[Investigation] VESC DRV8302 failures

Hello and welcome,

I’d like to investigate why the VESC keeps failing sins it’s a very expensive problem and very dangerous.

the most common failures of most VESCs I’ve seen accrue from the DRV8302 just being fried up after or during a ride not even maxing out the current or voltage setting of the VESC. this mostly happens when people use FOC settings(but was kinda “FIXED” after adding a capacitor) but also happens of the other setting.

Does anybody know why this happens and what can you do to prevent it?

This will just help improve our only option of motor controller and prevent further dead VESCs for the graveyard.

Thanks for the contributers.

I am also very curious on this topic. It would be awesome to have a data base of all the possible ways the DVR can fail, in order for VESC users to diagnose/trouble shoot.

From what I can tell and what makes sense is that the DRV dies because of high voltage inductive spikes from the motor. Adding more capacitance to C18 like what Ollin VESC’s do helps smooth those voltage spikes.

Always install new firmware on any new VESC. My VESC is fried from a DRV8302 failure because Torqueboards/DIY sent me a VESC with bugged firmware. Pretty disappointing when I purchased it months after the current ramp bug had been reported. Unfortunately there is no way to tell if you have the bad firmware or not, because Benjamin Vedder did not change it to 2.19. They are both 2.18

1 Like

I know inductive spiking in speakers and DC motors can be solved bu adding a shotkky diode parallel to the load in reverse. couldn’t that be the solution for the VESC as well?

How can you be sure it is the firmware? the VESC needs to be protected hardware wise as well…

There are multiple failure modes which can lead to a DRV error. Bugged firmware is one of them. Unknowingly running a maxed out current ramp step of 50, instead of the default 0.04, is a surefire way of destroying your VESC. It’s a precaution, not an end-all solution…

The VESC already has a Shottky Diode in it, its D4 Shottky 60V 2A. the diode does help with reverse voltage spikes but clearly not enough. @jmasta was just saying its one thing that can lead to a dead VESC or DRV.

Another scenario that can kill a vesc is coasting downhill too fast and then applying the brakes. Seems that you can exceed the erpm limit with negative as well as positive torque.

1 Like

inductive voltage spikes can occur because of lengthy power wires between the VESC and the battery, correct? In my build, I have a split enclosure in which there is some space(about 10 inches) between the batteries and the VESC, and I just experienced a fried DRV. I also, by accindent, had the ERPM limits set to - and + 100k. Could the long wires could have been the culprit of the ruined VESC? Or was it the ERPM setting It happened after braking for the first time after going 15 ish mph. (This was all with a torque boards VESC)

Thanks:)

Also, what would be considered the safe length of power wires from the battery to VESC?

If doesn’t help enough why not add another one in parallel?

I don’y think it has anything to do with the length of your wires.

Some of the errors that caused for failure are if you soldered it by hand and the ground pad wasn’t connected, it happens. Aside from that, most DRV errors are a result of wrong settings or running on 12S and some spike in current or voltage. Obviously, the VESC wasn’t designed to take extra current and burn it off. The brake function wasn’t designed to work as a real brake on long down hill stretches.

1 Like

Do you have an example for bad settings besides FOC?

Also the design of the VESC needed to assume the brake would always work as hard as it can without damaging it…

1 Like

I run my brakes continuously down steep hills and sometimes on a full battery and have had no issues. I do use moderate settings for motor and battery min. And I’m running dual setup so the load is divided. Motor min -40a Battery min -10a

1 Like

I think my first one originated from this vreg which blew pin 29 on the DRV

My second one is from freewheeling downhill and then braking while at a higher then max erpm(39mph, on my setup would be about 67k erpm). No fancy pictures for this one(no burns) just DRV fault.

Downhilling and holding the brake and staying at a speed that is under the erpm limit to your gearing is fine. But freewheeling faster and braking popped my DRV in half a millisecond. As the above poster said running brakes down a hill all the way on a full battery is fine as well. Its just braking while higher then 60k erpm ie I would be at 60kerpm at 35.4mph but I was going 39 when I braked…poof :slight_smile:

The good thing is if you built your board right you should never even be close to max erpm of the 4.12 while motors are under power(throttle) within your geared speed - unless your running 12s

mine is longer than 10 inch.

Hahaha I read this wrong and forgot the thread topic…lollll

Anyway mine is longer then 10" as well. You see with brushless motors and speed controllers the length of wire from power source to ESC doesn’t matter so much, just the ESC to motor. The longer the motor to ESC wires the easier for the back emf to dissipate Down the wire and cause desyncs

My power to ESC wires are about 34 inches total Vesc power leads are about 10" and my liion packs about 14 inches from the first pack to the terminals

And then 20.5 inches up top

Power to vescs doesn’t matter to much in length. Just motor to escs.

1 Like

I’m not into how the vesc is designed but on RC groups there is a long thread about too long battery wires will kill your escs. I did some scope measurements to this with running a ESC completely without CAPs, then with stock caps and then at last one with 6xlow ESR capacitors wired in parallele to catch the spikes. It’s not the capacitance that matters here hence why I did use many small instead of one big. Without caps I think spikes hit 35volt on a 6S system (25,2V charged). In comparison that would mean 70volt spikes on a 12S system. Could this be related to failed VESCs? It’s quite common in big custom quad copters with remote installed ESC to add capacitor banks out close to the ESC to prevent failures caused by this.

1 Like

On my builds, the negative main is fairly short but the positive main is about a foot long 12ga silicone wire. That has not caused any problems with over a year of operation. Running dual Vescs 4.12