Anyone Experience Thermal Throttling with Trampa VESC 6 mk3?

#1

So yesterday, I took my scooter out for a ride around my neighborhood, giving it a decent amount of throttle for about 30 minutes.

On my way back, the throttle response started sputtering at higher throttles, at first I thought the parallel connections of my battery had broken (again), but this was not the case.

After checking a few others things, I noticed that the heatsink of my Tampa VESC 6+ and my 63100 motor from Saite were both too hot to touch, so I ultimately ruled the sputtering throttle response as due to thermal throttling on the VESC.
It is not thermal throttling on the motor because I have that disabled on my VESC, and I did not have this issue with a Flipsky 6.6 running the exact same settings (maybe I screwed up and the VESC is throttling off the motor temperatures).

I took my scooter out for a ride later after everything cool down and the sputtering was gone.

So an interesting thing is, when I ran my scooter with a Flipsky 6.6 ESC, I never had this problem.

I run a 10S battery, average voltage was 38-39. Motor amps 80A, battery amps 40A, the sputtering happened under heavy throttle. Some back of the hand calculations with the DirectFETs show that switching losses and conduction losses are about equal at 40A, with conduction losses dominating above 40A motor current. It’s 4W per phase leg, or 12W in total. The thermal resistance of the heatsink is around 10C°/W based off the surface area and the thermal conductivity of aluminum. It sits on top of a piece of polyurethane foam.

I was wondering if anyone else has encountered this phenomenon, and if not, if someone could show me what I am doing wrong.

The Flipsky 6.6 uses MOSFETs with a better figure of merit (Rdson * gate switching charge), so they dissipate around 60% of the heat in similar operating conditions, if my estimates are correct, and would likely explain why I never encountered this issue previously with the Flipsky ESC.

Or maybe I screwed up the motor temperature settings.
Or I just need to not put this thing on top of a THERMALLY INSULATING piece of foam.

1 Like
#2

40A battery max shouldn’t be a big problem for all kind of hw6.xx based vescs.
Thermal throttling is usually a reduction of output power which means you notice it in form of not being able to accelerate as strong and fast like usual.
If I understood in your case it was more like a sputtering.

To be sure I would just check the temp data from your vesc. Either via bt module or just via live data connected to the vesc tool after a similar ride.

1 Like
#3

So, I think I had already exceeded the soft cutoff, and was already getting a reduction in output power, and I was actually hitting the hard cutoff when I experienced this.

It is the current through the motor and battery voltage that determines how much heat is dissipated in the ESC.

It cooled down relatively quickly when I stopped riding, so I might need some kind of BT module to get to the bottom of it.

1 Like
#4

If you’re trying to troubleshoot without logging, you’re gonna have a bad time.

4 Likes
#5

I have a few HC-05 modules lying around, will those work?

1 Like
#6

I found those to be a diceroll depending on the exact variant, but it should?

1 Like
#7

wait, crap, so I’m using twist throttle (adc input), with a brake button, this directly conflicts with the UART pinout…

#8

How many more curve balls you got hiding there? Hah

Sounds like you need to remove variables and isolate your testing. Analog input is highly susceptible to noise- how confident are you that you’re not introducing noise at higher duty cycle, signal EMI, etc.

There are other ways to drive throttle if you need to remove that from the equation.

4 Likes
#9

Motor resistance value off, due to hot motor.
Causes cogging… That’s why we monitor the motor temp. A hot motor simply has a higher resistance. Resistance needs to be ±5% accurate. Once the resistance drifts beyond that scope, trouble is ahead.

#10

So it wasn’t cogging with a very hot motor (I actually did run into that issue with a flipsky ESC, but not the Trampa VESC)

But you are correct that the resistance being off causes less efficient operation of the motor and the ESC, so that might play a part into what I encountered.

#11

It’s usually fine unless we are talking about startup cogging. Thermal thottling doesn’t present as jerky. Think about it logically the thermal time constant of the system is far too high, it’s not like the temperature is jumping around wildly effecting the torque.

1 Like
#12

Can also be a cooked motor due to thermal stress.

2 Likes
#13

We have experienced it. A hot motor can defo cause cogging.

1 Like
#14

Yeah, I’ve seen it before as well but I think it is less common. My money would be on that adc throttle getting corrupted, if it moves close to the phase leads it can couple in noise to the throttle very easily depending on the impedance of the throttle pot.

2 Likes
#15

Troubleshooting is always fun… and you always learn from it.

5 Likes
#16

Sometimes the tolerances ad up, sometimes they work for you. If the VESC measures the resistance close to the upper tolerance, then you might get cogging when winter arrives but no cogging on hot motors. If we hit the lower tolerance, a hot motor all of a sudden starts to cogging. It might be worth having a slider to adjust the resistance ±5% to fine tune the resistance. First test we do: Step on the board an give it full beans from stand still. If it coggs, we lower the resistance a bit and do the same test again. Some users have a hard time to do the maths and enter a wrong resistance value. That can ultimately kill your ESC. A fine tuning slider would help.

5 Likes
#17

So that I don’t derail too much, I feel like I’ve had similar experiences. My direct drives under heavy load after a while can start acting a bit funny, giving limited throttle response (I know due to throttling) but also some quircky cogging sometimes or sudden drops are changes in throttle … I usually take it easy from there and let things cool down a bit and things return to normal. If resistance due to variances in heat and motor detection is the issue… my (derail) question for @Trampa and @Deodand is what is the best temperature to run motor detection at? Say my motors range from 45-75C on average during a ride… should I measure them at room temperature, at 45C, somewhere in between average, or (probably not) max temp? You said vesc firmware compensates for heat to an extent, does it compensate for cooling if we were to measure when they are at a higher temp?

Also, for firmware changes… would it make any sense at all to re-measure resistance real quick automatically when say stopped at a stop light and the delta between initial detection temps resistance and current temperature has surpassed a certain point to reset the baseline and re-establish resistance accurately for higher temp motors mid ride… or maybe store a range of detected values throughout the temp range and compensate based off detected values at a range of temps? … almost like for your motor detection you do an initial motor detection then ride and measure at various temps to establish detection values, and store those values to not have to keep measuring resistance all the time… have like a smooth transition between all detected values, and dynamically calculated in-between values that are used in real time as temps drop and rise… just spit balling here since I’ve had odd experiences when detecting/riding and vice versa at cold temps/hot temps, etc. this may be a useless suggestion…

3 Likes
#18

If you run hubs, you can try to figure out the upper tolerance of the resistance. You run detection and check if you see cogging on a full throttle acceleration from stand still. You must apply full throttle to check for that.
Now you can increase the resistance a bit and see if everything is still good. There will be a point where you get cogging on a cold motor. You want to be a bit below that point.
Alway stay within ±5% of the original detection result!
There is a feature in VESC-Tool to compensate motor temperature. Itis called Temerature Compensation. It doesn’t always work for each motor the same so by default it is switched off. YOu can try that if you like.

3 Likes