I think I let my A123's starve to death

A

ALLVLTS

100 mW
Joined
May 10, 2012
Messages
41
Location
MD, USA
Hello ES, I haven't posted here much because I'm trying to build something first, then talk about it later... in case I end up not being able to finish for whatever reason. However, I need to ask about this because it's somewhat specific.

So... along those lines of not being able to finish something for various reasons... I started my project almost two years ago and many things have happened to put it on the back-burner, mostly health related. Unfortunately I thought it a good idea at the time to jump on a lot of new A123 26650M1 cells for a reasonable price, so I did. I bought 74 3S packs from migueralliart, which totals 222 cells, with the plan of breaking the packs apart and building a 216 cell pack with them and to have 6 extras for testing and/or backups. Each of these 3S packs came installed with a BMS board (you can probably see where this is headed). When they got here I checked out each pack and they almost all had a nice consistent voltage of 9.5x.
So long story less long, they got shelved before I had a chance to break most of them up which didn't concern me because I've read very good things about their shelf life. Well, I completely forgot about the BMS' ability to suck them dry even though I KNOW better. Stupid, stupid, stupid. They were draining for 1 year and 9 months.
Fast forward to the near-present, I dug them out of hibernation because I was reading something unrelated that made me remember and realize what I had done. I went through them all... and have been kicking myself in the ass ever since. :x :x :x

The BMS boards are now all gone so they're individual cells again. Actually, 24 of them have been since the beginning because I broke those apart at the time to mock something up. Those 24 are all resting right around 3.2V, all fat dumb and happy. The other 198 though...

Strangely, they're all over the place... mostly not a good place. There was no consistency between cells even in the same positions of each pack, so I don't know what the deal was with those BMS boards.
Rather than posting each individual voltage here, I'll group them.

3.00+V............x4 (not worried about these)
2.80V-2.99V......x55
2.50V-2.79V......x59
2.25V-2.49V......x29
2.00V-2.24V......x29
1.50V-1.99V......x19
1.00V-1.49V......x3 (lowest being 1.09V)

Obviously some of these should be okay, obviously some others should be paperweights. My question is... What is the best way to go about seeing how many are salvageable and what their reduced capacity and discharge capability is and seeing how many are entirely gone? I did slow charge one of them from 2.10V to 3.45V, it didn't get warm and it's resting now at 3.32V. That's a good sign I think, but of course it means little as to its usefulness and I haven't tested further than that.
I've been looking at some of the larger RC chargers that do IR measurements among others, but I'm wondering if it's even worth bothering with it because these won't be made into a pack now, and that makes the endeavor almost pointless. I'm just super bummed about it. I can think of a couple uses for a few here and there, though, so maybe all is not lost. If nothing else, I'm really curious to see the data, but I really don't have time to play around if it's not worth it.
It's not a quick process to go through 198 cells any way you go about it, but is there a way that makes it reasonably doable? I'm willing to spend a decent amount on a good RC charger and a PS because I can still use it in the future, but I'm not sure what the best one would be for this situation.

As you can tell, I'm endlessly toiling over the situation, so I'd like your thoughts.
On the bright side, battery technology is always improving and even now there are some very impressive 18650 cells out that have higher capacity, higher voltage, weigh a lot less, and are of course smaller, and are equally as robust and safe as the A123 cells. I won't make the same mistake again though, I'll wait to buy into it until I'm much closer to using it.

Even if you have no input for me, I appreciate you reading about my $1332 epic FAIL. :oops: Maybe someone can learn from it.
 
Sounds like a123 is lifepo4 with a 2.8V minimum. Making 55 more perfectly fine. I would prove the batch of 19 are rubbish by testing 3, expecting each to fail on a high discharge capacity test. About the only test I can do easily. That would give me some direction. Perhaps next I would sample 3 from the lower 29 batch. If all 3 fail again, the 19 are certainly going in the bin. Then test the upper 29. If all 3 fail again, the lower 29 are all scrap.

Hopefully such a system will paint you a picture relating to standing voltage within as few tests as possible. Maybe scrap the 3 samples idea, and just do one till you find the batch where the action is.



It's an odd spread of voltages. It often is. I would happily believe there is some sort of migration effect between cells.
 
Cells seem better able to recover from a deep discharge much better when that discharge was very slow. Plus, lifepo4 seems to survive better. Plus, A123 batteries seem very durable, going on other peoples' experiences. So I think there is a good chance you can recover many of them with little harm.

I have read rumours that many of the grey market A123 20Ah pouch cells were discharged to zero volts before A123 disposed of them. If that's the case then it seems they can be recovered to approximately full capacity.

Definitely worth buying a charger/discharger and testing them :) Discharge test will test for capacity, IR will test discharge rate and you can also test how much they self-discharge.
 
the cells that discharged down to 1V are the only ones at real risk of being lost, but only if you try to charge them individually with a large single cell charger. so not much damage done.

so first explain what you have to use for charging them back up.

if you have a large single cell charger like a 10A 3.65V charger then you will have to gang up all the cells together in parallel that are under 2 V and then charge them up with a big power resistor inline with the charger until the cells reach about 2V. then you can charge all of them up to full voltage. if you use a small cell phone charger to charge them up then you can charge up a few together in parallel instead of a lot, to 2V and then after that charge as you wish.

did you throw the BMS away? you should retain them so you can use them later if you build a pack.
 
friendly1uk said:
Sounds like a123 is lifepo4 with a 2.8V minimum. Making 55 more perfectly fine. I would prove the batch of 19 are rubbish by testing 3, expecting each to fail on a high discharge capacity test. About the only test I can do easily. That would give me some direction. Perhaps next I would sample 3 from the lower 29 batch. If all 3 fail again, the 19 are certainly going in the bin. Then test the upper 29. If all 3 fail again, the lower 29 are all scrap.

Hopefully such a system will paint you a picture relating to standing voltage within as few tests as possible. Maybe scrap the 3 samples idea, and just do one till you find the batch where the action is.



It's an odd spread of voltages. It often is. I would happily believe there is some sort of migration effect between cells.
Click to expand...
That's a good way to go about testing these...some number of samples from batches. I should have thought about that, duh. Thanks. :)
Actually I remembered that there is one interesting thing that stood out about the voltages. The most extreme cells on either end were in some of the same packs. IOW, the most extreme pack had a 3.16V cell, a 3.15V, and the 1.09V. It's like the one sacrificed itself to save the other two, heh. There were one or two more similar packs. Other than that, they really were all over the place.

Punx0r said:
Cells seem better able to recover from a deep discharge much better when that discharge was very slow. Plus, lifepo4 seems to survive better. Plus, A123 batteries seem very durable, going on other peoples' experiences. So I think there is a good chance you can recover many of them with little harm.

I have read rumours that many of the grey market A123 20Ah pouch cells were discharged to zero volts before A123 disposed of them. If that's the case then it seems they can be recovered to approximately full capacity.

Definitely worth buying a charger/discharger and testing them :) Discharge test will test for capacity, IR will test discharge rate and you can also test how much they self-discharge.
Click to expand...
Good to hear that the slow discharge may help them recover better, I wasn't aware of that.

Yeah, A123's are some of the most durable and reliable cells out there, that's why I went with them. I agree that if anything can be saved, it's these. If those rumors are true, then these are even more amazing than I thought. I was under the impression that pretty much any chemistry would die a permanent death if they were 100% depleted. If it really did happen, I wonder how long they sat at 0V before they were revived.

It's not so much the level of discharge that I'm worried about as it is the length of time they've been at that level. While I don't know where they were at say, a year ago, it's probably wasn't too pretty either. In reading about it, I was finding several people (here and other places) that said the longer they stay at a too-low voltage, the less likely they are to recover, at least to reasonable capacities, and they were speaking in terms of only weeks and months, not years. I haven't been able to find any real-world experiences though. That's partly why I'm interested in getting the actual data.
One other thing going for mine is that there were no temperature fluctuations as they've been in my house the whole time. And ~70*F might not be the perfectly ideal storage temperature, but it's not bad.



dnmun said:
the cells that discharged down to 1V are the only ones at real risk of being lost, but only if you try to charge them individually with a large single cell charger. so not much damage done.

so first explain what you have to use for charging them back up.

if you have a large single cell charger like a 10A 3.65V charger then you will have to gang up all the cells together in parallel that are under 2 V and then charge them up with a big power resistor inline with the charger until the cells reach about 2V. then you can charge all of them up to full voltage. if you use a small cell phone charger to charge them up then you can charge up a few together in parallel instead of a lot, to 2V and then after that charge as you wish.

did you throw the BMS away? you should retain them so you can use them later if you build a pack.
Click to expand...

All I have to charge with right now is a Chinese 700mA 3.65V wall-wart that was listed for LiFePO4 (cc/cv). I don't know how accurate the 700mA spec is so I plan on checking with the DMM shortly, but it does put out 3.65V . Under normal circumstances, these cells are rated for a standard 3A charge, which is about 1.3C. 700mA is about a .3C charge, does this sound low enough? I can certainly parallel some closely matching cells if lower would be better, or I could also throw a resistor in series, we have plenty of options at work. If slower is better, then I don't care if it takes a few days to charge some of the lowest cells.
I'll have to figure out a good load to high-discharge single cells. Of course that could mean as much as a constant 70A (definitely some stout cells).

So what kind of 'real' RC programmable charger/tester would you guys recommend? Ideally I'd like for it to have a 110VAC input but I realize that that is pretty rare, so I'll get a stout 12V PS if I have to. What's the highest series cell count available in one of these? I've looked into it a bit but haven't had a lot of time to get down to any details yet. I've seen 12S, and maybe 16S? dual port or four port would be even better if they existed with a high cell count. We're talking real money now, though. I'll spend as much as I need to, but it needs to be something that can do what I need it to do in this particular case. For future use, something like that would be more than enough.

I still have some of the BMS boards but they only have a max discharge rating of 9A which doesn't come anywhere close to what I need. I guess that would make sense if I can find some (relatively) low draw uses in 9.9V increments.

Thanks for all the help so far, it's much appreciated!
 
I have taken the a123 26650 cells in my packs down to .6 once and .8 volts another time by draining them to LVC and then leaving my controller on for a long weekend. I only lost a couple of cells. I am still using the rest some year plus later with little if any reduction in range. I would think charging the weaker cells up to match the 2.5, 2.8 units with the wall-wort would be a good idea before you charge the whole pack up. Might want to drain the 3.0 units down to near what the others are standing at as well. But to be kind to the pack you should balance charge after you get them close to get all the cells back to normal before you bulk charge again. Me I just bulk charged them in 20 minutes or so like usual and replaced the couple mentioned earlier that died.
 
Biohazardman, wow, that's pretty low. Still, they were down there for a weekend, not over a year and a half. The problem has to do with something like the anode and/or cathode material decomposing over time when there isn't a proper charge keeping that from happening. I forget the details so I need to go find and re-read the research papers I found somewhere. (My short term memory is not good, it's the quickest thing to go when you have sleeping problems [among others]).
These aren't in a pack yet, I wouldn't bother before I fully test them. I'll have to temporarily wire them up in xS1P strings to charge/test with a good balance charger, but it won't be a permanent configuration.

dnmun said:
nonononononononononnono

do not put these cells into a 20S string and bulk charge. do not do that.
Click to expand...
Don't worry, I'm a little more edjumacated than that, largely thanks to this forum. ;)
I think he meant that he would hope for 120 cells to be good so he could later make them into a 20S pack. He then meant that prior to making it into a pack, I should parallel charge them (still, I won't parallel bulk charge them from this state either... that's why I'm asking about a good programmable balance charger.)
 
It sounds like the iCharger 4010 Duo is the cat's meow. It has great reviews everywhere I've looked (including here). Looks like fun... Expensive, but fun. :D

http://www.icharger.co.nz/Products/4010-Duo.aspx

The power output capability is serious overkill for me, but it's the only one I've been able to easily find that will do everything I want it to... and accurately, reliably, and with 2 channels of 10S each. I think I might try to put together 2 10S1P solderless balance-tapped charging fixtures, then all I have to do is add 20 cells at a time and let it do its thing. After 12 total cycles, every single one of the cells will have a full data set associated with them. I haven't calculated how long each cycle would take with a specific power input, but I wouldn't need to max the thing out to do it all in a reasonable amount of time I think. I need to figure out how long this would take with my two dinky Lambda 250W 18-30V DC power supplies wired in parallel for 500W, but they might be enough for now because the charging current will be low.

Thoughts?
 
Oh wait, who am I kidding? Just one of my power supplies would be laughably overpowered for this task.
Even if I charged at only 500mA (per channel), charging from nothing might take as little as 5 hours. I don't know how long balancing would take though, or the discharge current I should set to get a capacity measurement. Then of course it would have to recharge again. I would think that that cycle could happen within 24 hours, no? If so, I could do one set per day and have this knocked out in about 2 weeks. In the meantime I can pull up some of the lowest cells with the wall charger. Does this sound feasible and reasonable?
 
If the cells were drained slowly, there is a good chance they will be fine once charged. Just be sure to charge at a very low current until they get over 2v. I have a bunch that sat around for years and most of them recovered just fine.

If you charge them to the nominal 3.7v full, then let them sit for overnight or a couple of days, then measure the voltage, any that drop back down to below 3.3v are suspect.

I put one cell in a solar light I have in the back yard. Every night, it runs down to about 1v before the light goes out. It's been cycling like this for about a year now and seems to still have the same capacity.
 
dnmun said:
you need JST plugs to use the balancing chargers. if you have that single cell charger then just charge up to about 3.35V and leave them alone.
Click to expand...
I have to wire up something regardless because I can't spend months on end slow charging individual cells, which is all that I have the ability to do right now. Wiring JST plugs is cheap and easy (I do something similar for a living, only this is a whole lot simpler).

You're saying to parallel some or all of the cells to charge with the single cell charger? I guess that might work, but wouldn't it be iffy to parallel charge cells that are so far out of balance? Even if I group them into like-voltages, I'd still need to do many of them at a time to make it worth my time because that is my most precious commodity. That puts the voltages pretty far apart from highest to lowest in any group, which will essentially quick-charge the lower voltages (which are already too low) and quick-discharge the higher cells (which are also too low in most cases). I suppose I might be overly-worried about it.
Plus, it won't give me any idea of how damaged some of the cells are or at least how much capacity they lost. Without going into details right now, I'll say that my plans require a super-reliable and very capable and powerful pack that will last a long time (hence the A123's). If all of the cells were similarly discharged I could do my own tests on several random samples and be reasonably sure that they were all similar, but since they are all over the place, if I'm still going to build a pack from these, then I need to know where they all stand.
 
fechter said:
If the cells were drained slowly, there is a good chance they will be fine once charged. Just be sure to charge at a very low current until they get over 2v. I have a bunch that sat around for years and most of them recovered just fine.

If you charge them to the nominal 3.7v full, then let them sit for overnight or a couple of days, then measure the voltage, any that drop back down to below 3.3v are suspect.

I put one cell in a solar light I have in the back yard. Every night, it runs down to about 1v before the light goes out. It's been cycling like this for about a year now and seems to still have the same capacity.
Click to expand...
(I was writing the last post when you posted this)
How much is a 'very low current'?

With each experience I read, I am getting less and less worried about it. However, as I mentioned just a bit ago, I need to know where they stand.
 
split the 500mA between the 3 low cells and it will be ok. those at 2V are ok to charge at 500mA. imo.

the A123 seem to be remarkably forgiving if you read duane's post about pushing them into the grave and then they come out again. however he has killed some too. pop pop goes the pressure relief top.
 
Nobody told him to build a 20s pack and bulk charge. Sooooooo. There are many steps to building a battery pack. As is being said.
But first you must slowly bring them back. Them let sit to see if self discharge.
Now this is where to start on a long road to hopefully building a 20s pack. Or slow your roll.
 
ALLVLTS ,

I'm glad you posted about this here so that people understand to disassemble these packs and leave bare cells in storage voltage.

What I usually did with these cells was cycle them once before sending them out. At the time I had a massive harness I used to charge them and then I had a resistance to discharge each pack quickly. What I sent you was a fully proven pack with the stated capacity so I doubt you have dud cells. Even if you left them in storage and they got low in voltage please fully charge em and discharge once to make sure they are good or bad. I'm pretty sure the majority of your cells are still good.

Here is an example on how tough these cells are;

When people ordered packs from me this was my usual procedure-

Went to storage and obtained the packs that were sitting at nominal voltage, something like 9.2-9.4.

I always took the packs that had less than 9v and stored them separately.

Out of hundreds of packs only 10-15 packs were actually below 9V.

When I got bored and I wasn't selling too many cells I got around to those packs that were low in voltageand disassembled them to build myself a smaller pack for my ebikes.

Surprisingly NONE of those cells were bad they even put up with 10C discharges easily. I even had cells in the 2V range which I charged back to 3.65v and still had the capacity in my pack.

If you happen to need some other cells pm me and I'll send you some to help you out on my dime :D.
 
Thanks, I appreciate the offer! I'll keep that in mind in case I only need a few. Thanks for the additional info about these packs.

I tested the single cell charger. Dunno why, but it puts out 3.796V now. I thought it was 3.65V. Under normal operation, it does indeed put out 704mA. However, if the load is low enough, it will go as high as 1.6A, but at a greatly reduced voltage. I don't think I want to use this charger to charge the 198 cells.
BTW, I am using only Fluke DMM's. The ones at work are calibrated and my personal Fluke 179 matches those exactly.

Something strange happened with 2 of the 'known good' 24 cells that have been loose. I thought I had checked through all 24 a couple weeks ago when I pulled these all off the shelf, but I might have missed a couple... or not? Yesterday before work I grabbed test cell #2 (T2) to bring with me to test, along with that one from the other 198 that I charged from 2.10V to 2.45V (we'll call that A1). A1 is still resting at 3.32V (good news) and I expected T2 to be at 3.2x like the others in that group, but it was sitting at 0.738V! I then had to leave it and got back to it 2 hours later, and it was at 0.499V! I have no idea what happened to that, but I decided to charge it. It was low enough that the charger didn't kick in until I lightly touched the contacts a couple times, to create some resistance, before it started charging and I hooked it back up. I followed that one all the way. Here's T2's charge data, in case you're interested:
Starting voltage- 0.499V
2.81V charging at 704mA, after 30 seconds
3.00V, 704mA, 2 mins
3.20V, 704mA, 7 min 40 sec
3.30V, 704mA, 11 min 50 sec
3.40V, 703mA, 22 min 30 sec
3.44V, 687mA, 44 min 0 sec
3.50V, 574mA, 2 hrs 52 min
3.60V, 376mA, 3 hrs 46 min
3.65V, 280mA, 3 hrs 52 min (max charge spec)
3.70V, 190mA, 4 hrs 2 min (just figured I'd push it a little further since it's a test cell anyway)

As of right now (20 hours after charge termination) it's sitting at 3.35V. I'm going to leave it alone for a couple days then check it again.
When I got home I went through all 24, and T3 was sitting at 2.11V. That one just finished charging, but I didn't follow its progress. The remaining 22 were at 3.2x
I just put together a discharge setup using 0.5 ohm for the load, and I'm going to discharge A1 (3.33V) and follow it. Will keep posted.
 
also, don't use just one charge discharge cycle to evaluate how far gone the cells are. if you do several charge cycles on these i think you will find the capacity increases with each time you do it. these were new cells and only had the formation charge plus the cycle he put on them.

make sure you charge them to full charge of 3.65V and discharge to 2V. you should have to monitor the cell voltages during the discharge to prevent any of them from dropping through the 2V floor. then you would remove the low cell and connect the remainder together again and continue the discharge until the next cell hit 2V and then rinse and repeat until you have a handle on how strong the remaining cells are in terms of capacity. all from the total Ah discharge from the watt meter which will continue accumulating the Ah until you disconnect the auxiliary battery power source.

it is easiest to do the discharge cycles by assembling a number of the cells in series. use what ever number is convenient to fit on your workspace. i use 12 at a time since i can charge 12 at a time in my single cell charger and set of six twin lead jumpers on the ping pouches i work with.

then link them in series and discharge through a good wattmeter into a heater. i use a radiant room heater and oil bath type radiators, pelonius type. do not use the switches to turn the heaters on and off during the discharge. it will burn up the contacts and they will not function after that.

i recently bot some of these and like them a lot. half the price of the turnigy watt meter but are identical inside. the best part is that they do not start the timer until there is current flowing but the turnigy timer runs all the time even when there is no current flowing so i consider these to be superior in design, but half the price. get a little 2 wire JST plug with wires and use an auxiliary battery connected by the two pin plug.

http://www.ebay.com/itm/191010954651
 
Thanks dnmun for the advice and for the reminder that cycling new cells a few times increases their capacity a little.
As I mentioned before, I don't have time to 'manually' cycle through all 198 cells, even in groups (and I do want to go through all of them). That's quite an involved process.

So, overkill or not, I just ordered an iCharger 4010Duo and a few extra 10S balance extensions to make harnesses out of. I read a lot of info about it and read through the manual, and it will do all I'll ever want it to and then some.
As I mentioned the possibility of before, I'm now in the process of designing two simple 10S solderless charging fixtures so I can simply swap in 20 bare cells at a time and just hit a button so it can do its thing while I'm doing something else (namely work). I'll just use one of my 250W power supplies since this won't take a lot of power. It'll be exciting for me to find out how many of these cells are still good and will be excellent to know the nitty-gritty on every one.

I did do a discharge test on A1 which is the one that was originally at 2.10V. It's looking good as far as I can tell. I'll have to get into the details later though, I've got to get back to work.
 
you can wait to cycle the cells when you finally build the battery. there is no need to do it to the individual cells if they are all new and the same lot. but you have to bring them all into balance first before assembling them into a pack. but that is easy to do by connecting all the cells together in parallel so the cell voltages are all identical.

cycling the cells one at a time is just too time consuming and liable to operator error. you are just better off building the pack up and when the BMS is installed it will protect the pack from your testing.
 
So here's what I did with A1...
I topped it off again and let it sit, it settled to 3.36V.
I put a .5ohm resistor load to it for a 2C discharge and put the first Fluke to the terminals for voltage and the second Fluke inline for current.

1 second........ 3.25V, 5.68A
1 min........... 3.19V, 5.67A
5 min........... 3.16V, 5.56A
10 min.......... 3.14V, 5.47A
15 min.......... 3.11V, 5.42A
20 min.......... 3.04V, 5.24A
23.5 min....... ~2.00V, ~4.00A

Man, these things really drop off a cliff after ~2.5V. That's not news, but it really puts things in perspective when it's in real time. The last measurement (right when I disconnected) is pretty close, but it wasn't possible to catch exact values when they were dropping that quickly.
The cell got up to about room temp by the end. That is to say, it wasn't cold to the touch any more, but not warm to the touch either.

So tell me how screwed up my math is... This is, of course, a rough estimation, but it tells me that this cell, that sat at 2.1V for up to 1.75 years, still has some sort of usefulness to it. :)
The average of the first 5 measurements (ignoring the last) was 5.47A (aka 5740mA), and that went for 20 minutes, which is .3333 hours. If you multiply 5740mA by .3333 hours, you get the equivalent of about 1820mAh for a cell that was at nominal voltage and had been cycled essentially two other times.
The voltage sag seems a bit much compared to the datasheet, but not by a lot. It looks like it should hold 3.20-3.25V at a 5A discharge.

I don't know how that all compares in real life to a new new cell, but it shows that it's not toast, at least. I plan on doing the same test to one of the 24 that stayed at a nice storage voltage for a direct apples to apples comparison, but I might not get a chance for a few days. I certainly won't do any more 'manual' tests beyond that, considering that I have an automatic portable battery test lab on its way. :)

Here is my testing-in-progress ghetto discharge setup.
 

Attachments

  • sIMG_1632.jpg
    sIMG_1632.jpg
    237 KB · Views: 1,117
When the cells have been sitting around for a long time, they sort of get lazy. After you cycle them a few times, you should see the capacity increase a bit.
 
You must log in or register to reply here.

Similar threads

sgtkahuna117
I think my scooter lead acid batteries died, I want to go with Lithium
2
Replies
35
Views
856
RTLSHIP
R
M
Help with first pack build
Replies
10
Views
547
mynameisjoe
M
W
BMS compatibility with new controllers and battery build 16s9p
Replies
14
Views
258
WonderDrift
W
nefles
Slim solderless 216 Wh 13S1P pack knocked me off my horse (21700 Moli P45B)
Replies
2
Views
225
nefles
nefles
N
    • Spicy
    • Like
I had a lithium fire (my own fault)
2
Replies
39
Views
2,493
JRP3
JRP3
Back
Top