A123 20Ah Pressure Thoughts--and bits for experiments.

I thought it might be productive to start a new thread dedicated specifically to this topic. The big(moose) A123 build thread below seems to have gone quiet with the revelation of just how important it is to put at least some nominal level of evenly applied compression force to these cells. Before everyone just lose heart looking at that 3/4 of a ton compression force number, I think it's important to step back and look at the larger picture here.

This topic has been the matter of debate inside A123 for a long time....probably forever. We've reached a point where I think everyone can agree that some compression is absolutely necessary, and it needs to be distributed as evenly as possible. In my mind at least, these two points have been proven to be non-negotiable--both with data from in house and from places like this. The remaining debate is about how much pressure is enough. The answer, I believe, lies at least partly in the application. Lower pressures should be acceptable in lower-demand applications. I would consider an e-bike to be a very light application for these cells. However, in a full-size car, or a big industrial hybrid vehicle (where A123 is the market leader) cells can regularly see spikes of 20C or greater (50C is not unheard of)on a regular basis. Higher compression force will be needed in these applications to keep cells healthy.

Now, I can't put the verbiage of A123 specs on the internet. I do think I can tell you that pressure specified is to be measured after 72 hours of rest (after banding). My guess--and this is just conjecture on my part--is that the upper limit was derived by assuming there would be no relaxation of pressure at all. This may well be what the machine applies. However, the lower limit is much lower than that. The lower limit I recommend is 10 psi, the number I was using when I first got involved in this conversation here. However, there are those who believe as little as half of that is OK, and after more thought, I can't say for sure that they're wrong. It may be that below 10 psi is OK, but I would expect that the evenness of it's application would be more and more critical at those lower levels.

I'm getting ready to scrap a bunch of stuff in my lab this week. I can salvage some endplates and stainless steel banding material and send it to anyone who wants to experiment with them for the price of shipping. Send me a FedEx account number, and I'll ship by tomorrow. I've offered this and other things in PM's, but nobody has responded. Looking around the board, I see there's a long history of scamming that's gone on, and in that context perhaps that's understandable. But I'm for real. Like many of you, I just want to see this stuff succeed.

Anyone interested in some bits for experimenting, PM me. First come, first served. Just endplates and bands right now....no electronics, no cells--yet.

===========================================================================
start bigmoose edit: I just wanted to add a reference in this thread to the compression numbers required for A123 pouch cells, so it doesn't get lost in the future.

Here is the thread reference for the numbers: http://endless-sphere.com/forums/viewtopic.php?p=773809#p773809

I (bigmoose) calculated the stack compressive force from the "stack pressure in psi" given by wb9k. He (wb9k) said the spec calls for 10 to 25 psi (pounds per square inch) compression on the cells.

bigmoose said:

So folks since a cell is 57 square inches in area that says that the A123 20 Ahr pouch cells want: 570 to 1425 lbs of compressive FORCE on the stack!

Click to expand...

wb9k was questioned by major:

major said:

So you are confirming you need to have almost 3/4 ton force on a cell?

Click to expand...

wb9k said:

As a max, yes.

Click to expand...

Now the data is all in one place
...end bigmoose edit
===========================================================================

OK, Bigmoose has laid claim to some material. Still two sets of endplates and some banding left....

This explains why the a123 OEM packs are banded together with the metal straps aka 'compression band' (''stainless steel banding material''?)



A simple DIY setup would be two plates (one at each end of the pack) and bolts (passing though each plate), use a torque wrench to hit the required pressure point

Thank you for speaking up on this design consideration

I agree with the comporession method.

All the packs i built have compression on them,, LiPo, LiMn and LiFePo4. All are still 100% working today, none have puffed yet :wink: and RI always stay low. so i can charge them at nominal rate or more without problem :wink:

For the lipo i also think it help reduce the danger.

Like i explained many times on E-S in the past, problem with pouch is that once they inflate (puff), the corner and edges on the side will have deformation and this is what might cause them to catch fire if they puff too hard. When deformation occur the conductive anode and cathode layers inside might short together as well as when a cell have puncture theu the layers.

Compressing pouches cells keep them in original shape and also keep layers closer wich help for the RI to stay low :wink:

Most of the lipo fire occur with pack assembled together with ducktape or simple shrink and sometime just inside a nylon bag where they can move and hit edges of the bicycle frame wich is very bad.

Building a case to protect them and hold the pouch together and keep them under pressure on the tickness axis is the key and it worked for me for many years.

The guy o fthe RIMAC car wich i consider as one of the genious in the high performance EV developpment also use that method and most o fthe EV manufacture are also using that method.
now..

As for the cylindrical cells, they dont need compression because they already are under compression!.. cylindrical shape is one of the shape that is the most strong to sustain pressure.

Layers expansion occur when cells are charged for the lipo and they also have an endothermic reaction wich make them to cool down!. once they are discharged they shrink a little bit.

TRICK:

One of the trick i have for pouch cells is to make the battery box with dimension of the cells in their discharged state ( when their tickness is at his thinner state) then when they are installed in their battery box, if they try to inflate furing the charging process, this will make them to autocompress and will keep the RI low and keep them in original shape. then when they discharge, they recover their uncompressed original shape :wink: This cycle repeat over the lifetime of the cells :wink:

It is so constant than i just have to knock on the cell box and the sound pitch give me an idea of teh state of charge :wink:


What are you waiting for guys! :wink:

Doc

PM sent...

PM sent

Haven't actually seen 999's PM yet, but that will about do it for parts I have available right now.

T3sla, yes this is the reason for the 1/2 inch stainless steel bands. Your method sounds like it should work, but I would want those endplates to be non-conductive. A little thought should make it obvious why.....

those black pressure plates in the picture above, are they metal or plastic ?

Plastic.

"When deformation occur the conductive anode and cathode layers inside might short together as well as when a cell have puncture theu the layers."

"Compressing pouches cells keep them in original shape and also keep layers closer wich help for the RI to stay low "
I undrstand now how the heat can generate seperation between the layers when they try and lay down after being heated. How hot do cells need to get to form a gas ? We all see sla bubble during charge

This revelation is interesting to me-- I wonder, how much pressure does the shrink-wrap on Cell_Man's packs actually provide? 50lbs? 100lbs? That shouldn't be enough per 10PSI specs, but his packs have performed well for everyone on here as long as they've used them (right? I've never read a complaint, only praise!) Perhaps it is because our discharge rates tend to be much lower. I'd like to get a 28s1p 20ah A123 pack for a FOC controller/cromotor. But after hearing this I know I'm going to be super anal about getting the right compression... which makes the build process a little more complex, considering I'll be making a huge mess in the living room for the wife to look at the whole time... someday I'll have a garage.

Does laying the pack so that the cells are flat/horizontal (with the compressed face vertical) reduce the need for substantial compression?

etriker, those are 8ah a123 cells! Reasonable deal. I'd make a 28s2p pack. That'd last a loooooong time.

2-4psi through the Z-axis is often recommended clamp load. This adds up to many hundreds of pounds of clamping force for large format pouches.

The price was from A123 website, 70usd per cell, legit. I think you would need to check out enerdel spec sheet, u get 3k cycles and 10% decrease in capacity at 100% DOD. It is at .2C but still nothing close to 1k.
Just my thoughts. It is expensive to build a pack from these meaning properly not a wank job. Add compression to that. I do like these cells, even got the kit made but until they are in 40usd/pcs range I go to competitors.


Regarding compression, glass reinforced fibre board could be used. 6mm thick could be routed into trapezoid shape with thickes in the middle, two grooves milled out for the straps not to slide and strapping tool used for compression borowed from local manufacturer or something cause it would add some few hundreds $ to your build. Perhaps those come with tension gauge which could be calculated by math guru into preasure?!

agniusm said:

The price was from A123 website, 70usd per cell, legit. I think you would need to check out enerdel spec sheet, u get 3k cycles and 10% decrease in capacity at 100% DOD. It is at .2C but still nothing close to 1k.
Just my thoughts. It is expensive to build a pack from these meaning properly not a wank job. Add compression to that. I do like these cells, even got the kit made but until they are in 40usd/pcs range I go to competitors.


Regarding compression, glass reinforced fibre board could be used. 6mm thick could be routed into trapezoid shape with thickes in the middle, two grooves milled out for the straps not to slide and strapping tool used for compression borowed from local manufacturer or something cause it would add some few hundreds $ to your build. Perhaps those come with tension gauge which could be calculated by math guru into preasure?!

Click to expand...

Wow! That's pretty good. I know that A123 cylindrical cells go down 40 percent after 20,000x 100 percent DOD cycles. But that's at 1C discharge. Cellmans packs are less than $50/ cell.

Maybe using a tension gauge on a strap would work?

Cheapest method for that would be 'stainless steel cable ties' they are commonly used in packing to bundle things.
The tension tools lack precision control, but I'm sure you could figure out something.

I have a question for wb9k about the heatsink plates.
Do they wedge between each individual cell or is it a solid encasing surrounding the cluster of cells?
I assume they are only critical in high C rate applications, it does look like they have milled out two channels for the compression bands to allow even pressure

I consider this A123 Pouch Cell compression topic a very important one for us. Both for safety and performance. It may well be germane to all pouch cells.

Therefore I split the off topic posts on A123 purchasing to this thread: http://endless-sphere.com/forums/viewtopic.php?f=14&t=52272 Those posts that had a "significant" content in compression stayed here, even if they mentioned some purchase info. Those that went the other way moved. I would have copied them both places, but the software won't let me do that.

Let's keep this thread on topic in respect to wb9k's contributions.

t3sla said:

Cheapest method for that would be 'stainless steel cable ties' they are commonly used in packing to bundle things.
The tension tools lack precision control, but I'm sure you could figure out something.

I have a question for wb9k about the heatsink plates.
Do they wedge between each individual cell or is it a solid encasing surrounding the cluster of cells?
I assume they are only critical in high C rate applications, it does look like they have milled out two channels for the compression bands to allow even pressure

Click to expand...

The heatsink plates are arranged (mostly) to be between every other cell. If you look closely at the pic, you'll notice the heatsink at the near end of the module has shorter "fins" than the others. Only one cell in there. The sinks are stamped, e-coated aluminum, and yes there are channels for the straps to lie in. Every cell is separated by either a foam pad or a heatsink. These modules can lie on any one of the three heatsink surfaces, which usually is liquid-cooled by a cold plate. You would not want to use a naked module like this as a pack on a bike--the heatsinks are too fragile to be exposed during regular use. They are easily damaged by impact, and the cells are just beneath.

Another poster asked if orienting the cells in a particular way would reduce the need for compression. The answer is no. The cells at the bottom of your stack might do OK, but toward the top, the same problems will develop.

bigmoose said:

I consider this A123 Pouch Cell compression topic a very important one for us. Both for safety and performance. It may well be germane to all pouch cells.

Therefore I split the off topic posts on A123 purchasing to this thread: http://endless-sphere.com/forums/viewtopic.php?f=14&t=52272 Those posts that had a "significant" content in compression stayed here, even if they mentioned some purchase info. Those that went the other way moved. I would have copied them both places, but the software won't let me do that.

Let's keep this thread on topic in respect to wb9k's contributions.

Click to expand...

Excellent idea! :wink:

Doc

What I consider most relevant is tech dialog for these cells & their packs, tending towards expert consensus that might be applied by anyone building a pack, including folks like me. I'm someone who wants to enjoy the benefits of an ebike at lowest possible cost, without sacrificing performance, and in the realm of possibility. It was the combination of low cost grey market cells (victpower) and agniusm's kit for a buildout that led me down this particular path in the first place. So far, I have not regretted the decision. But if I have to re-manufacture my pack to apply requisite compression, and heat-sinks, etc. its just no longer "in the realm of possibility" for me at least, and likely most ES'ers. Unless there is a practical way forward, like a kit, much as was produced by agniusm, but inclusive of the parts needed to apply compression. By way of suggestion of where I hope this dialog might go.

The stainless cable ties I used were not a success, they are the ones with a half ball bearing that wedges in the clip when it is tightened.

They have loosened slightly with vibration. If I started over I would use the Band-it strap and clips, once they are tightened and the strap is hammered back over on itself, it is pretty solid.

The only other method I can think is spot welding the strap when when under tension.

A full width of big plastic cable ties ,say 20 across a pack would probably be quite effective.

I believe you can source notched stainless cable ties that are similar to 'hose clamps' where by once it reaches a certain tension it remains locked.
I've used some SSCT that didn't have the retention strength of some decent quality plastic cable ties

Cheap crap is cheap crap, don't go using bottom of the runt stuff on a high grade battery.

Since even pressure is required, my immediate thought is to put something inflatable to press evenly on each endplate. How hard would it be to use a couple of car or truck inner tubes to make a well sealed bladder for each end to fit the shape of a cell? What I like is that the pressure is adjustable, measurable, evenly distributed, and can maintain a near constant pressure regardless of SOC. Plus air doesn't weigh much at 5-10psi.

John

I have an opinion about the heat sink plates, it is just an opinion. For automotive use, they likely optimize for strong regenerative braking, as that will be a significant range extender in stop and go city driving. Those high currents that go in during braking and go out during acceleration are rising the temperature due to the internal resistance of the cell. Seeing as we operate on much lower average currents we may not have any thermal problem. This opinion is perhaps substantiated by there being no reports of any overtemperature issue with either the A123 20Ahr pouch cells or the cylindrical cells during use.

So all we may need to worry about is the compression aspects.

John I'll bet that the stainless bands are somewhat thin and "springy." I'll bet when they are pretensioned "right" that they operate as an effective spring. What we will need are endplates to evenly distribute the load. Then think of our "tie bars" we don't want the equivalent of 10 gauge spokes that are too "stiff" with no give for the bumps, we want some nice 14/15 gauge spokes that can be "tensioned" and will give with the expansion/contraction of the cell thickness with charge/discharge. I now see that this battery needs "tension" but needs to also "breathe."

Hopefully wb9k will correct any assumptions that are amiss.

bigmoose said:

I have an opinion about the heat sink plates, it is just an opinion. For automotive use, they likely optimize for strong regenerative braking, as that will be a significant range extender in stop and go city driving. Those high currents that go in during braking and go out during acceleration are rising the temperature due to the internal resistance of the cell. Seeing as we operate on much lower average currents we may not have any thermal problem. This opinion is perhaps substantiated by there being no reports of any overtemperature issue with either the A123 20Ahr pouch cells or the cylindrical cells during use.

So all we may need to worry about is the compression aspects.

John I'll bet that the stainless bands are somewhat thin and "springy." I'll bet when they are pretensioned "right" that they operate as an effective spring. What we will need are endplates to evenly distribute the load. Then think of our "tie bars" we don't want the equivalent of 10 gauge spokes that are too "stiff" with no give for the bumps, we want some nice 14/15 gauge spokes that can be "tensioned" and will give with the expansion/contraction of the cell thickness with charge/discharge. I now see that this battery needs "tension" but needs to also "breathe."

Hopefully wb9k will correct any assumptions that are amiss.

Click to expand...

Springs, like on big screen crts. Kinda stiff like valve springs. And bolts or threaded rod that will only screw down so far so the spring is not fully compressed.

(cookie free post)

John in CR said:

Since even pressure is required, my immediate thought is to put something inflatable to press evenly on each endplate. How hard would it be to use a couple of car or truck inner tubes to make a well sealed bladder for each end to fit the shape of a cell? What I like is that the pressure is adjustable, measurable, evenly distributed, and can maintain a near constant pressure regardless of SOC. Plus air doesn't weigh much at 5-10psi.

John

Click to expand...

A mechanically constrained inflated bladder system sounds pretty good. That's clever, I like it.

If you've got a closed cell dense foam of a material that remains elastic, and you crush it, it behaves similar to a bunch of little bladders all constrained in a sheet of silicone rubber (or whatever material).