What are good motors and what are bad? In general I mean at least since QC varies
Hard to say without opening the motor and looking how the sense wires are routed. Having the gnd and temp sense wire run next to each other is what you want.
5 Likes
GOLD
would twisting these two pairs help? shielding?
2 Likes
Wouldn’t be the issue solved in just taking the middle value of the last 5-10sec input temp values? I mean the temp value is not a super time critical value. It’s not like the motor temp will rise 20 degrees in 1 second and even if the motor temp would rise that fast the 80/100 degree cut off limit is still so low that every quality motor should be able to handle a 10 second overheating of maybe 110 degrees.
2 Likes
That’s already what is done. Do some reading on aliasing and niquist frequency if you want to better understand the sampling issue, but it breaks down to a continuous time sample being taken discreetly. So with a coupled noise it’s possible to always sample at the incorrect interval to gather biased data even when averaging the discreet samples.
Twisting the motor sense wires wouldn’t hurt anything but probably shouldn’t be needed. I think it’s just when you have this kind of worst case routing where the wires are encircling the coil of the motor. This is literally how transformers are made.
8 Likes
So am I good if i use sensorless motors or what? This thread is way above my head.
Just don’t use the temp sensors if you worry.
You still can use your hall sensors, no problem there
2 Likes
How do I not use temp sensors? Can I disable it in the vesc tool?
Disconnect the temp wire from your jst plug.
Unfortunately there is no way to disable only the temp sensor via sw.
1 Like
this is what I presently do…
depin it at the JST
I’ve heard others use the No Limits version of the software and set the cut-offs to some ridiculously high number, but I myself have not tried that option
Don’t know if this has been mentioned here or in other threads or not, but I think safety is good enough of a reason to warrant repeating information:
-
@kevingraehl discovered a bug on his Unity that I later reproduced on a Trampa Vesc 6 running FW 5.1 where connecting a UART device such as a Metr module causes a remote cutout. That’s the observed behavior, what it probably is is a short reboot or something. @ducktaperules says it’s because of “safe start” turned on. So if you have a Unity, use some glue or something to make sure your bluetooth module doesn’t disconnect & reconnect mid ride!
-
Many people (including myself) have discovered that having the “Maximum duty cycle” and “Duty cycle current limit start” values equal will cause the motors to abruptly stop once they reach this value.
Found on FW 5.1, according to @DerelictRobot this also affects FW 5.2. If you use the “duty cycle current limit start” feature, make sure to test it on the bench before you ride! Having it at 100% disables it I guess, but then you have to be prepared for a sudden loss of acceleration when you reach top speed, just like you would before.
Sorry for tagging you and putting words in your mouths guys, I hope my language was vague enough that you can step in and clarify anything I got wrong.
16 Likes
Good write up! I hit the max duty cycle error you mentioned. It was not fun. Good to see some diagnosis on it.
When I set it to 100% it caused my motors to get really wobbly like they were oscillating back and forth at full throttle, like they were out of sync or something. Interested in seeing if anyone else has the effect.
6 Likes
Sorry to revive this dead ass thread but I have this oscillating at 85-95% duty cycle on Acks3.103 (Sensored, FOC) and Vedder 3.62/65 (Sensored, FOC) and have heard that it is due to the default sampling rate of 30khz and potential solution is raising it to 35khz where you’re now susceptible to the watchdog error/fault from the MCU running out of resources. If you ever found a solution Justin let me know.
1 Like
I never got a follow up solution I found acceptable.
I got told to limit it to 85% and had poor top end speed/acceleration control. The board felt sluggish compared to normal.
Bumping the duty limit start number up to a value that felt normal speed-wise gave me the weird motor wobbling.
I got told to just leave it at 85% then and what I was feeling was “normal” 
2 Likes
We can’t be the only ones with this problem though, every single one of my dual motor builds does this.
2 Likes
Agreed. I saw a couple of other people mentioning it but no pinpoint on why it’s happening to some and not others.
2 Likes
Ha… you got wobbling, i got motors cutting out at completely different speeds. Next thing i know my back wheels start turning when i accelerate further. 85% seems a good balance in the 60’s down shitty roads
@itsrow any topics i can read on the 35khz problem? It it a erpm issue?
it’s not dead. it will live forever! but low activity is good.
It’s been int he back of my head to get my head round the problem that @rusins wrote up that I presume you’re also experiencing and whatever the whole set duty cycle limit start lower advice is about. ie to know what it was and when in the versions history it was introduced and what conditions triggers it.
as it stands each time I’ve read it I’ve gotten confused and didn’t have enough time to focus on understanding it better.
It would help if that topic had it’s own thread where that stuff got worked out.
And then I’d update the first post with a succinct summary.
4 Likes
No not an ERPM issue, providing you weren’t having ERPM related issues beforehand. When you increase the sampling rate the STM32 has to physically refresh more time per second and if you run out of processing power when youre riding the esk8 it will fault and cut motors completely which is arguably worse than oscillating/inconsistent power delivery.
2 Likes