Guides/tips/etc for building pouch cell packs?

[ns-esk8]

I am interested in doing a lipo pouch cell battery for my next battery... I have built 18650/21700 packs, but this is new to me and I have had a hard time finding any sort of semi-comprehensive guide/tutorial/etc (and nothing in the stickied post). At a glance, it looks like this should make assembly easier, but I want to make sure I don't miss something that would have to be optimized or reworked later.

I have seen tidbits here and there, like using compression straps, tab cooling, etc, but don't know when to use what and have seen some mixed info (separating cells vs not, etc) and am trying to learn more. Like... I see lots of references to ultrasonic welding, but don't know if my spot welder would also work (and if so how much power/duration, what method to use, etc). I also want to be able to gauge what is the "minimum" vs. "good idea/worthwhile" vs. "maybe more than necessary"?

If anyone can give me any pointers where I might be able to get more info (or even just more tidbits/things for consideration to start), I would appreciate it!

 

[From-A-To-B]

You’re well on your way! And there’s plenty more good stuff on ES that will further your journey. You are correct - there’s no comprehensive guide on pouch pack building. At least, I haven’t found one. But the info is out there for those to find.

I dove into this topic before building my first pouch cell pack. Since then, I’ve built a dozen pouch packs - mostly for clients (I’ve offered battery services on the side.)

Most people can not weld the tabs on pouch cells, and soldering aluminum is a joke. At least, I laughed and cried when I tried to do this. Instead, spend a lot of time perusing the options for CLAMPING pouch cell tabs. You wind up spending a lot of time fabricating clamping plates and trying not to drop small screws and washers into spaces they really oughtn’t be.

Compressing pouch cells is an absolute must. There’s good advice here on that topic, and some good info carrys over from compressing prismatic cells. Too much is bad. Equal compression across the face of the pack is important. You’re looking to resist expansion more than compress the crap out of a pack. Again: time spent fabricating.

Tab cooling is beyond me. Really, any cooling is beyond me. I just over spec my batteries so they don’t get hot in the first place.

After building a dozen pouch packs, I prefer to build with round cells over pouch cells. For me, it’s so much easier to weld nickel or copper than it is to sort out tab clamping and pack compression. I’d rather build batteries than spend hours building nice boxes that go around batteries, if you know what I mean. But that’s just my style. Some pouch cells really do deliver on C rate, and they have their role!

Lastly, there’s a good ecosystem of parts and knowledge that’s been developed around Spim08HP batteries, as sold by BatteryHookup and others… likely because there’s so many of these cells floating around in the world. Look there for some build concepts, then copy.

You’re asking good general questions - I encourage you to keep reading and come back as your questions get more specific. There’s enough out there to get you set in the right direction.

 

[ns-esk8]

Thanks for the reply. I appreciate all the info.

I guess nickel aluminum sandwiches, etc, don't work .
I'll look for examples of clamping. I'm not much of a fabricator, though :/, so will probably see what's available w/r/t those parts you mentioned.

I also tend to go the overspec route ... batteries overspec'd for BMS which is overspec'd for controller, etc. (I guess that is another question I have... do people bother with discharge BMS for newer pouch cells or just for charging?)

I'll start digging around, I'm just worried I will have blind spots without starting with a good baseline set.
Some of the battery characteristics of newer pouch cells piqued my interest (w/r/t thermal runaway, life cycles, power density, etc). +I'm planning for a pack about 2x the capacity of the largest pack I've built previously, it seemed like they might be a better fit... but worst case, I can always fall back to my cylindricals .

Thanks again!

 

[amberwolf]

The simplest way mechanically to build a pouch battery is to repurpose a used EV battery module made with the type of pouches you want to use. They'll already be mounted, compressed, and interconnected for the most part, requiring less fabrication and making connecting the rest easier, usually. It might end up larger in volume than a custom-built module/pack depending on the module design and how much of it you can remove and still leave the stuff you need.

THe used EV cells will probably still be better than most of the stuff you'd find to buy to build your own with from scratch, as they will be well-matched cells (even if older and no longer perfectly matched) where the ones you buy separately will probably not be matched at all.

Batteryhookup and the like often have such modules for good deals, and sometimes you can get decent deals from automotive breaker sites (but probalby not as good as BH and the like).

REgarding BMS, it depends on the cells you use. If they are well-matched (whcih is atypical for non-large-EV packs or cells you can buy individually), then they will all charge and discharge equally, and monitoring main pack voltage is sufficient until it ages enough to no longer be well-matched, assuming no sudden cell problems. This is what I do with my ancient (>decade old) EIG packs. The catch with this approach is you haven't much idea if there *are* cell problems, until you do manual testing of them. So if a cell does suddenly fail in some way, you won't know other than by whatever symptoms it causes in the system (if you're watching for them, assuming its' not an obvious "doesnt' work").

There are many types of BMS, including those that have bluetooth to an app on a device, or other ways of checking the cells and whatnot in realtime, and/or logging data during use and charge. Some dont' have any balancing, so they just monitor HVC and LVC and shut off output/input based on those. Some are programmable for various parameters. For high current applications there are even contactor-based BMS (so they use a contactor instead of FETs to control input and output).

Overbuilding doesnt' hurt, unless you have volume, weight, or budget constraints.

 

[ns-esk8]

Great, thanks. Volume will be a big consideration, but that could be a viable path, as well.
If I buy cells, I would purchase name brand cells new/together from a reputable Alibaba vendor... am I likely to run into matching issues going that route, as well?

 

[amberwolf]

I'd guess that you're unlikely to get truly matched cells from anybody, even a factory, from the results various people have reported with the packs built from cells separately. (same results the prebuilt packs also often have, of cells that become unbalanced during usage or charge or both, even when the pack is new).

Cells from the same box from the factory are likely to be closer in characteristics than randomly selected cells, but they probably still aren't matched (unless the factory actually matches them to box them up...probably not).

Between the technical and language barriers, based on exchanges people have posted here, it seems to be hard to get across to them that matched doesn't mean equal voltages (whcih is really all some of them do). Really, it means that all of the cell characteristics (internal resistance, capacity, etc) be *the same* (not similar, not within some range, but *the same*). The best a recent poster could get them to understand was the same voltage (which doesn't matter) and less than a certain milliohm resistance (which isn't matched).

If you can get any of the sellers to actually do it, you should post your results and links to them since this is something that's highly desirable for better-behaving packs.

Cell-testing equipment for more than a few cells at a time is expensive, and labor costs money, so either way it's done it adds to cell cost. How much more, I don't know, but it's apparently enough to keep it from being done as a general rule, both for cells sold separately and for prebuilt packs, based on the results people have with those cells and packs for imbalance problems.

But the EV cells do get tested by the EV pack builders / EV manufacturers, for the large-scale big companies selling EV and hybrid cars, trucks, etc., because they don't want problems they have to fix under warranty. So used EV packs / modules are a good source of matched cells. If the module has all the same voltage cells in it, when full, empty, and in the middle of the capacity range, then they are still probably matched in characteristics. (testing under load and during charge gives further evidence of or against this; individual cell testing of characteristics can be done too, but is probably not needed in these cases).

If cells are all different, then they're probably not matched any more, and have aged too much or had other problems. If just one cell is different, you can swap it out or remove it, so the remaining ones are matched.

 

[ns-esk8]

Thanks for the info. That helps a lot.

I think I would always use a BMS for charging and usually active balancing (at least lately)... but wondering if I can get away with using a smaller, relatively cheaper one (EG: ~40a JK BMS) vs. a big high discharge (EG: 200a-300a ANT BMS) w/ quality pouches (not random, purchased from the same production run, at least, but also perhaps not perfectly matched) or if that comes with any risk, imbalance problems, etc (BMS bypassed for discharge). I have used a BMS for discharging with cylindrical cells for safety (mostly paranoid about thermal runaway->fire), but wondering if that might be a waste in this case.

I had seen some test with newer LG cells that didn't do much other than get the wire really hot during a full short test, but don't know how unique that is among modern pouch cells (I think that is what planted the original idea as I used to have a relatively negative impression of pouch es compared to circles).

Also, I don't know how common it is, but I did come across one seller that can honor requests to have the aluminum tabs replaced with nickel, which could simplify things a bit for me . I've got at least a few weeks before I have to figure out which way I wanna go.

 

[From-A-To-B]

Thanks for the reply. I appreciate all the info.

I guess nickel aluminum sandwiches, etc, don't work .

Every cell I’ve worked with had tabs that are much thicker than the .1mm copper we are all accustomed to “sandwiching.” More material to get through, harder to weld.

Great, thanks. Volume will be a big consideration, but that could be a viable path, as well.

If I buy cells, I would purchase name brand cells new/together from a reputable Alibaba vendor... am I likely to run into matching issues going that route, as well?

Amberwolf gave great advice. I always appreciate how well he can structure a post to be to-the-point and pointedly helpful. Good on you, pal.

Used EV modules will likely be better matched than even new cells from Ali. Matching requires extensive testing and sorting. Ali vendors are absolutely not sorting their cells into batches with like characteristics. Oftentimes this winds up being *fine* because we aren’t using cells to the edges of their capacity like big electric cars are. But in terms of value for dollar, used EV packs can be hard to beat (depending on the pack, it’s history, price, and how well it fits your needs, etc.)

 

[ns-esk8]

Ahh, good to know, thanks.
I've done .15 and .2 as well, though .2 takes much longer (waiting for probes to cool). I'm sure it varies quite a bit, but I'm how curious much thicker they will be on a ~40-60ah cell.

I definitely wouldn't expect a vendor to bother with cell matching, but if the cells are from the same production run, how much variability is there and how sensitive are the cells to those differences in a battery pack. EG: if there is a BMS for charge and I have a reasonable low voltage cutoff (total), how critical is cell matching in that context? I want to make sure it doesn't end up not fine . Can I (safely) get away with charge only BMS or does this make the case for discharge, as well?

I will definitely look for what's available with EV packs, thanks, but I also will want to min/max the space available to a reasonable degree, so I expect that could be a big limiting factor w/ options (short term application I'm considering this for is a QS273 'Enduro eBike' build, roughly 4 maybe 5 kWh @ 200a+ 'continuous').

 

[amberwolf]

I think I would always use a BMS for charging and usually active balancing (at least lately)

If the cells are well-matched, balancing isn't going to be used until the pack ages enough to become unmatched (or a cell (group) fails for some reason).

Monitoring (BT via app, etc) is always useful in troubleshooting or just plain keeping an eye on things over time.

... but wondering if I can get away with using a smaller, relatively cheaper one (EG: ~40a JK BMS) vs. a big high discharge (EG: 200a-300a ANT BMS) w/ quality pouches (not random, purchased from the same production run, at least, but also perhaps not perfectly matched) or if that comes with any risk, imbalance problems, etc (BMS bypassed for discharge). I have used a BMS for discharging with cylindrical cells for safety (mostly paranoid about thermal runaway->fire), but wondering if that might be a waste in this case.

Depends. If there is a failed cell or other problem, a fire is possible with any battery (even lead-acid). Some chemistries and especially some manufacturing processes (or really, lack of QC processes, and/or using the parts that failed those processes anyway), are more vulnerable to this than others.

A BMS, even if it controls charge and discharge, cant' stop a thermal event that's happening, but it can prevent a pack from discharging below any cell's LVC point (as setup in the BMS's programming or hardware), or charging above any cell's HVC point (as set in BMS). Since those are two of the things that could lead to damage that could lead to a fire, a BMS can minimize the possibility of a fire.

So...what to use for a BMS (or to use one at all) depends on your cells, usage, and end-results desired. If the cells aren't well-matched, then a balancing BMS is a good idea to keep the pack "usable" for more of it's capacity (otherwise the growing difference in SoC between cell (groups) will increasingly decrease the available capacity of the pack since some groups will be charged less than others).

If you're not using the BMS to control discharge, only charge, then any BMS that supports sufficient charge current and has the other functions you want will do, regardless of it's discharge current capability. If it could support the necessary charge current, even a "5Amp"-discharge BMS would work, if anybody makes one.

You could even get a contactor-based BMS, without the contactor, and add your own as needed for whatever current it has to support, for charge or discharge.

I had seen some test with newer LG cells that didn't do much other than get the wire really hot during a full short test, but don't know how unique that is among modern pouch cells (I think that is what planted the original idea as I used to have a relatively negative impression of pouch es compared to circles).

Good quality cells that are not QC rejects have a lower likelihood of causing a fire should something go wrong, but there are still ways to damage a cell that can lead to fires later on, though because a fire destroys the source cause evidence, much of what definitely causes fires is unknown. Some things are known to cause fires based on experimental lab information, some is (at least partly) speculation based on chemical and physical changes that happen under some conditions that could or do cause instability that can lead to reactions that cause fires.

So even if a brand new cell, charged up to normal and then shorted out, doesn't burn, that same cell if it is then used for a while normally afterward could develop problems that have undesirable results, because of the excessive current thru (and probably heating of) the cell itself, beyond it's design limits, that may cause chemistry changes or physical alterations in the cell.

Or a cell of the same model that is overdischarged below it's safe limit, then recharged, the changes to the cell that happen during the overdischarge could lead to problems (including fire).

Etc.

Cells today (new, good quality, not QC rejects ) are better designed with each new version, model, etc., but nothing is ever perfect, and there is always variation in the manufacturing process and the stuff used to make them from (which is why there still has to be QC and sorting).


Liveforphysics could speak to all this much better than I can, as he does battery testing (to destruction!) for a living. He's posted some results of some things, but mostly I think his results are for the companies contracted with and probably private, so I don't think there's threads about specific cells, batteries, etc. about his results.

 

[DogDipstick]

Look through my old posts.

Built a ton of packs, all polymer all pouch. I do build with used, and new, but all in sequential ly numbered serial order as they came off the production line. I dont mix and match. I do compress, I have shorted the bus for over 2000A and nothing burned, I have watched them last for years. Evy cell gets isolated form the next, and thew whole thing has a metered tension on the pack. I use pure copper only in the construction ( copper plates, copper rivets, and three types of bus insulation. Some Penetrox goop so it does not lose conductivity over time ( corrosion on the bus).

Not one has failed yet.

Maxxed out every controller I could afford. 20kW +.

 

[DogDipstick]

Moar pics

 

[DogDipstick]

Three bikes.... that did (well ) over 10kW on EV pouch cells.. for less than 300$ in cells. All of these regullarrly ( 20-100+ times per ride) hit 10kW output ( to upwards of 20kw peaks) or more. Thousands of times a week, hundered thousand times a year, or more. They just dont die. no matter what you hit them with.

205v3, NBPower 5kW, 205v3.

 

[DogDipstick]

Pictures.

Cut, stamp, clamp, press, rivet, tape and move on to the next. Wrap the whole thing up i the end, strap it with tension members on plates. Whoolah. Indestructible ebike pack.

 

[frostnova37]

Evy cell gets isolated form the next

Do you have a preferred insulator for doing this? I'm about to build a 26s1p pack of 45ah cells and not sure if I should use foam pads, card, tape or other.