Potting Batteries, Thoughts and Experiments

justin_le

Administrator
Staff member
Joined
May 28, 2007
Messages
2,299
Location
Vancouver
Background. Last year we had a number of custom batteries made up in a flat 10s X 1p cell layout arrangement for 2 reasons. #1 is so that we'd have convenient low profile 18650 batteries for mounting on powered skateboard projects (see here http://endless-sphere.com/forums/viewtopic.php?p=904577#p904577), and #2 was to have a supply of BMS protected 36V battery packs that were <100 Wh so that we could take them when travelling on airplanes, then just plug them back in parallel on arrival to have a typical 300-400 Wh ebike pack (See here http://endless-sphere.com/forums/viewtopic.php?p=1021791#p1021791).

All good. We ended up submitting a number of these to a test agency to get UN38.3 certification for shipping approval thinking it could become a handy little product. But as an actual product I'm loath to supply just a bare naked battery and BMS covered in shrinkwrap, as that offers no protection against water, impacts and drops, or solutions for mounting to a bike. Making a padded fabric sleeve can go partway to addressing those deficiencies but it feels like just a stop-gap solution. And since it's a pretty niche item there wouldn't have close to the sales volumes to justify the very high expense of tooling up a custom plastic enclosure or extrusion profile.

Then after the success of some of our other potting experiments with controllers we wondered heck, why not just try potting the entire battery too? We have a CNC machine to make a nice mold so it would cost almost nothing in tooling capital. There was a pretty lively debate in the controller thread about both the pros and cons of encapsulating ebike electronics (have read here: http://endless-sphere.com/forums/viewtopic.php?p=1035601#p1035601), and a lot of the exact same arguments would come into play with batteries.

Having been involved in many many 100's of lithium ebike battery pack repairs over the years, the reasons for actual pack failures typically stem from:
1) Packs being allowed self discharge in storage to the point that individual cell voltages are dangerously low and not safely recoverable.
2) Water damage either corroding/disolving cell tabs or causing erratic BMS behavior.
3) Impact damage from batteries being dropped or involved in accidents or mounted in a way that results in high stress on some individual cells.
4) Vibration causing eventual fatigue failure of tab weld connections.
5) Various random/unexplained BMS circuit failures.
6) Various random cell or cell groups failling (typically developing a soft short).

Of these things, potting the battery would have prevented almost all cases of 2, 3, and 4. Item 6 was an issue in the past but is almost unheard of now that we're dealing only with brand name 18650's. That leaves just items 1 and 5, and both of those should be resolved by robust circuit design and BMS parameters, like ultra low sleep mode quiescent current.

The big downside of potting is of course that should you have a failure it is difficult if not impossible to repair, but hopefully those issues needing repair would be way less frequent. And more interestingly, with this arrangement of a <100 Wh battery pack, it would cost less to replace with a new 100Wh battery than you would spend just on the hazmat shipping alone to send a normal ebike sized battery back to a vendor/manufacturer for repair. I could see being really nervous about having single 500-1000 wh battery be potted and unserviceable, but if this was instead 5-10 individually potted 100Wh packs, then it's no big deal if one module falls out, you still have the other 4-9 batteries to keep you going.
 
If it won't add too much to the cost of the battery, (I wouldn't think it would be too expensive to do) I think it's an excellent idea. The modular idea is excellent too. In moderate to high volume the price might go down enough to be competitive with the larger battery packs out there.

Another innovative product from Grin. :idea: :idea: :idea:
 
Would the potting add significant weight?

If so, would it help to foam the potting?

Some foamed products are stronger than some solid ones....
 
venting? Maybe you could add little blow out "freeze plugs" like on an engine block
 
So to try this out, we machined mold that had ~2mm extra width on all sides over the bare packs and poured a number of prototype batteries in this. The first just with pure epoxy, and then another 3 while wrapping or surrounding the pack with fiber reinforcements, including regular 5oz woven cloth, random fiber roving, and kevlar.

View attachment 1
Mold with Fibers.jpg

And here is what the first batch assortment looked like
View attachment 3

The standard drop test for ebike batteries (see both Batso and the proposed UL2271 specifications) is a 1m drop onto concrete, and all 4 packs handled that just fine. So we upped things a bit to do a 5m (16 foot) drop test from our mezzanine, which would have an impact speed to the concrete of about 35 kph.
5m Drop Test.jpg

The pack with just pure potting resin and no fibers cracked bad on the first drop and then on the 2nd one broke a huge chunk off:
No Filler 5m Drop Test.jpg

However, all the ones with reinforcement stayed remarkably in tact even after tossing them off the ledge 3 times. The fiberglass has some small internal cracks and separations visible, especially when we dropped it on the end.
5Oz Glass Reinforced post Dropping.jpg

The Kevlar had a long crack along the top but that's only at the seam where I had 2 pieces of cloth pieces that didn't overlap. Had it wrapped right around like the glass cloth then it surely would have been fine too. But I learned that even though kevlar sounds cool, it's a real pain to work with the fibers (nearly impossible to cut), and it doesn't wick up the resin nearly as well so the potting/impregnation had to go more slowly.
Kevlar Post Drop.jpg

Oh, and of course nothing is complete without a control. Here is the bare battery with nothing after the 5m drop test. There's no activity of any kind from the cells even after some crushing blows like this.
Control.jpg

I haven't tested any large format ebike batteries in a plastic or metal case from the same drop height, but I have a feeling how it would turn out. We normally get a few end-of-life packs a month coming through our doors for recycling so next time we'll subject them to a little 5m adventure first.
 
I wonder if a sheet of FR-4 on each side would be useful - maybe with a silicone pot instead of epoxy. Even thin FR-4 is pretty rugged and has nice dense fabric/glass penetration. A durable epoxy bond to the smooth FR-4 might be an issue but the silicone should work. I was thinking of the FR-4 on the face but fully embedded is another way to go. More expensive than cloth wrap, but just a thought...
 
Non rechargeable lithium batteries have been potted for use in flooded water meter pits for many years with good results. These were very low power batteries to run electronic meter recording and radio transmitters. This application was only for protection from water intrusion. For resistance to water permiation epoxy is best followed by poly urethane followed by silicone. Two part epoxies and urethanes are available in various hardnesses. I think for you application a urethane would be good. A two part silicone would also be good. I can't remember any exact specs or sources.
 
amberwolf said:
Would the potting add significant weight?

This was a significant initial concern, since in theory you'd think that filling in all the interstices between the cells with resin is just adding extra weight with no structural benefit. But here's the end result (this battery potted with green UV fluorescent dye, in case you're wondering).

Weights Compared.jpg

The potted enclosure adds about 120 grams to the 500 gram battery, so end result is that the final pack weight 80% battery and 20% casing. When we look at most other plastic enclosure battery packs they typically have in the 15-20% casing weight. Here's a nicely engineered Bosch battery which is at 16%.
Bosch Weights.jpg

So it's a little on the high side but in the same realm of what we see with conventional packs.

If so, would it help to foam the potting?
Some foamed products are stronger than some solid ones....

I think the easiest step for further weight reduction would be to use a lightweight filler that is smeared in all the spacing between the cells, and then the fibergerlass and resin only warps around the smooth section.

flathill said:
venting? Maybe you could add little blow out "freeze plugs" like on an engine block

Yeah I want to try and see what the consequences would be to a cell needing to vent. What's the recommended controllable and reproducible method to force an 18650 cell venting situation? Is it best just to directly overcharge to 5-6V or is there another way, like say first draining it down to 0V and then recharging at a high rate? I imagine that the fiberglass would just crack and let the smoke out rather than contain it and burst like some kind of bomb, but it could contribute to the flames if the cell contents are going pyro.

teklektik said:
I wonder if a sheet of FR-4 on each side would be useful - maybe with a silicone pot instead of epoxy. Even thin FR-4 is pretty rugged and has nice dense fabric/glass penetration. A durable epoxy bond to the smooth FR-4 might be an issue but the silicone should work. I was thinking of the FR-4 on the face but fully embedded is another way to go. More expensive than cloth wrap, but just a thought...

We've got a bunch of FR-4 sheet kicking around since our first plan was to just line the thing with FR-4 under the shrinkwrap, but it doesn't give quite the design freedom that I was liking with pouring the resin right over the cloth.

I didn't mention but I have been putting a thick silicone potting on the BMS circuitboard and letting that fully cure before doing the epoxy resin pour, not just to protect against thermal stress on the circuit but also so that in principle if something does fail, it wouldn't be too difficult to cut away the epoxy/glass on the top layer and then peel off the silicone layer in order to expose the PCB again for analysis.
 
Short the string when full charged. All should vent in about 5-20s with no fire

Do control and then compare vs pot

Also do an experiment when you clamp the cell pack so it cant vent and film it for info only. You might see fire and a pop here

Overcharge could be a separate experiment. More likely to see vent and fire even with the vent not blocked but the time it takes to happen is all over the place and each cell reacts diff

Better to test in string as the internal cell protection(s) may become less effective the more cells in series
 
A couple observations/ideas.

The epoxy didn't stick to the batteries in the first test. Roughing them up bit with a scotchbrite pad and cleaning them with alcohol will let the epoxy adhere completely.

Another idea that would give them some protection without compromising repairability could be potting in hot glue.

When working with kevlar, I've found it easier to pinch it between two metal bars along the cut line and use a cutoff wheel. Respirator needed of course. If you run a thin line of quick set epoxy down the cut line, let it cure, then clamp it and grind it off, there's much less mess and the grinder wheel is less likely to snag stuff.

The environmentalist/tinkerer in me would want them potted in something that could be removed for repair but I realize that the failure rate on these cells is so low that the likelihood of a repair being necessary once property potted is slim to none.

Hope that helps!
 
Thermal cycle the material and flex it under extreme cold and vibration as well as high temps (whatever you cells can handle).

Research the materials glass transition points and analyze what effects it could have.

Test the failure-mode effects in over-charge to confirm they aren't more hazardous.
 
flathill said:
Short the string when full charged. All should vent in about 5-20s with no fire
Do control and then compare vs pot
Also do an experiment when you clamp the cell pack so it cant vent and film it for info only. You might see fire and a pop here

Thanks Flathill. I think we'll get a chance to do these experiments and capture them on video this coming week, just need to pot a couple more packs with leads connected that bypass the BMS circuit so that we can run the full short circuit.

ErnestoA said:
A couple observations/ideas.
The epoxy didn't stick to the batteries in the first test. Roughing them up bit with a scotchbrite pad and cleaning them with alcohol will let the epoxy adhere completely.

Ah the reason you see the clean bare cells there is because the cells are wrapped in pink shrink tube. The epoxy bonded fine to the heatshrink, but the heatshrink is not bonded in any way to the cell can so when the resin popped off it pulled the heatshrink with it and left the bare cells exposed. I suppose in principle we wouldn't need the cells to be shrunkwrap once they are potted, but it would make handling of the packs before they get poured quite hazardous since the bare cans can easily short against each other and to the long tabs running up to the BMS board.

ErnestoA said:
When working with kevlar, I've found it easier to pinch it between two metal bars along the cut line and use a cutoff wheel.
Thanks, that's a nice tip. I've since learned that it also cuts OK with a CO2 laser which we have so if there are future kevlar experiments we might be able to cut it that way. It was a real surprise to see how futile even sharp sewing scissors were with the fabric.

liveforphysics said:
Thermal cycle the material and flex it under extreme cold and vibration as well as high temps (whatever you cells can handle).
Research the materials glass transition points and analyze what effects it could have.

I'm just curious is this mostly out of concern with stresses between the cells and the encapsulation, or with the BMS circuit, or just looking for potential mechanical failures of the cast enclosure? We'll be getting some dry ice to run cold temp cycling of the controllers so I can do some batteries too while we have them. The BATSO and UL2271 specifications both have well defined crush testing requirements too which will be easy to do in our shop press.
 
Different materials expand and contract at different rates. If something needs to swell, but can't, something's got to give. All of the potted electronics from major manufacturers that I've attempted to repair use potting materials with some flexibility to account for this.
 
I've been thinking about potting batteries for some time, especially in terms of ruggedising the package. However the added weight gain usually turns me away from it. Plus I'm using pouch cells, and these tend to swell a bit with race duty cycles (20 C bursts). However some level of potting would probably make a difference. I'm not a fan of active cooling of cells as I don't believe they need it, so potting to create a thermal path to the outside of the pack would only add even more weight.

I also don't know if the polyester resin would dissolve the protective film on the pouch cells :lol: Even just wrapping my battery sub-packs in a few thin layers of woven mat and painted with resin could be worth a try later on... :?:

Hey Justin - I'm going to be in Vancouver in August-September so I'll have to make sure we catch up. Been about 5 years I think!
 
Am I mistaken or do I remember from somewhere Zero motorcycles began using potted battery packs or maybe potting cells? I see it as a great solution to an often nagging long term problem.

Refining materials, methods and preferences will require some time and learning from field experiences but it’s probably the ideal way to package battery packs for mobile applications.

As a tinkerer, it’s always nice to have some form of “easy” rework capability but in the overall scheme of things probably not a critical option given high volume and low failure rates.

Truly a pleasure to see you folks work. Thanks for sharing and teaching an old dog a "trick" or two.
 
Helluva good idea, and especially if you do it in the 1 p configuration, able to parallel up to 10 ah or so easy.

Happily, in a way the days of unreliable cells is about over. So now potting the battery can make perfect sense.
 
May or may not be a factor, but nitrogenous materials like epoxy and urethane are inherently flame retardant.
 
Cells vibrating against each other are an electrical short waiting to happen, I think potting would dramatically reduce inter-cell vibration wear and improve shock-resistance (the 5m / 16-foot drop test is clear evidence to me).

Even for an Ebike battery, a bunch of 1P sub-groups would not bother me as a potential purchase. In fact, potting the BMS might be very helpful in reducing issues with an unexpected rain and humidity. If the military was using an E-bike in a rainy area on a mission, I can guarantee the connectors would be water-proof to the extreme, and the battery/BMS would be potted.

I have seen electric scuba tow-vehicles...what are their guts like? (Google "Sea-Doo Seascooter")
 
justin_le said:
I suppose in principle we wouldn't need the cells to be shrunkwrap once they are potted, but it would make handling of the packs before they get poured quite hazardous since the bare cans can easily short against each other and to the long tabs running up to the BMS board.
Maybe use molded (or printed) endcaps? (like the ones for headway/etc)

Then no sw is needed, no shorting, spacing maintained, and potting can bond to the cell itself.
 
This is awesome !!

I messed around with expanding foam in a can years ago ( insulation foam ) with PSI 10ah cells, biggest mess EVER..

Instead of making it into a brick, how about an epoxy dunk and drip, would cut down on weight quite a bit ( obviously would not be as strong.. but would keep the cells dry ) then box that in expanding foam rubber mix for impact absorption and would be lighter than epoxy ( i think.. )
 
Just pot the negative and pos terminal ...
About to finish a pack and I'll show an example in a few days
 
Here is how I pot them with the Chinese holders

Also about to make another pack with our holders that are 10% more dense, only pos and negative potted.

Yeah this one got a little messy, was experimenting with a 200ml cartridge gun and static mixers, don't use epoxies with particulates in cartridges!!
file.php
 

Attachments

  • 20150718_225049.jpg
    20150718_225049.jpg
    158.4 KB · Views: 13,396
Justin your issue with cutting Kevlar with scissors is oddly enough, because they are sharp and have a smooth edge. You can make awesome Kevlar shears with cheap scissors by lightly pressing both the cutting edges into a disc sander or grinder as of you were trying to blunt the edge. This creates thousands of micro serrations that prevent the Kevlar from slipping away from the edge and it will cut aramid like paper. When it stops cutting as well, "re-sharpen" on the sander.

However, my recommendation would be to skip Kevlar. For what you are doing, a woven fiberglass (not mat) on the bias will get you closer to what you are looking for in terms of real world vs. ideal world production.

Also you may want to try making a DIY syntactic foam. Mix a crap ton of micro ballon filler with a wee bit of epoxy to create a featherweight dough that you can do some cool stuff with.

LFP's comment on a heated cure is to be noted as well. I have built many types of chambers for composites curing and it can range from a styrofoam box with some light bulbs to PID controlled IR heat lamps. My favorite though is a 12 hour soak in a closed car on a summer day. This step is generally only worth pursuing if your epoxy and product actually requires it.

Epoxy aside, how about a silicone fill?

The batteries have no need at all to be stiff, they have their own shell to provide structure. You need to have compliant cover that protects the existing battery casings. Also, if designed correctly, you can drop the cells into the molded silicone and the pack will be 100% repairable. Silicone is cool like that.

My thought is that you don't need to design a cell phone case, just the bumpers. You don't win the egg drop contest by fiberglassing the egg to an anvil. :)
 
I am interested to see what happens to these potted battery's when they fail as a fiberglass potting would surely keep everything trapped in pretty well.

If the cells cant vent they may end up exploding especially if being overcharged by a failed charger.

I remember reading how someone managed to make a headway cell explode with a forklift charger he eventually realized he had blocked the release valve on the cell with too long a bolt.
 
Back
Top