I have reports from 3 independent sources about the following DAVEGA behavior:
While resting, all is working normally.
Under throttle, UART connection becomes unstable and DAVEGA receives little to no updates from the VESC, resulting in a very poor display refresh rate.
I had this reported by @glyphiks, @Oliver, and now also by another customer who’s not on the forum.
This appears to be a UART issue on some VESCs. I’m creating this thread to:
Get more insights into the issue.
Compile a list of affected VESCs.
If you experienced a similar kind of issue please chime in. According to @rpasichnyk, this was also reported for Metr Pro (UART) but not for Metr Pro CAN. Robogotchi users might have also been affected (tagging @DerelictRobot).
Here’s a list of VESCs, for which the issue has been reported (I’ll keep updating it):
I wonder if the construction of those ESCs has anything to do with it, maybe the higher voltage is inducing noise onto some of the components? I have not had issues with Stormcore or Tronic ESCs
That’s what I thought, too. Nevertheless, @Oliver also tested with his Trampa VESC 100/250 (even higher voltage) and no issues there. So maybe it’s actually not related to the input voltage but rather to a particular design of the ESC.
I have had the odd issue reading and writing settings to the VESC over CAN on the D75 with metr.pro. the VESC it’s plugged into is fine but the one over CAN sometimes errors
We haven’t seen anything like that. Actually the 100/250 and 75/300 are logic stage wise pretty much the same. We did lots of UART stuff and also our customers use UART a lot - so far without any issues.
Personally I would check the wiring from the ESC to the UART device and watch out for cable routing and possible interference.
It was exactly the same setup in my case. Vesc 75300 R3 had issues, Vesc 100250 absolutely no problems until now. Cables and cable routing all the same. To my mind it’s a hardware issue on 75300, there is no other explanation.
Please do. That would be very useful info. If you can reproduce the issue with the BT module connected via the DAVEGA cable, could you please also check if bypassing the DAVEGA cable changes anything (i.e. when you plug the BT module into the slave UART port using its own shorter cable)?
In the cases where failure has been observed, where is the ESC and DaveGA mounted?
Nose and tail of a board can be 1m+ apart and for a normal uart operating in a high noise environment it’s not that surprising to see interference.
If they are mounted like this it might be worth trying shielded cable maybe with differential pairs, and making sure you have a solid low resistance ground connection.
I’d like to mention I’ve seen this happen with the makerx D series mostly @YUTW123
Mine has this issue too. Robogochi not working well. Maybe it’s time for every device to go CAN
I just had the issue reported from another customer who’s using Maker X D100.
There’s also a report from someone using the DAVEGA on Kaly XLR, which (according to Ernesto) uses the Ubox 100V.
So far it looks as though only setups above 12S are affected.
The strange thing is that above 12S setups are nothing new. People have been using DAVEGAs on Kaly XLR for quite a while a no issues reported earlier.
This may be a valid point but please note there’s currently over 1000 DAVEGAs out there. Vast majority of these use either 100cm or 125cm cables and I don’t remember anyone reporting a similar issue over the last 3 years that DAVEGA has been around.
There also haven’t been any recent changes to the DAVEGA hardware. I’m still running off the last batch that was produced in early 2022.
This is a mystery. I hope we can collect more info and get to the bottom of this.
For the record, I just created this page and linked it from the DAVEGA X e-shop item to make people with high-voltage set ups hold back on the purchase until this is investigated.