You are here

Failures so far and how to prevent them

8 posts / 0 new
Last post
benjamin
Offline
Last seen: 21 hours 5 min ago
Joined: 2016-12-26 15:20
Posts: 369
Failures so far and how to prevent them

So far I have repaired three VESC6 that have failed for various reasons. Here is what went wrong and how to prevent similar failures. I will update this thread if I repair more VESCs that have failed for different reasons.

Dead MCU

I got two VESCs where the STM32F4 had died. The reason was that the ADC inputs for an analog throttle have been connected in parallel. When VESCs are connected to the same battery there will be a voltage differential between them under load due do inductance and resistance in the cables. This voltage differential can reach quite high values and the MCU pins are very sensitive to that. Negative one volt is sure to kill the input and short the whole MCU. The PPM input is slightly less sensitive to this than the other pins as it has a low-pass filter, but it still is a bad idea (the VESC4 had a slower PPM filter and was slightly less sensitive). The CAN-bus can handle around +-60V difference, which is why it should be used on a multi-VESC setup. The CAN-bus also provides other advantages such as cruise control with several motors and traction control.

Solution: Always use the CAN bus to connect VESCs in parallel and set up the apps accordingly. Never connect the other inputs in parallel in VESCs that are connected to the same battery.

Dead voltage regulator

One of the VESCs that I repaired had a dead 3.3v regulator. One way to kill the regulator is to feed it backwards with a 3.3v source, which has happened to me a few times when connecting a programmer.

Solution: Never feed the 3.3v net with voltage, only current draw is allowed. If you, for example, have a stlink v2 from ebay that outputs 3.3v on a pin then don't connect that pin - power the VESC from the input while the programmer is connected.

Roger Wolff
Offline
Last seen: 3 months 1 week ago
VESC FreeVESC Original
Joined: 2017-05-24 12:27
Posts: 202

Solution: Always use the CAN bus to connect VESCs in parallel and set up the apps accordingly. Never connect the other inputs in parallel in VESCs that are connected to the same battery

The alternative is to have two separate batteries, and have the VESCs be completely separate except for the ground-and-ananlog signal. (ground-analog-3V3 on one VESC, ground-analog on the second). 

Dead voltage regulator

Would a $0.05 bat54 connected between 5V (kathode) and 3.3V (anode) work to prevent this?

(there is a typo in the subject for -> far). 

benjamin
Offline
Last seen: 21 hours 5 min ago
Joined: 2016-12-26 15:20
Posts: 369

The alternative is to have two separate batteries, and have the VESCs be completely separate except for the ground-and-ananlog signal. (ground-analog-3V3 on one VESC, ground-analog on the second). 

Yes, that should also work.

Would a $0.05 bat54 connected between 5V (kathode) and 3.3V (anode) work to prevent this?

I think a diode could help to prevent reverse voltage, but 0.8V drop at 3.3v is a bit much.

Roger Wolff
Offline
Last seen: 3 months 1 week ago
VESC FreeVESC Original
Joined: 2017-05-24 12:27
Posts: 202

Oh, no not like that. I mean: The regulator probably fails due to the 5V line (its input) being lower than the 3.3V line. So if you put a diode there so that this will not happen, the regulator will survive. 

Jebe
Jebe's picture
Offline
Last seen: 2 years 2 months ago
VESC Original
Joined: 2017-05-23 16:37
Posts: 3

Have had canbus failure on my slave vesc. Unlike vesc 4.12, this vesc still powers up and has been reconfigured to run via ppm. Am running with a split ppm signal for now. Dual vesc 6 canbus 12s trampa dual mtb

tuloski
Offline
Last seen: 1 year 6 months ago
Joined: 2018-09-11 11:53
Posts: 12

So using PPM input (from the same autopilot) on 4 VESC for a multicopter is a bad idea? One single battery powers everything.

TechAUmNu
Offline
Last seen: 1 week 5 hours ago
VESC Free
Joined: 2017-09-22 01:27
Posts: 382

As with all other escs, just cut the red wire on all but one of the ppm cables. So the 5v power only comes from one VESC.

frank
Offline
Last seen: 3 weeks 6 days ago
VESC BronzeVESC FreeVESC GoldVESC OriginalVESC Silver
Joined: 2016-12-27 20:19
Posts: 367

Don't do that! Y ppm causes ground loops, even if you cut the voltage supply. These loops can kill a Vesc 6 pretty fast. HW 4 designs can tolerate ground loops better, and most of them survive it. Its still no good idea to use Y ppm.  Always CAN. HW 4 wants to be powered up simultaniously using CAN. Otherwise you can fry your CAN chip. Just make sure they run of the same power source and you won't have any issues. Vesc 6 can tolerate to be switched on individually while being connected via CAN bus cable. 

Only connect CAN L and Can H, GND and 5V are unconnected.