Battery Mounting in 2024 for Shock Load? (Lost a cell in Chezy QM71)

Technical Technical Discussion
,
1

Hey Chief, how are y’all mounting your batteries these days?

We stopped moving with 27 seconds left in our last qualifying match at Chezy Champs. It was far from the only reason we didn’t get picked for an alliance, but it definitely didn’t help us get picked.

Here’s the match video, timestamped to a couple seconds before the hit near the red Source:

Unremarkable hit, followed by zero lights on the robot. All main power wiring untouched, main breaker closed and cold. Actuating the main breaker with the match battery on the robot does nothing, robot stays off.

Take the battery off, battery reads 10.7V. Plug it back in, the robot still doesn’t turn on. Check resistance from + and - to ground, Megaohms, great.

We then checked the resistance of the main power wiring + to -, got 143 Ohms (not a typo, under 200), and chased that for the next two hours

It turns out, because the radio bridges the Vbatt into the RIO Ethernet port when used with an off the shelf Ethernet cable, we were actually measuring the resistance of the decoupling caps in the RIO ethernet port. This took a consult with @kiettyyyy to figure out, after I had isolated the 143Ohm issue to the fuse feeding the new radio, asked for a spare radio, and then replicated that connecting the new radio at the RIO port resulted in the same 143Ohm showing back up. I gave the spare back at that point…

This needs to be documented for CSAs - of three other teams I later checked main bus resistance on, two were in the 3.7kOhm-4kOhm range (at least one was using a REV cable to isolate the RIO from PoE) and one was measuring 513Ohm. I would have loved some community lore that “143 Ohm is possible, disconnect the RIO’s ethernet cable before you measure”.

When we dropped a new battery in, the robot ran fine with no changes.

Turns out, that battery was DEAD-dead. One cell totally gone, likely damage to others. We put it on a charger overnight, it still read about 10.7V and wouldn’t even turn on a Battery Beak the next day. Album attached: https://photos.app.goo.gl/2DtkX6AwHWoF9sQv7

I figure this has to be a progressive failure mode, given how unremarkable* the final impact was that triggered the complete failure.

Has anyone else done this to a battery before?

Has anyone established a method to detect impending failure?

Have y’all put foams or other impact dampeners around your battery mounts?

I thought we’d come up with a pretty good battery mounting strategy, but now I’m questioning it from a shock load perspective:

image
image526×576 59.1 KB

There’s about 1/64-1/32" of clearance all around, so it’s not completely rigid in there, but it definitely doesn’t explicitly protect the battery from shock loading. Whatever makes it past the bumpers is definitely transmitted into the body of the AGM battery.

We’ve never seen this prior to this season. These are new April 2024 Duracells with just EBR, Sunset, and Chezy on them. It feels like a really short lifespan. I don’t need to spend $hundreds on batteries every six months, this isn’t supposed to be Battlebots™.

*(Was it actually an unremarkable hit?)

6 Likes
2

This isn’t Battlebots, but padding around batteries for shock reasons is pretty standard practice for battlebots. Common implementations are gum rubber lining of a rigid cavity, battery boxes mounted on rubber sandwich mounts, or printed TPU battery mounts.

Never personally experienced a battery failure clearly tied to physical impact like this. My teams often used thick polycarb for structural elements of our battery mounting, in a variety of ways. This probably helped dampen things a bit compared to a rigid aluminum setup, as it would have a bit of spring to it compared to a rigid box frame.

I’m particularly concerned by the lower edge of your setup, where it appears the battery is constrained left to right just by the edges of thin gusset plates, not a full height backing. That strikes me as something that could place concentrated stress on the battery rather than spreading out the load. Especially since the left/right constraint is only at the bottom 1/4th or so of the battery, and the momentum of the battery is going to want to torque it in there, potentially generating point loading at the corners of the gussets and battery. Did the exterior of the battery show any marking at this spot? Was the hit from 90 degrees relative to the battery like this? Though like you said, it could have been progressive from other hits in this direction.

4 Likes
3

Is it possible that this was caused by mounting the battery directly against the frame? From the image it looks like the frame rail might even be bent in towards the battery. I’m not an ME but I’d think that’d cause the battery to directly absorb the force of a hit, rather than it being disappated if the battery was mounted e.g. in the center. But I’ve seen that mounting strategy before so not sure.

3 Likes
4

Lead-acid batteries can suffer from fatigue fractures of the straps between cells and the internal connections to the posts due to vibration. I’ve seen it multiple times in motorcycle batteries. It shows itself as a sudden high internal resistance.

5 Likes
5

For the past two years we’ve lined our battery boxes with a thin foam from McMaster. It has seemed to work good for us the past two years.

2 Likes
6

I knew I wasn’t crazy! I thought the frame looked bent.

2 Likes
7

We often save our scrap pool noodle cut offs from making bumpers and make a thin (1/4" or so) puck to squeeze into the battery box. Keeps the battery nice and snug but has enough give to provide some cushion. We sometimes replace them after they get squished or sometimes we just let them go all season figuring something is better than nothing.

Never had a battery lose a cell from impact before, but then again we never had to replace our frame 2x in a season until this past year either.

1 Like
8

I’m thinking this root cause, since our battery dressing design isolates the posts pretty well from getting pulled.

9

Nothing obvious, but I’ll look at our whole rack Friday to be sure. I was not particularly worried about the thin al because we’ve used thin al rings for battery boxes for years, and this battery box is pretty secure.

It’s definitely a little bent, so there’s some pinch action in that orientation now. Fair point that with greater rigidity is greater fatigue.

We replaced the intake frame rail multiple times per Regional, but I think left this one all season.

10

My primary theory is at some point this battery was dropped, and nobody noticed or said anything. This injects a latent failure triggered by impact at some time.

Secondary theory is these two red corners are poking the battery pretty hard as it jostles around and twists due to the blue gaps. At likely only 1/8" tall the stress would be pretty high on contact. Are there any markings there?

image

4 Likes
11

Yes, we generally print a PLA+ enclosure (3-4 sides around the battery) with around 1/16" or 1/32" offset between the battery and the print. The enclosure bolts into the belly pan (rivnut or nylock underneath) and has some slots in it for a strap to go through, under the belly pan, through again, and above the battery. Then apply generic adhesive foam strips from Amazon around the inside perimeter of the print.

image
image1015×1353 227 KB

This pic doesn’t have the foam in yet but you can see where it would go.

I’m not sure how much it actually helps, but it’s relatively trivial to add, and I’d recommend it to everyone. Like Adam said, those red corners he circled are probably stress points.

My rule of thumb is that the battery should be completely constrained in XY by the robot, and a strap only needs to constrain it in the Z direction; although this can change if you’re doing a brain pan or whatever with a bolted on cover plate.

2 Likes
12

I will second with the idea of adding foam. On LigerBots we always used some foam between all metal surfaces and the battery if possible.

2 Likes
13

I have mounted just like you pictured with 1/4-1/2" of very dense foam (e.g.floor tile) sorrounding the battery. Thickness is based on battery fit, not some special calculations. Basically operating under the rule of " something is better than nothing". I will admit last year this wasn’t really followed, but we lucked out.

I haven’t had an issue since 2017 iirc. That is over the 4 teams I have been on.

14

9401 had this happen their last match on newton. Battery would not create any meaningful power ever again. 1/8" frame dented in substantially. Never had it happen again so bad luck I guess.

15

“dented in substantially” -
Did you have similar mounting where the battery is pinched vertically against the dented tube?

16

Battery actually was mounted in another location so the frame by it did not dent but was for sure a hard hit.

1 Like
17

I’ve seen this type of problem, both as CSA and mentor. Worst case was a somewhat loose battery that collided with edges of battery enclosure, concentrating stresses. I’ve seen a battery actually leak after a hard hit. Multiple times, I’ve seen batteries die on the field this way. Accumulated damage is probably part of this, but I’ve seen a major hit correlate with killing a battery more than once. Video review is worthwhile for this type of thing.

I’d ensure the battery can’t move (at all) and that there are no places where there could be a stress concentration. But, I think shock and vibration can result in internal damage even with this type of mounting. For this reason, I think it’s good practice to pad the battery enclosure so as to take out the worst of this.

1 Like
18

These gussets are the same place I would suspect if the damage is from driving action rather than something such as dropping the battery; applying impact lateral forces over an area probably less than 1/8" by about 2" to a nominally 13 pound object with necessary internal find structure. When I’ve been part of the battery mounting decisions, I’ve always insisted that the support for the battery be fairly large faces of material, not edges of plate or extrusion. I liked the horizontal battery mount for the AM14U and disliked the vertical one for exactly this reason. If the vertical top support tray became a ring with cuts and flanges to provide real surface area [not sure you you manufacture this!] I would have been a lot better with it.
3946 used the horizontal in 2015 (first year available), 2016, and 2018 (my last year with the team). In 2017, it would have interfered with a floor pickup, so we rolled our own vertical mount, but the battery was held in place by at least five square inches of aluminum plate or extrusion on each face. We reprised some of our early and off-season designs by finally securing with a length of angle aluminum held in place against two faces of the battery (end face, and large face opposite the terminals) with threaded rod and wing nuts. We didn’t have padding, but the forces were widely distributed. Of course, today I’d want to move the battery in from the frame perimeter and add heavy closed cell padding, such as foam floor tiles, but much like bumpers, the key principle is to spread the force out over as large an area as feasible.

Also, I suspect that the rules will still mandate a single type of foam padding throughout a set of bumpers. It’s not readily inspectable, but if teams follow it, it is the simplest way to avoid “wedge bumpers”.