VESC Tuning Advice, Poor Acceleration compared to other controllers

carltoza

1 mW
Joined
Jul 22, 2022
Messages
11
Anyone with any experience using VESC controllers with DD hub motors please advise!

I am building a custom ebike with a 1500w direct drive hub motor. Before the flipsky 75100 that I am currently using, I had a Kelly KLS-S 48v controller on it. Specs on that were are sinewave FOC, 50amp cont. 120amp burst, 48v.

My issue is that the Kelly controller had way better acceleration and torque from 0 to about 20mph, when compared to the flipsky 75100.

I have done the phase angle calibration several times over and played around with a lot of the different settings in the app but nothing seems to affect the bottom end acceleration. It has a bit higher top speed than the Kelly (with the same 48v battery), my guess there is something to do with flux weaking, but it doesn't get up and go and I'm looking for more acceleration, more punch. (to anyone that says that can't be achieved with a DD hub motor on a 29in wheel I assure you it's not the case. I've had ample acceleration both with the Kelly and the square wave chinese controllers.)

I checked for resistance and heat along all the wires and have eliminated that variable. 8-12g wire wherever possible, shortest phase wires possible. I'm an experienced battery builder so the power source can certainly supply more than the motor can accept which is around 40 amps cont. 80+amps peak. from the battery data I gathered from the smart BMS while running the Kelly Controller it rarely got above 60 amps and was averaging closer to 35 amps. Motor rarely got hot. top speed was around 31 mph. yet it had great acceleration, never studdered, was always smooth off the line.

My guess is that it is something in the VESC app settings. I have cranked the amperage settings up even to 100, for both battery and motor to little effect. When you run the phase angle tuning there's even a selection for DD ebike hub motors that I used that pre sets all the motor and FOC settings...

If it's not the VESC setings, the only other thing I can think of is that the 75100 vesc, being a 72v controller with components rated for 100v isn't all that happy with the lower 48v input voltage. But that's just a wild guess and I think it's probably something in the FOC settings.

Could anyone provide any wisdom on the matter? It would be greatly appreciated.

FOC SETTINGS

zero vector frequency 25 kHz
Openloop ERPM 300
Motor Resistance 66.0 mO
Motor Inductance 74.35
Motor Flux Linkage 24.965 mWh
Current KP 0.0744
Current KI 65.97
 
Solution
Low end torque comes from phase (motor) amps.

So the Kelly was rated for 120A peak phase current and so is your flipsky? My first guess would be that the power stage (I think that is the correct term) is more robust (lower resistance) on the Kelly. I'm guessing the Kelly is larger/heavier?

Also, is the VESC utilizing hall sensors? IE: What "sensor mode" is it set to under the FOC settings?
Low end torque comes from phase (motor) amps.

So the Kelly was rated for 120A peak phase current and so is your flipsky? My first guess would be that the power stage (I think that is the correct term) is more robust (lower resistance) on the Kelly. I'm guessing the Kelly is larger/heavier?

Also, is the VESC utilizing hall sensors? IE: What "sensor mode" is it set to under the FOC settings?
 
Solution
Go to APP CFG , ADC section and Check your positive ramp time and decrease it to 0.1 or even less. And adjust your throttle curve to 15%-20% exponential. You should see acceleration improvement.
 
VESC using FOC with hall sensors and hub motors needs a little tweaking after using FOC wizard. Mainly sensorless open loop ERPM needs to be lowered to 50. Sensorless ERPM lowered to 500. Observer gain can be lowered to as low as .20, usually, it is set too high. You can also experiment with lowering the motor flux linkage, I see yours is 24.965, try lowering it to around 15.

Lowering observer and flux linkage was a tip learned from a post from forum member casainho, it helped but the biggest difference was lowering the ERPM settings I listed above. Wow, what a difference!

My application is an ekickscooter which Im pushing REALLY hard with a 16s6p pack. The motor is similar enough to an ebike hub motor that these changes I listed shoud work. I've used multiple implementations of VESC controllers starting with the original poorly designed 75/100. I experienced thermal throttling way too quickly. Moved on to lower voltage VESC versions, and thermal issues were solved but I was voltage-limited. I recently purchased an FS75/100 pro. Thermal issues have been taken care of. I'm highly impressed. Still, I was having issues with ABS overcurrent, desyncs, and poor performance before the transition from sensored to sensorless. Lowering those settings was the key. Powerful fast acceleration, incredibly quiet motor, with field weakening set as high as 70A Im getting insane speed increases. Top speed on flat land is 47-50 mph! No tailwind, no drafting behind a tractor-trailer, or holding onto a bumper..Lol! Of course, I don't go that fast all the time but it's certainly capable of doing it for several minutes at a time. Sounds unbelievable but it is what it is, GPS measured BTW. Even more unbelievable is that the scooter and its motor are an ex-rental.. You read that right. A scooter designed using a 10s battery with top speeds of 15mph max!

As Im pumping 3kw through the motor I had to mitigate heat issues. I already experienced one meltdown. I have to say, after reading the hub motor cooling method thread here I decided to open the motor sideplates up on both sides. I milled out a good portion of the side plates and man what a difference in cooling.

Anyway, hope this helps.
 
PWD is likely on the money.
Phase amps should typically be 2x - 3x that of your battery amps; try sending that to the moon.
 
Just discovered this thread, and I had the same issue with poor acceleration from 0 to 10-15mph, I fixed it and now its a f***cling rocket! Solution found here
How To Configure Motor
and here

The fix is tweaking the values from motor detection manually

  1. Observer gain (FOC-General) = HALF of wizard value the magical fix
  2. Sensorless ERPM (FOC-Hall Sensors) = 2000
  3. Sensor Interpolation ERPM (FOC-Hall Sensors) = 250
  4. Zero Vector Frequency (FOC-Advanced) = Increase up to 30 max didnt tried that
  5. Motor Resistance (FOC-General) = increase to 105 can help. - didnt tried that
  6. Matching KI- KI factor to match their now higher Motor Resistance value. didnt tried that
Hope that helps
 
Back
Top