Limelight Caught On Fire (and other Electrical Woes)

Hello Everyone!

W.A.F.F.L.E.S. are in need of as much help as we can get. We’re competing week one and today at 5406’s practice field and today our Limelight 2 it got a little smoky and caught on fire. Before you ask - yes, we have contacted Limelight directly but haven’t heard back.

We saw a small wisp of smoke and then we smelled ozone right after we plugged a battery into our robot and turned on the breaker during practice. It turned out it was from the limelight, we opened it up and found damage on the camera board between the PI and the camera board.
See below for some pictures of inside.

Would anyone know why this would happen? We’ve got a spare that a team is going to loan us but if we’re doing something wrong we don’t want to put another one on. Used this model all year last year without an issue.



2 Likes

What was your breaker rated for? Is it possible you used one that was rated for a higher current and there was a short or something that caused too high of a current for the LL to handle.

It was a 20A breaker, which is within the recommended range as far as we know.

I’m not an electrical engineer, but I think that could be your problem. My team has been using a 10A breaker. And according to the website it should be a 5A breaker.

It almost looks like there was too much solder on C22 (or a bit of metal swarf?) somehow which caused it to short to a pin on the U1 IC. I’m not sure that’s what happened, but that’s just what it looks like to me.

A 5A breaker probably won’t save your Limelight if it shorts for some reason — at the point that it’s drawing more than 5 amperes, it’s probably already dead. The breaker is there to prevent your Limelight and the wires to it from becoming a fuse element — which is to say, to literally catch on for and destroy a lot more than itself.

8 Likes

We referenced this, and confirmed with other teams like 1241 and 5406 that 20A was okay.

It isn’t, and you really shouldn’t be running more than a 5A breaker to the device.

I mean, it’ll work just fine on pretty much any breaker, or none at all. It’s when things go wrong that you’re going to wish for the 5A rather than something bigger.

1 Like

Agreed. The oversized breaker didn’t cause the problem, but a smaller one may have prevented the problem from spreading into a fire or other additional damage. Please let us (and future readers of this topic) know what you learn.

5 Likes

Heard back from @Brandon_Hjelstrom and he was super helpful! While he said this was the first time he’s ever seen this issue, our current thoughts is that it’s a fluke accident of some metal getting inside and shorting out the board.

Getting a replacement from LL and sending ours back for hopefully a full autopsy.

3 Likes

If you look at the curves for the breakers, you’ll see even the 5A breaker will allow decent to pass for quite a while before tripping.

It is incredibly unlikely that a 5A breaker would have saved anything here.

47 Likes

I am an electrical engineer, and what is left of U1 (the little black box) looks like a voltage regulator that didn’t survive. As you say you used this same Limelight last year, there is a chance that sometime in the past an overvoltage event occurred (maybe rolling your robot by hand with NEOs) and U1 finally died. Sometimes circuits can last a long time after an event, and not have a catastrophic failure until much later.

As it failed when you were turning on your robot, I doubt you have anything wrong. Luckily this failed this week and not during a competition.

6 Likes

Just add some computer fans and call it a day. EZ

Or just slap one of those falcon cooling ports on it and run air through it.

7 Likes

Have you tried turning the limelight off and back on again?

15 Likes

Have you tried rinsing off the burnt parts?

Tbh if you just rinse off the magic smoke stain it should be fresh

10 Likes

Yes, absolutely. The best a 5A breaker is going to do in this situation is keep Very Bad™ from becoming Extra Worse™. Your Limelight is done for regardless.

3 Likes

We are really stumped, looking for additional help.

New info… Looking for some additional assistance here. Since the LL has died on us, both the VRM and the PCM have died.

What we know:

  • PDP breakers haven’t been damaged. Has no visual damage, and continuity exists on the breaker.
  • neither the VRM nor the PCM have any observable damage when opened up, not do they turn on with a 12V input.
    -All motor controllers are behaving fine.
  • The VRM likely died at the same time as the limelight.
  • The PCM died after less than two minutes of drive time after vrm and LL. No battery was removed. The robot was driven slowly.
    -The electronics seem to be in functional order, some less than ideal can wiring, but that shouldn’t cause issues.
  • The robot frame is isolated from the main breaker.
    -when the VRM and LL died, the robot was rapidly turning and the battery was ejected/disconnected from the robot.
    -We have approximately 6-7 hours of power on time/drive time prior to any of these issues occuring.
    -There is a very large amount of static buildup on the field we are currently at, to the point where the robot can turn 180 degrees and touching metal will shock the robot. Other teams (4039/5406) are not killing components.
    -the PDP PCM and vrm were brand new. The LL ran all 2019 season fine.
    -putting a new vrm in had no issues, but then the PCM died later

Not really sure what the common failure point is, and it really doesn’t seem likely that these components are all failing at the same time.

I’m not sure if you are serious or not.

@tmpoles the grey patch in the middle of U1 indicates the chip inside got hot enough to burn the black plastic on top of it. That is generally a fatal condition. It also looks like a piece of the black plastic has blown off the lower right corner of U1. The charring next to C3 may correspond to the heat damage on U2 (2nd picture) next to R26. Is R26 still there? It looks like only the solder on the two ends are still there. It may be possible that U2 is also damaged

Here are the dead vrm and PCM. Both look pristine, but neither turn on/have leds

I doubt that your failure was caused by static. Static discharge is most likely to damage signal input pins on IC’s because the are connected to very small transistors that don’t take a lot of energy to kill. It looks like U2 was a voltage regulator wich would have transistors that would be hundreds of times larger and muc,h much more resistant to static damage.

I worked for over 15 years in Ottawa and TO where I killed my fair share of chips with static but it never produced such dramatic results. The devices all died with just a whimper and no release of magic smoke.

Is there any evidence of loose bits of metal in the case? It could be in the form of metal shavings or chips or solder balls.