TP100 star rewind — Hall sensors working after pinout fix, two separate issues: ABS overcurrent threshold and observer instability at high ERPM

Following up on my earlier question about the TP100 star rewind FOC cutout issue. The root cause turned out to be a wiring mistake — the SS41F Hall sensor pinout (VCC/GND) was reversed throughout testing. Once corrected, VESC Tool confirmed valid Hall tables and the motor now spins freely past the previous cutout threshold.
Current setup:

Motor: TP100, delta350KV → star rewind, targeting ~200KV
ESC: Flipsky 75350, VESC FW 7.0
Battery: 14S LiPo

Background:

The motor ran without issues in its original delta winding (~350KV). Problems appeared only after rewinding to star — initially no Hall sensors, sensorless FOC was cutting out above ~20% duty due to low BEMF in star configuration. Hall sensors (SS41F) were installed to solve the low-speed observer instability.
Progress after pinout fix (log attached):

With Hall sensors correctly wired, the motor reaches ~28,000 ERPM. Through testing we’ve identified two separate issues:
Issue 1 — ABS overcurrent threshold:

FAULT_CODE_ABS_OVER_CURRENT (fault=4) triggers consistently at 101A+ phase current. At 100A limit, clean runs with no ABS fault. The boundary is very sharp — 100A is fine, 101A cuts out. Log shows 141 fault=4 events from earlier runs before the current limit was dialed in.
Issue 2 — Observer instability at Hall/sensorless transition:

Even at 100A limit (no ABS fault), the motor cuts out intermittently at high ERPM. Log shows RPM climbing smoothly to ~27,500 ERPM, then dropping to ~3,700 in a single sample before coasting to a stop — no fault code, just observer losing sync at the Hall/sensorless handoff region.
These are two distinct failure modes — the ABS threshold is a hard current limit issue, the high-ERPM dropout is a sensorless observer issue.
Additional note:

One Hall sensor was found to be angularly misaligned (shifted one slot) during the original installation. Pinout is now correct, alignment is also corrected — attached log predates the alignment fix, so next test will be with both corrections in place.
Questions:

The 100A/101A ABS cutout boundary — is this expected behavior for this motor/ESC combo on 14S, or is there a way to push the threshold higher without triggering overcurrent protection?
For the observer dropout at ~27,500 ERPM — is this a Hall handoff threshold setting or Observer Gain tuning issue on a 12N10P star config?
Could the previously misaligned Hall sensor have left residual errors in the Hall table that affect the sensorless transition even after correction?

The desync and the over current errors are the same issue.
Desync causes the over current spikes.

Other than that I have no experience with your setup.

1 Like

Thanks for the input! That makes sense — if the desync is the root cause, then fixing the observer stability should resolve both issues.

Do you have any recommendations for tuning the observer gain or the Hall/sensorless handoff threshold on VESC to prevent the desync in the first place?