Self Discharge after Charging

Shawndoe

Member
Joined
Sep 13, 2018
Messages
32
So as I mentioned earlier, I opened my first pack (Dell 11.1V 3S2P) over the weekend. I got 6 Samsung 28As in 3 pairs at 2.2V, 2.6V, and 3.6V. I charged them per the datasheet, and when finished I could see the voltage settling. Most lost 0.05V within an hour, 0.1V within the first 12Hrs, and after 96Hrs all have settled between 4.00V and 4.03V. I feel like these kinds of self discharge numbers imply EOL cells, but I'm not sure, what are good self discharge numbers?

The next pack I opened was a generic 3S3P packwithflesh pink cells. The cells have bar codes, with numbers, no manufacturer, and no part number Google could recognize. One set reads 0.073V, the others read around 3.7V. I haven't done any testing on these, I'm just curious if anybody else has seen these cells like this before?
 
Pretty much anything over .1V in a few days is too high. The settling should slow down over time. So really, if they dropped by .1V within hours of pulling off the charger (as long as they "truly" were at 4.2V to begin with), then they shouldn't be used in a parallel pack in a high power application. Maybe a flashlight would be good use for it.

Generally, if you let them set for a 2 - 3 weeks and they drop from 4.2V to 4.1V, then they are probably still just fine. I've had some cells sitting for months and not drop below 4.10 - 4.15V. But I also have some that dropped to 3.95V after a month. So those won't be used in the powerwall at all.

I've seen some with the barcodes, yes. I haven't done any research on these either. I would assume that "maybe" if you looked really hard, there might be some embossed numbering on the label. Or the real numbers may be printed on the cell itself.
 
Dropping from 4.30V to 4.00V in 4 days is a lot. I've seen some of those 4.3V cells do that where they drop to a voltage and then stay there. But usually it's like 4.2V or 4.15V, and those cells typically have a low capacity.

Post a picture of the generic cells you found. I believe I have a few just like that. Generic packs have terrible BMSs as you can see. You can use the Wolf method on the low voltage cells if you want to test them, but the 3.7V ones should be fine. And dont believe the capacity on the battery pack, it's never correct.
 
Dallski said:
Dropping from 4.30V to 4.00V in 4 days is a lot.

Where are you getting "4.3" from? He's not talking about HV cells, but standard laptop cells. They only go up to 4.2V. Yes, going from 4.3 to 4.1V would be a sign of a bad cell, but from 4.2 t 4.1V does not.
If you mean't 4.2 but those keys are dating each other, make sure to review your post before posting, to help minimize any confusion ;)


Dallski said:
Post a picture of the generic cells you found. I believe I have a few just like that. Generic packs have terrible BMSs as you can see. You can use the Wolf method on the low voltage cells if you want to test them, but the 3.7V ones should be fine. And dont believe the capacity on the battery pack, it's never correct.

I agree, clear pictures are great. Helps us see things you may have overlooked. Plus the camera sometimes can pick up more hidden text or damage that the eye may miss.
Any battery pack that is over a few years old, maybe even over 1 year, will not have the capacity that is on the label. Unless that pack was not used, or rarely used, like medical packs for example. So always check or even double check capacities of harvested cells to verify their true value. Otherwise building a properly balanced pack to begin with will be almost impossible.
 
Down to 4.1 or even 4.05 is not a big deal. Some do it after 1 hour and some after a week. Asko as it stays There its fine
 
Dropping to 4.05 would be acceptable depending on your application, IMHO.

If you want to charge up to only 4.1V to help extend cell cycle life, then that cell(s) that want to drop to 4.05 may become a pack discharger cell. This is why we make sure to test all cells and let them rest for a period of time before we use them
 
Shawndoe identified the cells as Samsung INR18650-28A, which are 4.3V cells. And he said he tested them according to the datasheet, so I was assuming that meant charging to 4.3V.

Also, I probably shouldnt have used the word capacity when referring to the generic packs. What I meant was that usually a 5200mAh label would have 2 parallel cells labeled as 2000mAh inside. And the packs labeled 4400mAh almost certainly contain 2 parallel 1800mAh cells. The best are power tool packs that say 2Ah on the outside and then the cells on the inside are labeled 1.3Ah. Regardless, all cells should be capacity checked.

Lastly, are these the cells you found?


image_recoao.jpg
 
Dallski said:
Shawndoe identified the cells as Samsung INR18650-28A, which are 4.3V cells. And he said he tested them according to the datasheet, so I was assuming that meant charging to 4.3V.

Also, I probably shouldnt have used the word capacity when referring to the generic packs. What I meant was that usually a 5200mAh label would have 2 parallel cells labeled as 2000mAh inside. And the packs labeled 4400mAh almost certainly contain 2 parallel 1800mAh cells. The best are power tool packs that say 2Ah on the outside and then the cells on the inside are labeled 1.3Ah. Regardless, all cells should be capacity checked.

Lastly, are these the cells you found?


image_recoao.jpg

Actually I charged to 4.2V, all other parameters were spec.

That cell is a veritable plethora of information compared to these cells. Check this out:


image_ayisnx.jpg


I can't identify any useful information, and neither could Google.
 
Oh, then I apologize to Korishan, I was wrong. I do actually test mine to 4.35 (opus) so I assumed you did, too, based on your statement.

170311 is most likely the date code, "2017 March 11." Same color cells, though! First number is probably a serial number. Can you see through the PVC wrapper to any markings underneath? Or if not, if one of the 0.07V ones is bad, you can take the PVC wrapper off and check too. Also, the top terminal might give some clues.
 
Dallski said:
Oh, then I apologize to Korishan, I was wrong. I do actually test mine to 4.35 (opus) so I assumed you did, too, based on your statement.

170311 is most likely the date code, "2017 March 11." Same color cells, though! First number is probably a serial number. Can you see through the PVC wrapper to any markings underneath? Or if not, if one of the 0.07V ones is bad, you can take the PVC wrapper off and check too. Also, the top terminal might give some clues.

I didn't see anything through the wrapper, I tested the first set, and 2 cells dropped below 4.00V in less than 12Hrs so I might peel the wrapper off one of those and see what's underneath when I get home from work.

The datasheet for the Samsung 28A is 4.30V max charging voltage, is there a reason you overshoot that 0.05V?

Thanks again
 
Shawndoe said:
............................I opened my first pack (Dell 11.1V 3S2P) over the weekend. I got 6 Samsung 28As in 3 pairs at 2.2V, 2.6V, and 3.6V. I charged them per the datasheet, and when finished I could see the voltage settling.................................

TheSamsung ICR18650-28AIt is a battery I have grown to love and hate. (Along with the ICR18650-30A and B)
I have recorded 297 of the mischievous cellsand justknow a couple of things.
Anycell >55m? is a crap shoot. I have had cells give me a 80% of capacity all the way up to 70m? anything above that is definitely bin material.
And to keep with the topic very likely to bea SD (Self Discharger)
Cells with an IR of <55m? and your chances are reasonablygood you will get a respectablecell. <50m? almost guaranteed you will have a good cell.

You notice I say crapshoot, chances, very likely, andalmost guaranteed. I have not pinned down the exact reason why these cells are all over the place The IR differenceis between49.5m? (my lowest IR testedSamsung ICR18650-28Acell so far) all the way to 70.3m?. My highest IR recordedSamsung ICR18650-28A is at 125m? with a whopping199mAh capacity :p . But As I said anything over 70m? is no good. One theory I have is that thesecells are pushed to 4.3Vmaxand 2.75Vcutoff which creates a high stress environment inside the cell and therefore theyperformpoorly in our tests as these cell have given all they got in a matter of 150 to 200 cycles. That ismy best guess. The degradation of the cells should show up in the IR value of the cell but that is not the case with theICR18650-28A at least not that I have found.

So I am perplexed as to why these cells behave this way. The SDs are more than likely highIR cells so If you could check them for IRthat would give you the potentialanswer.

Just to show you the variation of capacity percentages returned at the 56m? to 54.5m? spread is enough to scratch your head. I have other cells that do not show this behavior. TheSamsungICR18650-26*series andSanyoUR18650F*series and many more. You can give me the IR of those cells and by now I can just about give you the mAh results to within 100mAh. But those cells also run at 4.2Vto 2.75V and3V respectively.

Here is the sheet that makes you go hmmm. At least it does me.

Wolf

image_mgfhvp.jpg
 
Curious: with the IR readings, do you tend to measure at a know SoC or maybe after a charge?
Wondering if the SoC, voltage or a few "wake cell chemistry up" charge cycles makes any difference to the IR reading?
 
Redpacket said:
Curious: with the IR readings, do you tend to measure at a know SoC or maybe after a charge?

The IR readings in the chart are after a charge as you can see by the V to the left of the IR reading. Most of the time the cells have been pre charged and are sitting at this SoC for at least2 to 3 hrs before the IR measurement but more likely a day or 2. The 3.6* cells you see in the chart if you look at the datesit was early in my IR testing and experimenting. That was the V they came out of the pack with.
I now take an initial IR reading no matter what the Voltage is,before I go through the trouble of cleaning up the cell, to determine if they are worthy.
If the cellpasses my criteria then it gets charged by whatever method the cell can handle. Right now anything ~3V or below gets the CC/CV 4.2V, 50mA treatment.
>3V cells get300mA.
I do notice there is some change in IR at different SoC'es for cells. Most good performing cells whether at 1.5V or 4.2V will have an IR where the delta is pretty tight as in <10%. So lets say I have a Sanyo UR18650FM with 1.5V and an IR of 55.2m? this is a good candidate to charge. After charge and a restthe V is 4.15V and the IR is 51.3m?, then after the norm test C/D/C the V is 4.2 and the IR is 54.5m?. Subsequent tests have show that the IR does not change very much(~ 10% if that)on a good cell between different percentages of SoC.

On low performing cells it seems that the IR goes up by more than 10% from the initial "wakeup charge". With over 4000 cells recorded the amount of data is interesting to study. But also time consuming to data mine and find correlations between all these criteria IR, V, mAh, manufacturer, and model#.
I try to collect as much data as I can. The problem is that it seems each cell manufacturer and each part number have different criteria when considering good VS bad. So the search continues. But I will have"some" answers with some concrete measurement criteria for certain cell types in the not to distant future as there is irrefutable evidence piling up that has to be disclosed.

Wondering if the SoC, voltage or a few "wake cell chemistry up" charge cycles makes any difference to the IR reading?

Funny you should mention that. I am doing an experiment on some curious cells that have a reasonablecapacity outcome 75% to 80%but have a quite high IR
I am cycling them an my SKYRC at manufactures spec and they seem to hold and if nothing else increase their mAh results slightly every time. I will run them a couple more times and see what their IR is at the end.

Till then

Wolf
 
I notice you have a SkyRC, how does it's IR measurements compare to your 4 wire?
 
Shawndoe said:
I notice you have a SkyRC, how does it's IR measurements compare to your 4 wire?

The SKYRC is terrible just like the OPUS, Liitokala, Zanflair, etc. There's just no way to do IR right except 4 wire.
I even wrote SKYRC an email asking them how come such an expensive tester is so way off on IR. I have yet to hear from them and I expect not much of an answer when they do respond :p.

I am in the process of testing these 5 testers with 2 more on the way from china. all within the $50 range.All are a variation of the 1kHz AC 4 wire kelvin methode. ( I may include the XTAR Dragon as it has a separate IR function and found it to be pretty close)

image_snlyip.jpg


I will be using these items as a reference.
A 10m? reference resistor came packaged with one of the testers by mistake, I think, but neverthelessits anIsabellenhtteresistor 1%

image_uacrsi.jpg

A Voltage reference board with thehigh precision Analog Device AD584HKvoltage reference IC.

image_mitrer.jpg


Some brand new 18650 batteries and some random 18650s picked out of my stash.

So stay tuned.

Wolf
 
I assume they all measure higher than actual resistances, but are they consistent is the real question. If the unit measures 100m? higher consistently than you can work with that, but if it's 100 this time and 300 the next you can't get anything useful out of that.

What did your 4 wire cost you?
 
Shawndoe said:
I assume they all measure higher than actual resistances, but are they consistent is the real question. If the unit measures 100m? higher consistently than you can work with that, but if it's 100 this time and 300 the next you can't get anything useful out of that.

What did your 4 wire cost you?

I have given up on the consistency check of these charger/testers. I will put it on my list of things to check on the SKYRC though and report back to you.

All the IR testers in the above pictures are at the ~$50 to $60US mark.
The Vapecell is the one I have used pretty much exclusively. I like it. Easy to use and accurate. But I will report on all of them soon.

image_lrarcs.jpg


Wolf
 
Wolf said:
The IR readings in .....

Wolf

Thanks for the details responses, awesome info you're collecting there! :)
 
Just a follow up on those generic cells.

I bump charged one of the near 0V cells CC at 50ma to 2.80V until current fell to 20ma, and it dropped back to 0.14V in 12Hrs. So I removed the wrapping and it had nothing, just a blank cylinder. So it really doesn't give us any new info.
 
Back
Top