Intermittent voltage drop on 12s to <32V (possible short?)

Esk8 Electronics Guides, Support, & General Questions
#1

Setup:
Spintend Ubox V1 Micro USB, fw 5.2
12s5p P42A by @ZachTetra, rated for 125A continuous
LLT Smart BMS discharge bypassed
TorqueBoards DD 75kv
+/-80 motor amps per side
+60/-25 battery amps per side
XT90S loopkey

I’m having a problem where, at seemingly random times, my pack voltage drops to <32V for a few seconds, causing the ESC to cut all power. Here are some metr logs that show the issue:


2 sec cutout


4 sec cutout
1 sec cutout

After this log, I discovered that I hadn’t been plugging my loopkey in all the way due to debris in the socket, so I cleaned it out and made sure the loopkey was all the way plugged in and took it for a test ride. The ride went smoothly with no cutouts, even though I ran over every bump I could find to see if the issue was linked to vibration. However, on the next ride, another cutout happened.


14 sec cutout

After this cutout which lasted much longer than the previous ones, I wasn’t able to unplug the loopkey easily. With some force, it did unplug, leaving a ring of green plastic from the negative side of the loopkey attached to the opposite connector. This leads me to believe there is an intermittent short (or at least massive current draw) occurring, making the battery sag significantly and heating the loopkey up. However, there aren’t any burnt or discolored spots anywhere visible inside the ESC. By the second metr log, the BMS had recorded a min pack voltage of around 30V and logged an undervoltage fault (I didn’t check it when the first cutout happened). I reset session data before the third log to see if it would detect another voltage drop, and it’s now saying it has recorded a min pack voltage of 1.50V, which isn’t possible, and did not log a pack undervoltage fault. I did notice that the brakes still worked. I’m not sure if I just happened to hit them right when the cutout was ending or if braking somehow was the event that ended the cutout.

Before this issue started happening, there were some other funky things going on. I was on a ride with friends and the board felt like it was rear steering more than usual at some points. I checked during the ride to make sure both motors were working and they were, but after the ride was over, one motor was not being driven. I can only assume that intermittently running on one motor during the ride was what was giving me the rear steer feeling. Unfortunately my metr is very temperamental and wasn’t staying connected for that ride, so I don’t have logs to confirm that.

After a power cycle, the motor came back up, but the brakes felt off. I tried riding back and forth to figure out whether it was just me or if something was actually wrong, and the intermittent motor made a very loud sound and stopped being driven. My metr refused to connect at all at this point. When I got home, I opened the enclosure without removing power and found that one side of the ESC had powered off. I was unable to connect to it over USB or Bluetooth to read faults. After a power cycle, I tried to run the motors and they weren’t drawing enough current to spin at full speed and the battery was sagging much more than it should. I found out I only had the loopkey half plugged in and power was running through the resistor. After plugging it in all the way, everything seemed normal, but I’m not sure if I damaged something in the loopkey. It doesn’t spark when plugged in, so I assumed I hadn’t. For good measure, I reflashed firmware and reran detection without issue.

On my next ride I noticed the battery felt more saggy than usual, dropping from 47.8V to 40.5V during a 100 battery amp load on a 60ºF day, and acceleration felt affected. The ride after that is when cutouts started happening.

I’m not sure what next steps for troubleshooting would be. I don’t have any extra parts I can swap around. Is this likely an ESC issue, a motor issue, a combination of both? I don’t have any idea how to proceed, especially since this seems like a very dangerous situation if high enough amounts of current to sag the battery 10V and begin to melt an XT90 are involved.

2 Likes
#2

TLDR.

My guess is a bad series connection somehwere between the + and - to the ESC, could be anywhere, nickel, cable or connectors. unless you’re really abusing the cells above thier limitations.

2 Likes
#3

loop key resistor is burned out, due not properly plugged in. This is also only part of the problem, as was already said, there is a dodgy connection somewhere.ALso whats call and p-count are you running? this could play into the problem aswell.

1 Like
#4

I’m running 5p of Molicel P42A.

Thank you both for your input. I hadn’t considered a bad series connection but it would make a lot of sense.

#5

Okay a 5p p42a can deliver more than enough current without dropping. Somewhere in your system load increases resistance, that means something which shouldn’t get hot gets very hot.

1 Like
#6

Sounds to me like you have some issues with the connectors.
There is are small region in the XT90S connector where there is no connection if the mating connector isnt fully connected.

when connecting … It’s Resistor–open circuit–fully connected. Thats why it needs to be pushed in quickly.

If you fail to connect it fully, it can end up in the region where there is barely any connection.
It will run crazy hot in there if you push high amps.

I dont know how you have it mounted, but my guess here is that you run the XT90S as a connector on the battery (not as a series connector with a loose bridged XT90 connector for start up).
Correct?

You then have pos and neg in the loopkey. Then if you have enough arcing and melting going on inside… its possible you could get a short between positive and negative showing up as your massive voltage drop.

The whole situation would be a very stressful for the ESC which explains why isnt behaving now.
I don’t think the resistor is part of your problem. It would have blown pretty quick if it was conducting while you spin the motors.

Investing your XT90 connectors should give hints… should be burnt in that case.

I could be wrong, Happy fault tracing! :slight_smile:

#7

This is likely your problem.

#8

If you are able, I’d go back through the entire power chain and reflow every solder joint. That means all series and connector joints. If that doesn’t fix the problem, then it will most likely help you locate it.

edit: just occurred to me: are the cells well balanced? What is the health of the pack like?

1 Like
#9

I’m not sure quite what you mean by this, but i have the XT90S run as a break in the positive wire between the battery and the ESC.

I was under the impression that if the resistor blew, the only negative effect would be a spark when the loopkey is plugged in. That isn’t happening, so I assumed the connector was fine enough to continue using. I could be completely wrong about this though, so I’ll definitely try a new XT90S.

The cells are just over a year old and stay within 0.012V of each other at all times. Right now they’re reading within 0.006V.

1 Like
#10

excellent. That rules out bad welds. Usually when bad welds are the cause of bad series connections, they are also the cause of bad balancing.

so the issue may still be solved with a reflow of all your series joints, including connectors.

2 Likes
#11

Ah gotcha. Then my arcing theory between pos and neg was a swing and a miss. :-1:

#12

Yeah, I’d check all of my solder joints before taking the battery apart. It’s not likely an issue with the battery construction.

1 Like
#13

Man this thread was super helpful to me just now. My nearly identical pack tried sagging over fifteen volts tryna pull ten amps through a Ali battery capacity tester just now, I was super worried. And just about to start doing some of my famous investigatory work aka taking it apart never to resemble. But instead I think I’ll just construct a proper connector instead of alligator clipping my xt90 lol. If this solves the issue, I’ll be back to thank y’all!