I will give this a try
Jeff and Vedder are hard at work trying to solve this problem
Has anyone discussed using toroids for high frequency noise suppression? This appears to becoming common when using variable frequency drives in industrial machinery and automation, to deal with the spikes… Might be necessary in conjunction with the current low-pass filtering.
Generally you use the toroidal cores around the actual wires to/from the motor and controller, with inductively suppress high frequencies very effectively. The spikes themselves are inductive I think, and shielded wires aren’t enough to stop them showing up in the motor windings, and are bad for the longevity of the motors regardless.
This is outside of my wheelhouse, so I’m simplifying and probably getting some of the jargon wrong, but, worth exploring. I know the toroids available “specifically” for this application tend to be pretty expensive from domestic suppliers, but there’s been some videos recently on the effectiveness of similar import cores that are pretty inexpensive ($20-30 for larger ones) and testing effective suppression in the frequencies necessary. Let me know if you want some links.
1 Like
to me, it seems far easier to do analog-digital conversion adjacent to the sensor itself, and convey the measured temperature to the esc over a relatively noise tolerant bus like i2c. parts like mcp9808 are inexpensive, have acceptable precision, have libraries already available (because they are commonly used by hobbyists), and are 5v tolerant.
i don’t expect everyone to share my view, since most people ride with hall sensors instead of as5047’s
4 Likes
Benefits seem clear to me, wish more motor manufacturers used it, especially since vescs have support for it.
3 Likes
So I am also running 3.65 and experiencing this fault. Hesitation and cogging around 18 mph and then this motor temp fault hits. Seems random, motor is not hot to touch…
Is this a fw bug, time to upgrade? I’m running a Flipsky Dual and heard 3.65 is latest stable for this ESC
Maybe. can you show that you have noisy temperature sensor with any kinda telemetry?
would have to be a jump to 5.2. there was mitigation in 5.0 (stronger software low pass filter) but a new temperature related median filter bug in 5.0 and 5.1. 5.2 is still beta
also worth checking out later part of that thread:
4 Likes
Figured I would need telemetry to rule that out. I get my metr pro this weekend, I’ll hang tight and see if I can replicate it again with more data
1 Like
Alright, I got the fault to trigger a bunch more - I’m not sure if it’s really my motors overheating or caused by noise.
It seems I can feel my motors hesitate and buck when I approach 18.5 MPH. If i really push this on the throttle, it will trigger a fault and cut out. Here are faults from Metr:
Wondering if my best course of action is going to be either to upgrade to new F/W or remove the temp sensor.
1 Like
hey who stole “Stoner Ram” ???
easy enough just to depin the temp. sense wire from the plug and test it…
I had some wonky riding situations, disconnected the temp. wire and no mas problemos…
I may try FW5.02 when it comes out of beta… or may not… everything runs purrrfect now…
1 Like
Alright, I may try this then. So i just yank it out, or cut it clean? And I assume I can re-attach if I’d like thereafter if needed?
Also trying to find a schematic to determine temp sensor wire for these Haggy Motors -
1 Like
De-pin the temperature sensor pin from the JST connector… pull “gently” up on the little plastic tab and the female pin will slide out of the JST… insulate it from shorting anything, and you can always just slide the pin back in the connector
Most of the time the Temp. wire is adjacent to the 5v, find a pin out of your VESC and it will show the location…
1 Like
You’re getting overtemp motor at 43c?? What are your values for
motor temperature cutoff start
motor temperature cutoff end
Acceleration temperature decrease.
if they’re anywhere near defaults. ( 80c - 100c - 15% ).
then you’re definitely getting spurious readings to have it say 43c in the fault and not be anywhere close to those settings.
but since you have metr you should be able to upload a record of you ride and share the url with us. you can see the noisy temp readings on the temperature graph.
1 Like
Lift the correct plastic tab with like an exacto blade and the female pin will easily pull out…
Then you can simply push it back in again when you wish to use it again
2 Likes
So here is the graph and motor temp marked right before a fault at 18:53
Defaults -
Motor Temp Cutoff Start- 85c
Motor Temp Cutoff End- 100c
Acceleration temperature decrease- 15%
looks like wild motor temps to me. not 100% sure of the scale, but I think it looks worse than mine above, which looks worse than other people’s it should ideally move very smoothly. no big jumps in temp.
Looks to me like you have the issue.
1 Like
Thanks, I will give this a go tomorrow. Anything else i should be weary of while running without these sensors?
1 Like
You’re planning on running with all your sensors except temperature, right?
ride it like you stole it, hand check the motor temps occasionally until you’re sure you’re not cooking the motors.
2 Likes
worth noting you can “upload” your metr record and share the link on it’s own line and it will get embedded.
Also worth noting that there’s a little split icon you can use to get both motor temps if you were connected via CAN or have a unity.
This is a record from a ride on a second board with the exact same setup as the one I have problems with above. motors/vesc/battery all the same. you can see the temperatures are much smoother.
even in the noisy spots it’s not really jumping by more than a degree. that’s more like what yoyu wnat to see.
1 Like
So I’ve got em pulled out and heat shrinked
Sanity check me just in case, but metr is no longer displaying motor temp so pretty sure I got it right.
Diagram
Right
Left
Thanks for all the help on this guys, and @fessyfoo for metr tips and logs too
2 Likes
and… i triggered it one more time. because I haven’t gotten around to a better fix. and the 120c cutoff isn’t “safe” enough.
would love a disable temp sensor boolean option in vesc tool the way focbox ui has.
still fw3.65
1 Like