This is how A123 care about quality control

[flathill]

really intelligent. quoting a lawyers brief.

if you know that the wrinkles are proof of defect because you read it, then show us the link or produce the paper. or even an author so we can google it.

Just read my previous posts in this thread and google it to find the source. You have to remember for any startup company that uses lithium cells, large or small, whether it is Tesla or Zero Motorcycles, Fisker or Lightning Motorcycles, a battery recall is often a fatal blow to the company. This is serious business, even for the "hobbyist" which is a word I try never to use. Tesla seems unstoppable right now but if they had to recall even a small percentage of their battery packs it would be game over. This is the real reason for the gigafactory. To control their own destiny.

Choosing a cell supplier is THE most important decision you will ever make if you ever sell or give away a product that uses lithium cells.

 

[flathill]

http://www.google.com/url?sa=t&rct=...DA05AE7vxNqElfh-Q&sig2=EpH5qkbGukRsRXsEH0LqXg


That .pdf provides some good stack pressure info, as well as the effects of uneven pressure distribution.

So this pretty much supports good practice is to use soft silicone separators with low stack pressure, especially since cells swell and contract with usage, but swell more with age.

This paper not really go into too much detail about uneven pressure but it did hint at non uniform current distributions. From your linked paper:

"This also suggests that the me- chanical stress within the constrained pouch cells is not uniform. Despite the apparent importance of local environment, no clear preference for formation on either the flat faces or rounded edges is observed when comparing all of the disassembled cells from this study (including cells not pictured in Fig. 6). At present, the chemical composition and formation mechanism of the anode surface films are unknown. However, because the stack pressures in the constrained pouch cells are relatively low from a thermody- namic perspective, it is likely that these films form through a ki- netic effect, possibly due to nonuniform current distributions caused by the localized regions of high deformation within the stressed separator observed in Fig. 7. Full characterization is beyond the scope of this work and will be the subject of a subsequent study."

See the ealier paper I quoted from the Boeing fire investigation:
"Dendrites likely nucleate from areas of surface inhomogeneities or disruptions in the surface layers which lead to non-uniform current distribution during charging. In this case it appears the dendrites are nucleating at locations where there is significant variation in the local stateof-charge originating near the wrinkles seen in the electrode foils. This could induce localized overcharging which would not necessarily be detectable in
the cell voltage."

 

[regmeister]

http://www.google.com/url?sa=t&rct=...DA05AE7vxNqElfh-Q&sig2=EpH5qkbGukRsRXsEH0LqXg

That .pdf provides some good stack pressure info, as well as the effects of uneven pressure distribution.

Thanks LFP. That is an interesting paper.

I see the co-author Craig B. Arnold, has some further papers on his page here:
http://www.princeton.edu/~spikelab/research/mech_prop.html#109

I also ran across this set of slides, from a talk he gave re: the first paper:
http://www.slideshare.net/PECSweb/pecs-talk-batteries

So for the final application, is the cheap solution to compress with some elastic bands?

 

[Punx0r]

Be careful not to jump to conclude a causal relationship between "wrinkle in outer pouch" that A123 describe and things like "wrinkles in electrode foils" and "uneven pressure distribution" in the linked articles.

The likely reality is those two may have been found to have negatives effects on cell life, it's a separate issue to determine if they result from, or are caused by outer pouch wrinkles.

I would expect A123 to be the most likely source for research data on the subject of whether the wrinkles seen in AMP20 cells cause negative effects because they (obviously) have the most interest in the subject. Otherwise, you trying to apply general principles from independent research that is of varying degrees of relatedness. I mean, if research has been done on cells that aren't current design A123 AMP20s then there are automatically a number of factors you must determine whether or not they are equivalent. I.e. comparing apples to apples.

Also, regarding product recalls I would be very, very surprised if the script of a Hollywood film written 10+ years ago concerning alleged practice 40-50 years ago has any relevancy to common product recall. Leaving aside legislative motivation to do the "right thing", one grisly, avoidable death or accident could potentially ruin the company through bad PR. Actually, forget injury, in today's competitive markets just a loss of reputation for quality due a relatively small number of defects could push a company over the edge.

 

[docnjoj]

I guess from all the stuff I have read here, it seems that like myself, everyone is actually rooting for A123 to succeed. If this thread is considered a suggestion on how to improve quality and perception issues, then it should be taken seriously. Perception is really important, and is taken into account by every major company in the world, at least those that like to demonstrate that they "appear" to be customer worthy. This is certainly not from an engineering standpoint, but from a PR standpoint. Many times if a company fixes how they appear to customers, they fix any perceived problems, real or not.
Just my 2 drachmas.
otherDoc

 

[arkmundi]

Just my 2 drachmas.otherDoc

I'll see those 2 and raise you 1 drachma.... A123 is an OEM supplier of batteries, so its "customers" are the Dewalts and Fiskers of the world. A piddling little supply to a bunch of hobbyists is not really in their revenue chart. That they have made arrangements for such a continued supply through Stortronics is, at this time, still a courtesy. Its not our perception of them, but their perception of us that has me concerned. 1 million cells per year, you get a direct OEM relationship. Less than that and its through Sortronics. Less than 100, and its why bother?

 

[wb9k]

Be careful not to jump to conclude a causal relationship between "wrinkle in outer pouch" that A123 describe and things like "wrinkles in electrode foils" and "uneven pressure distribution" in the linked articles.

The likely reality is those two may have been found to have negatives effects on cell life, it's a separate issue to determine if they result from, or are caused by outer pouch wrinkles.

I would expect A123 to be the most likely source for research data on the subject of whether the wrinkles seen in AMP20 cells cause negative effects because they (obviously) have the most interest in the subject. Otherwise, you trying to apply general principles from independent research that is of varying degrees of relatedness. I mean, if research has been done on cells that aren't current design A123 AMP20s then there are automatically a number of factors you must determine whether or not they are equivalent. I.e. comparing apples to apples.

Also, regarding product recalls I would be very, very surprised if the script of a Hollywood film written 10+ years ago concerning alleged practice 40-50 years ago has any relevancy to common product recall. Leaving aside legislative motivation to do the "right thing", one grisly, avoidable death or accident could potentially ruin the company through bad PR. Actually, forget injury, in today's competitive markets just a loss of reputation for quality due a relatively small number of defects could push a company over the edge.

Thanks for your continued sensibility; I would rather others speak up on these matters than keep expecting people to take MY word for it all the time. I thought the "technical comparisons" to the Boeing incident were particularly inappropriate, given that those cells used a completely different (metal oxide) chemistry. I don't mind sharing that one very big aircraft mfg recently signed an NDA as an early step in potentially going with A123 cells in future designs. I had always thought the selection of a metal oxide was a strange choice for Boeing, but a 5-6 year development cycle will do that to you. Good quality LFP wasn't really well-established yet when they started designing the Dreamliner, and it wasn't an option to change to it once design had gotten well underway.

We had a nice all hands meeting this morning and all signs are overwhelmingly good at the moment. Our cell warranty defect rate is 5 ppm. That's 5 returned cells (with actual defects) for every 1 million cells sold. Not a bad number at all. Wanxiang continues to invest heavily in the company at all facilities, including those in the US. Livonia alone will see roughly $7.5M in capital investment over the next year, focused mostly on updating cell making equipment and processes. All factories globally will be operating at capacity soon, and capacity is growing like mad, everywhere.

Back to work!

 

[flathill]

Be careful not to jump to conclude a causal relationship between "wrinkle in outer pouch" that A123 describe and things like "wrinkles in electrode foils" and "uneven pressure distribution" in the linked articles.

The likely reality is those two may have been found to have negatives effects on cell life, it's a separate issue to determine if they result from, or are caused by outer pouch wrinkles.

I would expect A123 to be the most likely source for research data on the subject of whether the wrinkles seen in AMP20 cells cause negative effects because they (obviously) have the most interest in the subject. Otherwise, you trying to apply general principles from independent research that is of varying degrees of relatedness. I mean, if research has been done on cells that aren't current design A123 AMP20s then there are automatically a number of factors you must determine whether or not they are equivalent. I.e. comparing apples to apples.

Also, regarding product recalls I would be very, very surprised if the script of a Hollywood film written 10+ years ago concerning alleged practice 40-50 years ago has any relevancy to common product recall. Leaving aside legislative motivation to do the "right thing", one grisly, avoidable death or accident could potentially ruin the company through bad PR. Actually, forget injury, in today's competitive markets just a loss of reputation for quality due a relatively small number of defects could push a company over the edge.

The pouch is vacuum sealed to such the the case conforms to the underlying shape the electrode assembly. Like said before it is possible the pouch wrinkles are not indicating foil wrinkles but how can you be sure. Even if the foil wrinkles were not present before a vacuum was applied, when the pouch shrinks it can cause wrinkles in the foil. If the wrinkles are near the tabs there is less to worry about. Clearly if you applied a massive vaccum the whole pouch would crunch up. I'm sure in thae case you would use the cell right since only the pouch is crunched up haha. One way other compnies prevent wrinkles by using a "mold" that precisely matches the electrode assembly including the tabs, and applying a vacuum to one side when the pouch is not yet sealed just to remove excess air ans start the pouch deformation before injecting electrolyte, then perform a second vacuum operation and seal the case, but that would not work for a123 as they cant even seem to align their tabs repeatably. The other way around this problem is just to build thinner/smaller pouches.

 

[liveforphysics]

I thought the "technical comparisons" to the Boeing incident were particularly inappropriate, given that those cells used a completely different (metal oxide) chemistry. I don't mind sharing that one very big aircraft mfg recently signed an NDA as an early step in potentially going with A123 cells in future designs. I had always thought the selection of a metal oxide was a strange choice for Boeing, but a 5-6 year development cycle will do that to you. Good quality LFP wasn't really well-established yet when they started designing the Dreamliner, and it wasn't an option to change to it once design had gotten well underway.

If the cell uses a metal oxide cathode or not, everything except LTO uses the same type of carbon anode in the same way. It's the anode that films up locally from uneven cell pressure distribution. If you can create uniform surface pressures through wrinkles and things, my hat's off to you.

This same anode surface is where catastrophic LIB events typically initiate from as well, yet everyone tends to only talk about the safety of the cathode material that rarely plays a roll in starting an event.

 

[liveforphysics]

Clearly if you applied a massive vaccum the whole pouch would crunch up.

The geometry of the stack actually prevents them crunching up even with very high vacuum. There is so much more area on the sides than the edge, that they become quite rigid from the atmospheric pressure applying clamp load through the Z-axis (as well as all other outside surfaces of course, just with lesser net-force due to less area).

Slightly unrelated, but the solvents also handle extremely low pressures without boiling or off-gassing substantially. Amazingly one can leave pouch cells in a hard vacuum chamber for weeks with minimal if any effects.

Pouches are robust in some ways, and yet very fragile in other ways (like mechanical damage from handling them.)

 

[flathill]

Interesting. I thought they only used a low vacuum during the fill operation of around 40-400 Torr. I assumed if they try to fill it with an high, ultra high, or extremely high vaccuum setting (<.001 Torr, <.0000000001 Torr, <.0000000000001 Torr) the thin cell would just crunch up. Anyone would dumb to attempt this in practice, I was only trying to make the point you can't judge where the wrinkle is occurring simply by looking at the outside of the cell. In any case, even if the wrinkle is on the outside only it will still translate into uneven pressure when clamped making where the wrinkle is occurring a moot point to the end user.

 

[jmac]

wb9k, my hat is off to all of you for 5ppm defect rate. I've never worked in an industry that got better than 99.99 so seeing 99.9995% boggles my mind. In a good way.

I hope people are reasonably pointing at your department and hard work in helping reach those numbers.

 

[Hillhater]

wb9k, my hat is off to all of you for 5ppm defect rate. I've never worked in an industry that got better than 99.99 so seeing 99.9995% boggles my mind. In a good way.
.

."..."That's 5 returned cells (with actual defects) for every 1 million cells sold. ".
..don't get too complacent......It's all relative..
I don't know the expectations of the cell industry, but there are many industries where that would be a concern.
I have worked in industries where that rate would represent 50+ customer returns per DAY !.....and that where every single formal customer return was considered business critical.
EG:-- Think of the food/ beverage industries....a sliver of glass in a jar of coffee ?... A piece of metal in a hamburger pattie,... A chipped glass rim on your beer bottle ?...
And I wonder what the acceptable defect rate in the pharmaceutical industry is ?....one odd tab in the wrong pack ?

 

[liveforphysics]

Interesting. I thought they only used a low vacuum during the fill operation of around 40-400 Torr. I assumed if they try to fill it with an high, ultra high, or extremely high vaccuum setting (<.001 Torr, <.0000000001 Torr, <.0000000000001 Torr) the thin cell would just crunch up. Anyone would dumb to attempt this in practice, I was only trying to make the point you can't judge where the wrinkle is occurring simply by looking at the outside of the cell. In any case, even if the wrinkle is on the outside only it will still translate into uneven pressure when clamped making where the wrinkle is occurring a moot point to the end user.

Keep in mind, as atmospheric pressure stays ~14.7psi peak, even if you went from say 50torr (~1psi absolute) to <0.000000000000001 Torr, you only increased the net forces on the cell from 13.7psi to 14.7psi (~7%).

 

[whatever]

5ppm would include wrinkled cells, wb9k has stated wrinkled cells are fine by a123 standards, but who here would accept wrinkled cells for their own builds, at $70usd per cell?
Wrinkles are totally unacceptable, going buy miro13cars' purchase, were looking at a defect rate of around 40%, if your happy with some wrinkled cells buy from stortronics for top dollar. The wrinkles are a big issue and thus far no data from wb9k saying otherwise. Anyone got a infrared camera and a cell with wrinkles?

 

[Punx0r]

The wrinkles are a big issue

You keep making this claim: please substantiate it.


A degree of rationality is required when assessing any manufactured product. If a big industrial air compressor is delivered with a scratch on the paintwork, that's a defect - the spec for the paintwork doesn't permit scratches. Does it affect the functionality of the product in any way? Does it matter? Does the customer care? No.

Same compressor is delivered with a stub pipe slightly off-centre. The customer may not even notice and the unit works fine. Is it a problem? Yes. Because the manufacturer knows that misalignment could only be caused by the pipe being bent. Testing and analysis of customer returns shows this results in eventual stress failure.

Of course, as an armchair expert I could cite any number of examples with engineered products where a paint scratch or bent pipe either made absolutely no difference or resulted in catastrophic failure. It's very easy to make superficial comparisons to support any claim one wishes.

I think some of the commenters in this thread just aren't realistic. They only see black and white: "this is right, that is wrong, because that's what my intuition says". If FLIR examination of a bare, uncompressed cell shows a 0.5°C delta-T at a wrinkle/vein that shows a difference, but proves nothing. What must be then demonstrated is whether or not that difference matters, which requires testing and analysis. Who is most likely to have done that? To have that data?

None of us should be too credulous in life, but we should also exercise some trust of those who know more about specific subjects than ourselves. For one thing, life is too short to personally verify everything for ourselves. Unfortunately I know plenty of people whose attitude is "everyone except me is incompetent and an idiot".

 

[Hillhater]

Personally, Whilst I would expect the manufacturer to be the best informed and most knowledgeable about product design and manufacture,....but experience has taught me not to expect them to understand how those products perform or fail in any specific applications or general use.
Their expertise is in design and manufacturing, not long term use and maintenance to maximise component life, in a variety of applications.
Often as a "user" I have had to advise a component supplier how to modify his product to extend its service life or prevent failure .
Many manufacturers rely heavily on customer feedback to improve their products and manufacturing processes.

 

[Punx0r]

That is very true, however that information comes back to A123.

 

[liveforphysics]

If FLIR examination of a bare, uncompressed cell shows a 0.5°C delta-T at a wrinkle/vein that shows a difference, but proves nothing. What must be then demonstrated is whether or not that difference matters, which requires testing and analysis. Who is most likely to have done that? To have that data?

The temperature increase at higher pressure locations is caused by local increased usage of active material local to the areas of highest pressure. Anytime you have non-uniform material usage, you're aging the cell prematurely fast at the areas doing most of the work (the places under higher pressure with lowest ionic migration resistance due to having a shorter path between anode and cathode layers). If you always discharge some places first and always charge some places first those areas don't get even wear leveling and hence decay faster proportionately to their amount of use.

If you look at any surface closely enough, it's rough and uneven. That said, pouches on automated lines done right leave production without so much as a crease or wrinkle the human eye can see, just flat smooth surfaces. The pouches that have any visually apparent abnormality in them get scrapped, because it means something went wrong in the mfg process, even if that means it was manufactured correctly but handled roughly at some point.

I guess if your target market isn't EV's anymore and just gasoline engine starting batteries or whatever then it likely doesn't matter how relaxed QC standards happen to be, likely a good match for the application.

 

[wb9k]

Well, we still do HV. Our biggest customer at the moment is still ATBS, who is building packs for Chinese OEM SAIC. One is a PHEV, the other is a full EV. Those guys will probably be switching to the new NMC cells when they become available, but for now they use the Amp 20. Fisker too, and of course the Chevy Spark used these cells for the first year as well. Lots of other HV programs out there using them as well.

I have very good information that says we are not the only guys out there selling cells with wrinkles to EV OEMs. The biggest name in the biz is doing exactly the same thing. That's not to say we're not working hard to eliminate wrinkles, because we are. At the same time, the more testing we do with these cells, the less concerned we get. Their performance is slightly sub-par, but still WELL within spec--and that includes life cycle testing.

 

[arkmundi]

Those guys will probably be switching to the new NMC cells when they become available...

First time I've heard that A123 may be changing chemistry. Will that be nano-NMC, so a step up in performance & longevity?

 

[flathill]

Probably trying to copy Kokam Nano 1.0 and put their LFP experience to good use

Over 10,000 cycles at 80%DOD (2C charge, 2C discharge) and 10+ year calendar life, and higher charge and discharge rates than standard NMC cells. 3C continous charge. 5C continuous discharge. Energy density does not quite match Panasonic NCA but this would still make an awesome motorcycle pack. Imagine if your motorcycle got to 80% after a 10 min charge/shitbreak

It is a NMC hybrid (secret is NMC + LTO & LFP Coating).

(NMC based LiFePO4 Coating + Graphite based LTO Coating )

12,000 cycles at 80%DOD (1C charge, 1C discharge)

 

[whatever]

regarding wrinkling, I've spent quite alot of time searching the net for research papers on wrinkling. Its quite hard to find information on it. There have been some posted already on this thread. Someone previously posted the boeing fire due to wrinkling during manufacture was not relevant to a123 cells. But I disagree I think it is very strong evidence to suggest that wrinkling is a big issue. Lithium batteries are basically long sheets of copper and aluminium . No matter what the surface coatings or polymer separator, there are basically the same thing. If wrinkles have been shown to be detrimental in Yuasas lithiums supplied to boeing, it has consequences for studying wrinkles in other manufacturers of batteries.

This quote regarding beoing lithium fire:

GS Yuasa’s manual cell winding flattening process could create electrode foil buckling in the windings, due to the non-uniform distribution of stress in the windings. Subsequent swelling and contracting of the electrodes during charging and recharging further exacerbated these wrinkles within the winding layers
Read more at http://www.ferret.com.au/articles/news/manufacturing-flaws-led-to-boeing-lithium-ion-battery-fire-n2519525#oYo0mbKj18LezTE6.99

Its an extremely complex subject to determine what effect wrinkling has, some of the variables:
1. how does surface wrinkling relate to subsurface wrinkling, are they linked or not? Probably in some cases they are and some
cases they are not, who knows?
2. the degree of wrinkling would be extremely difficult to study and determine which surface wrinkles/subsurface wrinkles are detrimental and which are not. You would have to go into the length of the wrinkle, its direction, its height etc etc.

All I can say is wb9k is probably correct, minor wrinkles are not worth worrying about, but as a consumer/purchaser of a123 cells, I expect to get cells without wrinkles. Wrinkles have an effect, no one here is saying wrinkles do not have an effect.
There are basically two sides to this story:
manufacturers side: minor wrinkling is no problem, less cells are rejected good for business.
consumers side: pay top dollar you expect to get perfect cells with no wrinkles.
This is what the thread began with, a customer who paid top dollar and got wrinkles on his cells, were his cells useable, yes, was the capacity ok, yes probably was. Would the lifetime of the cells be effected, probably by a small degree, but not worth worrying about. Was the customer happy, no he wasn't. He wanted pristine condition cells, thats why he paid top dollar.

 

[wb9k]

The Boeing cells have a different internal structure than A123, which do not use a flattened wound "jelly roll". Since that was cited as a specific wrinkle-related problem, it does not apply to A123 cells, which do not have a wound electrode stack.

 

[Punx0r]

All I can say is wb9k is probably correct, minor wrinkles are not worth worrying about, but as a consumer/purchaser of a123 cells, I expect to get cells without wrinkles. Wrinkles have an effect, no one here is saying wrinkles do not have an effect.

I'm pleasantly surprised we've reached a point of agreement

Yes, ideally the cells would have no wrinkles in the outer pouch, clearly A123 thinks the same if they're actively trying to eliminate them. Even a theoretical risk is worse than no risk. Even if testing show the wrinkles have no significant effect, no testing is completely reliable and the wrinkles represent a variable that must be control, so best to eliminate them if possible. Hopefully this can be done without raising the price of the cells.

A123 cells are relatively expensive but I don't think they're expensive enough to justify perfection. As with all manufactured products we must strike a compromise between price and quality and we always suffer the law of diminishing returns.