Hi everyone,
I'm using VESC 75V 300A version from Trampa Boards to drive Motenergy ME0907 motor in sensored mode. This is used for an electric vehicle. We are testing VESC as a replacement for Kelly KLS controller. VESC works perfectly and in pretty much every way it's better than Kelly, however there is one problem - insufficient torque at low RPM which is a deal breaker. Kelly provides higher torque.
Is there any VESC settings that I could tune to fix this issue? Thanks
Increase the motor and/or battery current.
NextGen FOC High voltage 144v/34s, 30kw (https://vesc-project.com/node/1477)
Torque is pretty much only Amp related. What is the input method? CAN, PPM, UART?
Thanks for your answers, I already increased it to the maximum (450A I think). It doesn't let me increase it more. Input method is CAN.
Please note the issue here is the torque only at low RPM. As soon as the RPM goes beyond 500, the torque gets sufficient. The motor is operating in sensored more across all range.
Is there any parameters that might affect the torque specifically at low RPM? Kelly controller doesn't have this issue, so I assume it could be fixed with some parameter changes in VESC.
Maybe the motor is not dialed in very well. You could try and lower the time constant fom 1000uS to 500uS, double the observer gain or quadruple.
After these changes you should go way down in amps and observe the results in the RT-Data window. Open RT Data FOC Tab and observe Q-Currents
If you see big ripples or spikes, let go the throttle as fas as you can and re-adjust values. Go up in amps bit by bit and observe after each step up.
Another thing you can do to increase the torque is using an AS5047P encoder instead of Halls. The cable from the Encoder to the VESC 75/300 should be shielded and short (10 to 15cm).
The sensor should be operated in 3.3V mode.
https://www.mouser.de/ProductDetail/ams/AS5047P-TS_EK_AB?qs=Rt6VE0PE%2FO...
https://trampaboards.com/resources/manuals/207.pdf
Thanks a lot Frank, appreciate your help! I'll try that and will get back with the results.
Meanwhile one question - why using AS5047P can be advantageous for torque over Hall sensors? Thanks
AS5047P will create 100% torque from stand still due to 0.2° angle accuracy versus 7° for Halls.
Thank you for the explanation. Considering that Kelly produces sufficient torque with the same Hall sensors I guess the problem is somewhere else, and I'm trying to figure it out now, will be back with an update soon. Meanwhile there's a separate issue that I can see now related to performance decreasing in time, which I'll describe in a separate topic.
So this issue boiled down to two items:
1. The Hall sensors required tuning. I tuned those for another motor of the same model and assumed it would work on others, which was a wrong assumption. After re-tuning the torque increased, but still was somewhat lower than Kelly's.
2. Motor current measurement showed that Kelly produced up to 450A of current (although it's rated at 350A). VESC was able to produce 400A max (rated at 300A). Thus some difference in torque. This is something I can't fix.
Anyway I assume the problem is solved now. Thanks everyone for your help.
You would probably improve your low-speed high-torque performance even further by using an encoder.
The VESC seems to have a lot of trouble with interference/noise in the hall signals, which the Kelly controller easily filters out. So on the exact same sensored motor the vesc will not be as smooth or powerful as a kelly esc with similar specifications. Use an encoder though and the vesc is the clear winner.Revisited some problem motors and realized I was using the wrong hall voltage.
Some halls also like 5V instead of 3.3V as supply. You can try to set the sensor voltage to 5V.