Bringing Headways back from the 0V grave

[adrian_sm]

What does it take to kill a headway cell?

Here is the story.

A couple of us got a bunch of abused headway packs. Some dropped, crushed, shorted, water damaged, generally over-discahrged and left in a pile to rot. Our plan was to rescue the "good" cells, those that showed above 2.0V, and recycle the rest.

But I thought I would try charging up some of the "bad" cells that had <2.0V and see what happens. Some were at 0.0V, a lot at 0.4-0.5V, and various voltage up from there.

The thing is that both "good" and "bad" cells seem to be performing well.

What we have done is:
1) individually charge each cell to 3.6V
- charger is a Turnigy Accucel-8 150W 7A Balancer/Charger

2) Perform a load test to guage voltage sag.
- a couple of big fat resistors on a heat sink to provide a Fluke clamp ampmeter verfied ~24.8 Amp load
- charge up the cell
- let it rest
- apply 24.8 Amp load, measure voltage at 1min mark
- remove load
- measure resting voltage after 2 mins
- infer a DC internal resistance from this.

3) Self Discharge Test
- fully charge the cell
- let it rest for a week or two
- then charge it again, and measure the mAh needed to top it up.

4) Capacity Test
- use the Turnigy Accucel-8 150W 7A Balancer/Charger to discharge at 5Amps down to 2.0V
- record the mAh

The interesting things is that we have found a few duds with high Ri, but most of the "good" and "bad" cells are looking fine.
- DC internal resistance of between 8 and 11 mOhm (based on the voltage sag test above, not the AC 1kHz standard measure)
- No significant self discharge, still requires ~120mAh to top it up after a day, or a couple of weeks.
- All still have greater than 10Ah capacity at 0.5C

There doesn't appear to be any correlation to the DOD we found the cells at, and how good the are now.

We obviously haven't done any long term life cycle testing, but from the "quick" testing so far, it looks like these cells don't instantly die when you take them down to 0 volts, it may hurt the cells, but they still look very useable to us.

Would love to hear the experts opinions on what tests we should do to confirm the health of the batteries, and what it takes to kill them.


- Adrian

 

[dogman dan]

Couldn't be good for the cycle life, but still, a very interesting result. my guess is that if they have failry normal resistance, they might last more than just a few cycles. I suppose it might matter how they were taken down to 0v. Possibly triclking there might be less destructive than getting to 0v in a few min?

Lucky you, to get a stash of these recycled cells.

 

[Jeremy Harris]

Pretty much mirrors my experience. I let three Headway cells get down to about 0.5V (my fault, I relied on a BMS........) and thought they were dead. They sat on the shelf for maybe 6 months until I had an idle moment and decided to try and charge them, just to see what happens. They came up fine and two have been working away (admittedly at a low load) in a home-brew LED torch OK. They seem to hold charge, look pretty much the same as known good cells as far as I've bothered to check and generally don't seem to have suffered from their abuse.

Jeremy

 

[snowranger]

When I bought my Headway Pack, it had probably been sitting for a year. I thought it was toast, but after manually balancing one time, I got the full capacity back. I like that fact that these are hardy cylindrical cells that can stand up to mechanical shock and reasonably high discharge rates at a good price.

 

[fechter]

I have a large pile of Panasonic 18650 Li-Co laptop style cells that sat around for over two years. All of them read zero volts thanks to the "cell protection" circuit draining them.

I've tried reviving a number of them, and really only one was truly dead (it had obvious leakage around the seal too). Most of them showed over 90% of their original capacity, which is pretty amazing considering they were abused to start with and probably over 5 years old.

As far as I can tell from my observations, if a lithium battery is very slowly drained all the way to zero, there's a good chance it can be brought back to life.

I think its when you actually reverse one during discharge that it instantly dies. This is true for most chemistries.

 

[vanilla ice]

Niice! Good to know this is possible. Luckily have not had to try this yet with my fancy li cells, but I have used zombie lead and nickel cells and gotten tons of cycles out of them.

 

[Spacey]

I over tightened one of my 12ah Headway Cells and the top just span around inside itself....totally F*cked.

 

[scoot]

I over tightened one of my 12ah Headway Cells and the top just span around inside itself....totally F*cked.

I take it you are referring to the negative end with the epoxy fill. If so, then the epoxy joint can be be re-done with regular epoxy if you can remove the old material. I have done this myself and you can't tell the difference. However, depending on how much you continued spinning after the epoxy broke loose, it may be a problem.

Back on topic... thanx guys for sharing your experiences with over-discharges. I always have just assumed the worst would happen in these cases, and now maybe not so much as i feared.

 

[heathyoung]

Lithium chemistry recovers OK if the cells have not been reversed - this destroys a cell (even at 0V) instantly. I had an oops moment when charging some of the konion runts (from 0.2V). One charged up fine, but the one I reversed was dead as a dodo - it worked like a nice hand warmer - charge it up, take off the power, and it self discharged quite quickly

If the cells drop very low very slowly, it doesn't seem to damage them as badly as if it was done at a much higher load.

 

[adrian_sm]

Thanks for everyones feedback. It looks like the chemsitry of these LiFePo4 cells is fairly resilient.

The anecdontal evidence seems to be as long as they don't go negative, and they pass the simple Ri, self-discharge and capacity tests they are useable.

We are trying to work out what to do with the cells now. There is no point going to all the effort of building packs, if they are just going to die after a hundred cycles.

I would love to hear if they can hold up after a lot of cycles.


Has anyone been using a "revived" cell for a prolonged period of time?
- How many cycles?
- How are they performing?
- What abuse were they subjected to before "revival"?


Cheers,
- Adrian

 

[scoot]

Well, only my confirmed Headway cell damage experiences have been from passing too much current through a few of them... a little different from the thread topic. But in short, I paired up a couple of these damaged cells and I inserted them one of my 8s2p packs where I could get at them for easy exchanging (position 4 from negative actually). A true" balancing act" now (pun intended). I think they are getting better some over time. Happy to explain further if you like adrian_sm.

 

[adrian_sm]

Well, only my confirmed Headway cell damage experiences have been from passing too much current through a few of them... a little different from the thread topic. But in short, I paired up a couple of these damaged cells and I inserted them one of my 8s2p packs where I could get at them for easy exchanging (position 4 from negative actually). A true" balancing act" now (pun intended). I think they are getting better some over time. Happy to explain further if you like adrian_sm.

Scoot,

Would love to hear more.

I suspect some of my cells have been shorted, leading to a rapid discharge to 0V. So I am quite interested to hear how the cells behave over time.

Please tell me that you have you been logging their performance some how.

- Adrian

 

[amberwolf]

I have a large pile of Panasonic 18650 Li-Co laptop style cells that sat around for over two years. All of them read zero volts thanks to the "cell protection" circuit draining them.
<snip>
As far as I can tell from my observations, if a lithium battery is very slowly drained all the way to zero, there's a good chance it can be brought back to life.

That's interesting. What exactly did you do during the revival process? I have a bunch I'd like to try that with, once I get the likely-good ones sorted. But I don't want to blow them up or start a fire by reviving them the wrong way...I suspect the lightning rods, tesla coils, and other Frankenstein gear is right out.

 

[adrian_sm]

All I did was initially charge them using the NiMh setting at 1.0amps until I got up to about 2.0v.
Then switched over to LiFe setting and charged them up fully.

See the original post for the charger I used.

- Adrian

 

[fechter]

I have a large pile of Panasonic 18650 Li-Co laptop style cells that sat around for over two years. All of them read zero volts thanks to the "cell protection" circuit draining them.
<snip>
As far as I can tell from my observations, if a lithium battery is very slowly drained all the way to zero, there's a good chance it can be brought back to life.

That's interesting. What exactly did you do during the revival process? I have a bunch I'd like to try that with, once I get the likely-good ones sorted. But I don't want to blow them up or start a fire by reviving them the wrong way...I suspect the lightning rods, tesla coils, and other Frankenstein gear is right out.

I trickle charged them at around 25mA until they got up to around 3v (this takes several hours), then slammed them with the regular charger running at about 1/3C

 

[dnmun]

i can add 2 data points

have 2 of the blue 38120S headways that were returned after being severely over discharged under load.

one cell read 0V and was an open circuit, would not take charge at all.

the other cell read .25V and when placed on a 5V cell phone charger, was able to take 40->38 micro amps, cell potential climbed rapidly up towards 4V. removed charger and cell held 3.7 then 3.65, then kept dropping slowly through 3.24 over a few minutes, and then returned to .25V resting after 24 hours. no capacity test made. yet.

all the headways i have that self discharged over a year or so while in storage were restored to normal operation when i charged them. i have not found reduced capacity, but i suspect that the one that dropped to .35V may have lost a little capacity to about 9.5-9.8Ah. that was 1.7 years from manufacturing date without charge, or use.

 

[adrian_sm]

Thanks for the input.

Can you tell me more about why you think one 0.35V cell has reduced capacity. Have you been measuring it? Is it just the weakest cell in a pack? Guess I am just asking if you have some test data.

Did it self discharge to 0.35V over 1.5 years on a shelf? Was this in a pack with a BMS slowly draining it, or just a single cell on it's own.

Thanks,
- Adrian

 

[dnmun]

yes, it sat uncharged from the date of manufacture until i finally got it into a pack and charged it with a ping v2 signalab. the first discharge was only 9.5Ah but i think it has improved since i now have a v1 signalab on it and it seems to have more than 9.7Ah now before i stopped. i get anxious about taking the voltage below 2.1V so i think it is almost unaffected, but maybe significant. i don't consider it a weak cell. all imho

i have others which have never been charged and sat unused for 1.7 years and still have 3.24V.

about 3 were .8V and a few in the the 1.5V range and 4 in the 2.4V range. out of 128.

 

[adrian_sm]

Great data.

So about 8-10 cells out of a hundred appeared to self-discharge.
And so far everyone seems to have reported that a cell that has been recharged successfully from very low, and tests okay for internal resistance, generally behaves itself. Still no one with good capacity or life data for revived dodgy cells. But looks promising so far.

I might just build up some packs with these suspect cells, and give them the stress test on the bike. Doing capacity tests one by one at 5A would take for friggin' ever.

Thanks again for sharing dnmun.

- Adrian

 

[dnmun]

you could charge them individually, then discharge them into a small load, using an ammeteter to monitor the current and voltage with another voltmeter across the terminals to create a simple capacity estimate of the cells if you don't have a CBA. that will tell you a lot, and then you can load test them to determine the internal resistance and try to mix and match from among all the cells so you can spread out the ones that have a high internal resistance so they don't get concentrated in one row if you assemble them in parallel with others.

this is what makes a battery monitor like the CBA so useful.

 

[adrian_sm]

I am already doing the capacity testing with a Celllog8 to monitor voltages, and my Turnigy 8150 charger/discharger to set the load, and recharge.
But it just takes too long. Discharge @ 5.0 Amps, and charge @ 7 Amps, means ~ 4 hours per cell. x100 and that gets old pretty quick.
And I haven't seen too many dud cells so far.

Where as the internal resistance test is quick, about 10-15 mins per cell. And seems to be the best indicator of if the cell is damaged.

So I am tempted to just do the Ri test, build packs, throw them on a bike, and log their performance with the celllog8.
I expect a few dodgy cells will then be found, so I will just have to build the pack in such a way to be able to pull them apart without too much hassle to replace the dodgy ones.

I might just do a bit more capacity testing, on known good and bad cells. See if this tells me anything.

The CBA looks pretty nice. Can it automatically charge and discharge a cell for you, with out intervention?
At the moment I have to discharge, then physically press buttons to charge again. Would be great if that could be automated.

- Adrian

 

[scoot]

...The CBA looks pretty nice. Can it automatically charge and discharge a cell for you, with out intervention?
At the moment I have to discharge, then physically press buttons to charge again. Would be great if that could be automated.

- Adrian

For that kind of functionality in an "off the shelf" charger/discharger, you may want to look in to the icharger 3010b or the FMA Powerlab 8. They are pricey and need a separate power supply, but have some very flexible capabilities including regenerative discharging... meaning you can do some higher discharge rates on the batteries you are testing while filling up large SLA for instance.

Lotsa info on this kind of stuff that can help you out on this topic on RCG forum: http://www.rcgroups.com/batteries-and-chargers-129/

 

[Alan B]

I...

The CBA looks pretty nice. Can it automatically charge and discharge a cell for you, with out intervention?
At the moment I have to discharge, then physically press buttons to charge again. Would be great if that could be automated.

- Adrian

The CBA is handy but it is discharge only, with data acquisition.

 

[amberwolf]

What exactly did you do during the revival process? I have a bunch I'd like to try that with, once I get the likely-good ones sorted. But I don't want to blow them up or start a fire by reviving them the wrong way...I suspect the lightning rods, tesla coils, and other Frankenstein gear is right out.

I trickle charged them at around 25mA until they got up to around 3v (this takes several hours), then slammed them with the regular charger running at about 1/3C

Thanks--I'll try that and see. None of these cells should have died from abuse, just neglect, so it ought to work with at least some of htem.

 

[adrian_sm]

I performed some repeated discharge / recharge cycles on a "good" cell and a "bad" cell. With some interesting results.

The good cell - Cell 003
- 3.076 voltage when found
- Internal resistance of 9.2 mOhm

The Bad cell - Cell 021
- 3.327 voltage when found.
- Internal resistance of 18.0 mOhm

The test consistent of charging at 0.7C, discharging at 0.5C.

The good cell got a very consistent results in the capacity test.
#1 - 10.3 Ah
#2 - 10.4 Ah
Discharged at 0.5C.


The Bad cell had very inconsistent capacity results.
#1 10.7 Ah
#2 9.6 Ah
#3 9.3 Ah

Then for the final discharge test, once I had finished the 0.7C charge, I did a 0.1C charge.
And suddenly I got
#4 10.6 Ah
View attachment 1


So what did I find out from this:
1) My charger isn't the best
2) Internal resistance appears to be a very good indicator of cell health
3) To get reliable capacity test results, I need better control over the charge cycle.
4) A fast charger, combined with a cell with high internal resistance could lead to a much more imbalanced pack, then a slow charger. If the charger isn't any smarter than mine.

- Adrian