jumps or runs away on enable | Commutation, >90° | Resolver 180° off, or motor phase sequence swapped | Back-EMF sequence U-V-W vs resolver direction |
| Holds position but motor gets hot at light load | Heat, 20–60° | Resolver offset shifted during rebuild | Drive current monitor at standstill hold vs sister axis |
| Stronger in one direction than the other | Commutation | Offset error combined with reluctance or cogging effects | Current during constant-speed jog, both directions |
| Hot even with correct phasing | Heat, mechanical | Binding bearings, brake dragging, winding damage from rebuild | Hand-turn torque with motor disconnected; insulation resistance test |
Three routes to restore resolver phasing
| Method | Needs | Accuracy | Risk |
|---|---|---|---|
| Back-EMF vs resolver comparison | Two-channel scope, a way to spin the shaft, resolver excitation, an identical sister motor for reference | A few electrical degrees | Low; motor stays disconnected from drive power |
| DC rotor lock, then set resolver | Current-limited DC bench supply, scope on resolver outputs | Limited by friction and cogging | Winding heating if current is not limited |
If F1PC cannot be obtained, use the back-EMF method with a known-good sister motor of the same type from another axis as the reference. The drive's stored offset was correct for the original motor. Reproducing the original mechanical relationship between back-EMF and resolver output restores commutation without writing any parameter to the drive. The sister motor provides that relationship directly, so no Selema phasing convention needs to be known.
Back-EMF phasing procedure without the software
- Lock out the axis. Disconnect the motor power leads U, V, and W from the drive. Uncouple the load, or support any vertical axis mechanically before releasing the brake. Do not back-drive the motor while its power leads are connected to the drive, because back-EMF feeds the drive's DC bus through the output stage.
- Excite the resolver reference winding. You can leave the resolver connected to a drive with control power on and the power stage disabled. Alternatively, drive the reference winding from a signal generator at the excitation frequency and amplitude given in the drive or motor documentation.
- Connect scope channel 1 to motor line-to-line voltage, U-V. Connect channel 2 to the resolver sine output. The sine output is carrier-modulated, so trigger on channel 1 and read the envelope of channel 2.
- Spin the sister motor slowly and steadily in the positive machine direction. Record the phase relationship between the U-V zero crossing and the resolver sine envelope zero crossing, and the direction in which the envelope phase changes.
- Repeat step 4 on the rebuilt motor. Loosen the resolver stator clamps and rotate the housing until the relationship matches the sister motor. Remember that one mechanical degree equals the pole-pair count in electrical degrees.
- Tighten the clamps and scribe a witness mark across the resolver housing and the motor end bell. Recheck the waveform after tightening, because clamps often drag the housing a few degrees.
- Check phase sequence. The U-V-W back-EMF order must match the sister motor for the same rotation direction. If it does not, correct the motor lead order or the resolver sin/cos pairing before powering up.
The DC-lock method follows the same idea. Pass a limited DC current into U and out of V and W tied together, with the current kept well below the nameplate continuous rating. The rotor settles at an electrical zero. Rotate the resolver housing until its sin/cos outputs match the reading from the sister motor under the same lock. Friction and cogging make this method coarser than the back-EMF method.
Running F1PC if a copy turns up
- Use a period laptop or an industrial PC with a native UART. USB serial adapters often fail with DOS software that accesses COM port hardware directly.
- Wire the RS-422 link following the
SELINT-1documentation. Confirm the transmit and receive pair polarity with a scope before connecting. - Before running any phasing routine, upload and save the drive's full parameter set. On older drives, a failed download can leave the axis unconfigured.
- Run the phasing routine with the load uncoupled, following the
F1PCversion's own instructions. Record the new offset value, then save the complete parameter file again.
Confirming phasing under load
Reconnect the motor and set the drive's current limit low for the first enable. Jog slowly from the OSAI control and watch for smooth rotation in both directions. Then compare the rebuilt axis with the sister axis:
- Holding current at standstill should be within a few percent of the sister axis under the same load.
- Current during constant-speed jog should be symmetric between the two directions.
- Motor case temperature after a normal duty cycle should track the sister motor.
- Following error on the OSAI axis diagnostics should match pre-rebuild behavior during rapid moves.
If current is symmetric but high on both axes, the cause is mechanical or thermal, not phasing.
FAQ
How do I tell if a rebuilt servo motor's resolver is out of phase?
Compare standstill holding current and constant-speed jog current against an identical axis. Higher current for the same load, or current that differs by direction, points to commutation error. A jump or runaway on enable means the error is beyond 90 electrical degrees.
How do I align a resolver without the F1PC software?
Disconnect the motor power leads and spin the shaft. Scope U-V back-EMF against the resolver sine envelope, then rotate the resolver housing until the phase relationship matches an identical sister motor. Because the drive's stored offset is left unchanged, this restores the original commutation.
How do I connect F1PC to a DBC III drive?
SELINT-1 documentation. Back up all drive parameters before starting the phasing routine.
When should I stop and send the motor or drive out for service?
Stop if the motor still runs hot with symmetric current after phasing, if insulation or winding resistance tests fail, or if the drive faults at enable with a verified resolver. At that point, contact CMS for the machine or an authorized Selema/Vickers servo repair service, and send the motor together with a sister motor or the original resolver position data.