BMS Inductive load issue

I have issues when powering an inductive load such as a motor used in a hand vacuum. Originally the vac used nickel metal cells that eventually died so I attempted a 4s 18650 replacement. I discovered that with two different BMS boards I tried, neither would start the vacuum. The vacuum motor is fine starting and running without issues when the 4s pack is connected to the motor without the BMS. Apparently, the di/dt inrush and the reverse spike that is created caused the the BMS to immediately shutdown with the motor receiving just enough current to allow the motor to spin with a very Shirley duty cycle of about 1/4 sec. I also discover that if I continued to switch the motor on and off rapidly I was able to get the BMS to allow the motor to spin up enough to sustain constant normal operation. Thats why I believe the starting issue is related to the classical di/dt current issue. This would suggest the same issue with any other inductive load. Has anyone else experienced this and discovered a way around it?

Ive tried Schottky diodes in reverse in an attempt to dampen the reverse current but its not reliable using even a PN junction diode. Looking for comment regarding similar experience and mitigation suggestions.

Thanks

Use a different BMS that is rated for the surge amps required by the motor. It is quite possible the motor is pulling 3-4x the rated amps to get it started. If you hand spin the motor and then flip the switch, it probably will work.

Schottky diodes will do nothing as they won't limit the inrush current to the motor. What you could do is put some capacitors between the motor and bms in parallel. It would charge up and hold enough power to get the motor started.

You could consider bypassing the over current sense feature of the BMS & just use a fuse?
You want a BMS to monitor individual cells, but the over current bit is just a nuisance here.
Maybe find the current sense resistor on the BMS & short it out with some copper wire?

It's a good idea to leave the diode there for reverse spike handling.
The BMS might also like some noise suppression - maybe a snubber network eg for this DC motor, 0.22uF cap & 3.3ohm 1W resistor in series across the motor terminals?

Probably related to this:
https://secondlifestorage.com/showthread.php?tid=9777&highlight=speed+controller

The ideal solution is probably some sort of soft start circuit.
Perhaps get it running with an inline resistor to limit the inrush current, then activate a bypass relay after 0.5s or so to get full power.