Recycled 18650 testing - new method - more info.

[Skrzypas]

Hi,

My first post. I'm from Warsaw, Poland and this summer I started to prepare e-bike for the next season.
Well, I've started making a ~40 Ah (counting the rated capacity of gathered cells) battery from recycled laptop 18650 cells (3,7 and 3,6 V nom).

I would like to share my method for testing these (and probably suitable also for other) cells. Yet I didn't start to discharge cells so any opinions and comments are welcome.

The goal is to make a pack from recycled cells that will be reliable and repairable.

Up to now:
1. I've bought a few cheap 18650 chargers for 1 cell (works for pair pretty well)
2. I have access to an Electronic load on a daily basis, so I can make programmable discharge of cells

New method is proposed since the pack must be reliable as much as possible, and plenty of time is still left for the machine launch:
1. Charging cells to 4.20V-4,25V
2. Waiting, occasionally writing down voltages of all up-to-now charged cells
=> result - discharge graphs, V/month self discharge
3. Cells with excessive self discharge are put on another pile.

Well, I see that most of my original thought was actually not posted:

4. Equalizing cells at 4.200V (note 3 decimal places) +/- 5 mV
5. Discharge in a following manner (with 1s data acquisition):

• a) OCV (open circuit voltage) 10s (0A)
  b) CCL (constant current load) 12 min (1A) -> 1A is my system design point for ca. 2000 mAh rated cells
  c) OCV 5 min (0A) -> 5 minutes time will be adjusted, since I did not check how fast the OCV rebuilds in laptop cells
  d) CCL 12 min (1A)
  e) OCV 5 min (0A)
  f) and so on, up to voltage = 3 V@1A -> not to stress the cells too much

After such discharging a lot of information can be determined:

• - Capacity@1A [mAh]
  - OCV in function of drawn mAh (discrete - one point each 12 minutes = 200 mAh) [V] - for SoC determining
  - V@1A in function of drawn mAh (discrete, 200 mAh) [V] - for determining real battery operation time in the design point
  - Resistance in function of drawn mAh (discrete, each 200 mAh, R=dV/1A) [Ohm]

When I test high enough number of cells, I will arrange cells in a following way: each 20P sub-pack will get cells of
a) Similar resistance and capacity
b) if possible: Similar shape of discharge curve at 1A

And so:

• - Small variations in cell resistance between cells (in each 20P sub-pack) will allow equal current passing through each cell in the pack in each DoD.
  - Similar capacity will allow draining cells equally to 3.0V (BMS cut off).
  - Shape of discharge curve gathers all above information in a continuous (not discrete) way.

Any suggestions/corrections are welcome.

 

[e-beach]

Hello Skrzypas!

I know you said that you are from Warsaw but please follow these directions.



Ok, IMO the self discharge test is a waist of time. Simply charge the cells to their maximum voltage and then discharge them through a watt meter to see what capacity the cell has. Group them by discharge watt capacity. Cells with a high watt capacity will pass a self discharge test while cells with a low watt capacity will fail a self discharge test because those cell is worn-out.

Edit: Do not over discharge the cells.

If your used cells hold 4.2v you are lucky. Most used laptop cells will not fully charge so you will have to go thorough a lot of cells to get the amount of cells you need to build a high capacity battery pack.

Good luck with it and I am looking forward to your independent testing results.

 

[DrkAngel]

See - Bulk Capacity Testing

 

[amberwolf]

Most used laptop cells will not fully charge

Yes because their BMS stopping them, the cells have no brain to do that.

It is more likely (since this discussion is about testing bare cells, not while in a laptop battery pack with it's BMS) that E-beach's statment meant that cells simply do not reach full capacity (they don't accept as many Ah as they did when new, which is perfectly normal for aging cells). Sometimes aging cells also will not reach or hold a normal final voltage, either.

 

[999zip999]

I like to put a set of sense wires for checking voltage with a cellog, this is with a bms or not. One low or two in a string will weaken a series to the lowest ah. Just one or two weak cells.

 

[DrkAngel]

Ok, IMO the self discharge test is a waist of time.

e-beach has been on this rant before!
It seems he is trying to sabotage everyone's pack building efforts! ... ?
After much hemming and hawing ... turns out ... e-beach had never, ever! , built any battery pack using 18650 cells! - No experience!!!
He did try to test some cells, but didn't even finish.

Self-discharge testing of each cell is essential ...
unless you don't mind your pack constantly unbalancing itself.

Eliminate self discharging cells,
then build banks of equal capacity and IR
and you will have a stable, reliable battery pack ... with no "need" for any BMS!
(Do ... monitor pack and bank voltages though)

See - Bulk Capacity Testing

Failure to remove self-discharging cells will require the use of a BMS.
And!
Tragically, the BMS will be forced to be continuously discharging all banks not contaminated with self discharging cells, (and also banks with cells of lesser self discharge rates.)

Not recommended!
With diligent testing, you could "balance" all banks of cells by placing identical self discharging cells in each bank ...
However, self discharge, typically, becomes unpredictably worse, so pack would need to have a BMS and need to be re-built-balanced and constantly monitored and tested.

Much better-simpler to:
bulk charge all cells to identical voltage,
let set for several days,
eliminate all cells with noticeable self-discharge. 8)

Want to compare?
Not recommended! - UNSAFE!!!
Self-discharge indicates some degree of internal short.
Which might get progressively worse at any time.
Possibly resulting in heat >> smoke >> fire? ... !!!
Build a small pack using the self-discharging cells ... for comparison purposes only, in a fire safe location, with safety gear - not for use!

 

[Skrzypas]

Most used laptop cells will not fully charge

Yes because their BMS stopping them, the cells have no brain to do that.

It is more likely (since this discussion is about testing bare cells, not while in a laptop battery pack with it's BMS) that E-beach's statment meant that cells simply do not reach full capacity (they don't accept as many Ah as they did when new, which is perfectly normal for aging cells). Sometimes aging cells also will not reach or hold a normal final voltage, either.

They charge, but some do not keep voltage and/or deliver current. I already charged more than 180 separate cells from old laptop bateries (with cheap single cell chargers). Except of cells that were 0V and did not wake up (even if paired with good cells), I didn't have a single cell that could not be charged to >4.2V!

 

[DrkAngel]

Most used laptop cells will not fully charge

Yes because their BMS stopping them, the cells have no brain to do that.

It is more likely (since this discussion is about testing bare cells, not while in a laptop battery pack with it's BMS) that E-beach's statment meant that cells simply do not reach full capacity (they don't accept as many Ah as they did when new, which is perfectly normal for aging cells). Sometimes aging cells also will not reach or hold a normal final voltage, either.

They charge, but some do not keep voltage and/or deliver current. I already charged more than 180 separate cells from old laptop bateries (with cheap single cell chargers). Except of cells that were 0V and did not wake up (even if paired with good cells), I didn't have a single cell that could not be charged to >4.2V!

Don't waste so much time and effort trying to figure out what e-beach means ... last I investigated, he had no - never - zero - zip - nada 18650 pack building experience!!!

Ok, IMO the self discharge test is a waist of time.

e-beach has been on this rant before!
It seems he is trying to sabotage everyone's pack building efforts! ... ?
After much hemming and hawing ... turns out ... e-beach had never, ever! , built any battery pack using 18650 cells!
He did try to test some cells, but after 4 months, still hadn't finished.

It is essential to test cells for self discharge!!!

I built my 1st recycled 18650 packs back in 2008.
Yes, due to not taking self discharge into account, durability suffered.
I was a noob and there were no known guides, instructions or predecessors in recycling 18650 cells into battery packs.
So, hopefully, I might be forgiven for failing to test for self discharge ... ?

Subsequent builds, after eliminating self discharging cells and building banks of equal capacity have been durable and reliable ... without even needing any BMS!
See - Homemade Battery Packs (very large thread = use the index or spend hours browsing)
(My later posts typically reflect improved methods, procedures etc., but all my posts were retained as a historical - logical progression of experience and ideas.)

 

[e-beach]

And leave it to DrkAngle to try and highjack someone else's thread. DrkAngle sure screams and rants like a little girl. :lol:
In regards to DrkAngle, he can rant on his own thread all he wants, but, he should leave his rants off of Skrzypas thread. I will not engage with the little-grey-kitty fellow despite his rants, half-truths and outright lies. ES is about information, not confrontation.

Enough said about that.

@ Skrzypas, do your own testing and you will find, as I have, that there is no reason to waist time on bleed down test because a used and useful cell with it's near new capacity, will hold it's voltages very will. Healthy used cells don't bleed down in an appreciable manner. They hold their voltages much like new cells. Keep copious data on each cell and over time you too will abandon the bleed down test because it is not an accurate test and only waist's time.

Just charge your used cells, with a smart charger, to their maximum charge voltage. Then discharge them through a watt meter and record their capacity. Sort your cells by discharge capacity and you will know the ones you will want to use.

Keep a complete data-base on each cell so you have actual proof of your testing, and not just opinion like the ranting little grey-kitty guy who does not keep any data.

As for the cells you acquired that were zero voltages, yes they are dead. You will probably find that the cells with 3.65v or higher, when you can obtain them are the ones you will want to use. They are the ones that are probably still healthy enough to be highly useful.

I am looking forward to your own independent conclusions to your own testings and building!

 

[DrkAngel]

... Just charge your used cells, with a smart charger, to their maximum charge voltage. Then discharge them through a watt meter and record their capacity. Sort your cells by discharge capacity and you will know the ones you will want to use. ...

Then, ... when you get frustrated because your pack is constantly out of balance ...
Dissemble your pack, test for and remove self discharging cells.

Self discharge is largely unrelated to capacity!
Cells that heat up during charging have severe self discharge.
But there are varying degrees of self discharge.
Most always becoming progressively worse >> more dangerous.

I say get rid of them.

e-beach, did you ever finish testing your batch of cells and build your 1st 18650 recycled battery pack?
............... (started testing in November 2013) .................

 

[fechter]

Guys.... chill with the personal attacks here! I'm starting to get tired of it.


Back on topic...

If you have a large quantity of cells to test, I'd recommend charging to 4.2 then measure the discharge capacity as a first round. Anything that won't charge or deliver a decent capacity gets tossed into the recycle pile right off. Self-discharge could be tested with the ones that pass the first round. Self-discharge testing is very time consuming.

I had a large pile of Panasonic cells that were all near zero volts from being discharged slowly by the attached BMS. Anything below about 3v needs to be charged at a very low rate (like C/20) until it gets above 3v, then it can be charged at a normal rate.

Nearly all my cells recovered nicely and provided over 85% of their rated capacity, not bad considering they were well used before going into storage. I don't recall finding any that had significant self-discharge even after sitting for a month.

 

[dnmun]

you are deliberately lying about what ebeach said originally where you also responded by insulting him instead of proving your argument.

you do not have the capacity to make an argument that is supported by facts so instead you resort to insulting him as though that is of any value except to prove you are a bully.

the argument you have hidden in this trash you posted up was that the cells that dropped in voltage off of the charger were defective and did not have the stated storage capacity.

when i observed that your statement was incorrect then you started insulting me like you did to ebeach.

i have pointed out before on your thread that it is cell storage capacity that is the most important thing to learn about the cell but you never do capacity tests and you never use a BMS because you think that just by using only part of the stated battery capacity you will be safe from over discharge.

there is nothing to prove that either. in short you just run a self serving campaign promoting poor practices. anyone building a battery without a BMS is not constructing a reliable power source which can protect itself from damage either from over discharge or overcharge or from short circuit protection on the output.

 

[DrkAngel]

In my experience ...
I charge all prospective cells to ~4.20V. (4.17V for latest batch)
If they heat up or won't attain this voltage, I trash them. (quantity varies by batch)
Then I let them set for several days.
Many will bleed down a 1/100th V ... no problem.
But, some will self discharge several 100ths of a volt = big problem.
(Possibly 10% rejected, varies by batch)

These are the cells I insist must be eliminated.
They will constantly unbalance any pack.
Even if charged-balanced constantly, the BMS will be constantly charging the banks with self discharging cells then discharging the remaining banks.
No BMS? Banks with self discharging cells will continually drop to unusable.

I prefer to build packs with no self discharging cells.
Simply build banks of equal capacity.
No BMS.
(But I do monitor bank and pack voltages constantly cell level meter-alarm <$5)
No need to rebalance.
100s of trouble free cycles
Years of trouble free use.

As evidenced by 4 year 7000++ miles recycled laptop Lipo pack - no BMS, bulk charged
(Ran balance charge ~annually to meter capacity)
Patched 1 cell after ribbon cable damage)

As evidenced by re-re-recycled 1 year old 3000+ miles
cells were best capacity that I harvested from previous builds - Bulk charge only - no BMS
Possibly an average of 8000+ miles on each cell.
Presently at <50% of original rated capacity.
All banks still balanced within a couple 100ths V after 100++ bulk charges

Why build,
take apart,
retest,
rebuild.
Repeat as necessary ...
When you can do it once,
properly!

 

[DrkAngel]

Self-discharge testing is very time consuming.

I just gang together a large batches of cells
1s40p
Charge to 4.20V per cell (Modified MeanWell S-150-5 : iMax B6 or B8 would work, but slower)

or
3s24p
Balance charge to 4.17V per cell (iMax B6 or B8 etc)

Then separate and let set for several days
check retained voltage

I can process hundreds of cell per day.
Yeah, takes several days but only time "consumed" is several minutes of soldering - desoldering.

 

[fechter]

Either way you probably want to measure both capacity and self discharge to test if a cell is good. Which one you do first should be the one easier/faster for your test equipment.

If you want to run packs without a BMS, well, that's up to you. With careful pack level monitoring and conservative voltage cutoff points, you can get away with it until a cell group fails. I've done it with some of my packs.

If you're building packs that are going to be used by the general public, they really need safeguards to prevent damage. Why do you think ALL commercial electric car and laptop batteries have one? There are just too many ways you can mess up and damage an expensive battery or even worse, start a fire. No matter how well matched, eventually a cell will go bad and create a dangerous condition.

 

[DrkAngel]

Either way you probably want to measure both capacity and self discharge to test if a cell is good. Which one you do first should be the one easier/faster for your test equipment.

e-beach, rather strongly, insists that others should not test for-remove self discharging cells.
I recognize self discharging cells as possibly dangerous and at minimum, the cause of problematic, constantly unbalanced packs.
In my view, he is welcome to his opinion ...
But he is not welcome to actively advocate to others that they sabotage their battery builds by using self discharging cells!
(e-beach has tried testing some cells but there is no record, or claim, of his ever building any pack (18650's)

3s24p
Balance charge to 4.17V per cell (iMax B6 or B8 etc)

Then separate and let set for several days
check retained voltage

I can process hundreds of cell per day.
Yeah, takes several days but only time "consumed" is several minutes of soldering - desoldering.

Then I flip every other 3s4p
and
rewire as 18s4p and do a metered capacity discharge

75V DC through 2 x 100w light bulbs provides ~1A discharge, but Ah meter keeps good measure.
I used 100w light bulb and 150w mini heater for a 1.35A discharge.
I added external power to Ah meter so I could mark capacity and remove cells as they dipped to target low voltage limit.

Took about 7 hours to accurately rate 72 cells. (Next 72 tomorrow.)
Could have discharged faster ...
But I have to sit "working" LOL, 8 hours at a stretch and need something to do.
Computer repair is 95% watching anti-virus programs run.

 

[DrkAngel]

Red and Green 18650 cells, pictured in previous post ...
Used 75% to build 1kWh+ actual capacity battery (33.3V 31.2Ah) build for a battery cost of ~$90!
15mph = 60 miles (motor only)
20mph = 40 miles (motor only)
25mph = ~27 miles (motor only)

And LiFePO4 battery of same capacity would have to be twice as large and heavy ... and likely 10x the cost (~$900 ?)!

With my conservative charged voltage, used capacity will be slightly less than 1 kWh, but should last many years and many, many thousands of miles.

 

[liveforphysics]

Hello Skrzypas!

I know you said that you are from Warsaw but please follow these directions.



Ok, IMO the self discharge test is a waist of time. Simply charge the cells to their maximum voltage and then discharge them through a watt meter to see what capacity the cell has. Group them by discharge watt capacity. Cells with a high watt capacity will pass a self discharge test while cells with a low watt capacity will fail a self discharge test because those cell is worn-out.

Edit: Do not over discharge the cells.

If your used cells hold 4.2v you are lucky. Most used laptop cells will not fully charge so you will have to go thorough a lot of cells to get the amount of cells you need to build a high capacity battery pack.

Good luck with it and I am looking forward to your independent testing results.

Funny, the self-discharge test is the single one I would pick if only able to do one type of screening.

I would personally not build packs from used cells with unknown history, but if I were, your process for screening looks as good as any, if your time is worth less to you than buying new cells.

 

[John in CR]

Self discharge test makes for a decent keep or set aside for flashlights and stuff test, but there's no reason not to do at least a crude method to match based on capacity. I have Konion cell packs (from Makita tool packs that used Sony US18650V and VT cells) that have been in service since 2009 (infrequent use now, but daily for 2 years) that still maintain good balance, and all I did was a simple capacity test to match them in series strings.

First, I always keep them as close to the factory alignment as possible. Tab welds have been proven with use, and if the cells are balanced that means they are perfectly matched. I charge same size blocks in parallel and get all the cells exactly the same voltage. I then put all of the blocks in one long series string...be very careful with the high voltage...I've had strings of up to 800-900V for the discharge test. I use a series string of 100W lightbulbs as my resistors, 1 per 100V, set up at least 1m away due to the heat. I measure the current and discharge for 1/2 to 2/3 of the nominal capacity.

Then I disconnect the string and measure the voltage of each block. Blocks with the highest cell voltages get rated A, and then assign capacity grades for each. Those with same ratings get put together in series strings for my pack(s), and low rated cells are set aside. While I don't know the actual capacity, the relative capacity is good info, and a single discharge test only takes about 2 hours. The most time is spent getting the cell blocks charged to identical cell voltages.

Even at just an amp the high voltage DC will arc across a sizable gap, so be careful.

 

[Punx0r]

I've had strings of up to 800-900V for the discharge test.

 

[dnmun]

this is the misinformation that dank angle has repeated about what ebeach said:

"e-beach, rather strongly, insists that others should not test for-remove self discharging cells.
I recognize self discharging cells as possibly dangerous and at minimum, the cause of problematic, constantly unbalanced packs.
In my view, he is welcome to his opinion ...
But he is not welcome to actively advocate to others that they sabotage their battery builds by using self discharging cells!
(e-beach has tried testing some cells but there is no record, or claim, of his ever building any pack (18650's)"

ebeach did not say that. dank angle said that in replying to him and distorted what he had said while insulting him.

i repeated my observation that capacity of the lifepo4 cells i had evaluated was not a direct corollary to the self discharge rate and i pointed out that the capacity of the cell was the most important variable for the lifepo4 cells i had tested.

self discharge in lipo is distinctive and is easily observed to be able to eliminate it from the search because a lipo cell with self discharge will drain to 0V and is already removed from the group.

i think it is usually obvious which cans could be used and which are dead so his insults to ebeach were uncalled for then as well as now.

 

[DrkAngel]

........... Yo! - dum'un ...........

I agree with you on these statements, but I still think a bleed down test is a waist of time until one has identified the potential cells for usage.

The bleed down test is a useless time waster and should be abandoned.

Don't waist you time with a "bleed down test".

Sadly, e-beach seems willing to put obviously defective cells into his builds?

Con respecto, I haven't missed that point at all. And from my study it is also not 100% accurate as I have found a small number of cells that self discharge and still have a very useful capacity. I use my bike on a daily basis and their-fore charge my bike on a daily basis so one option for me could include cells like that,

But that is no excuse to advocate the same "madness" to others!!!

Ok, IMO the self discharge test is a waist of time.

@ Skrzypas, do your own testing and you will find, as I have, that there is no reason to waist time on bleed down test ...

Once more ... e-beach has never built any 18650 battery pack ... no battery building experience! ... he has tested some cells ... but never put any in service after his "testing". - (as of last query)

 

[e-beach]

With used cells a bleed down test is a waist of time. Especially a 30 day on the shelf test that was so popular for such a long time. A capacity test will tell you if the cells are good or not. It will also let you group your cell in a configuration that works well for you. With used cells I prefer to group my cells in parallel strings from high capacity to low so over time I can rotate the low capacity cells out of the pack more easily.


There is a direct correlation between a cell with high capacity and it's voltage stability. A used cell with high capacity will not bleed down significantly because the cells chemistry is in good shape.

This is why new cells don't bleed down significantly.

As for the several 18650 several pack I have built, I have not published any info on them because I am not done with what I am doing with them. As I have stated in the past, it could take a while. When I am satisfied with my work with them I may publish or not depending on my time constraints and mood.

As for what I have done or not done in regards to ebikes or battery packs or any other project, DrkAngel has never been to my garage/shop and has no idea of what I have done or not done.

 

[DrkAngel]

There is a direct correlation between a cell with high capacity and it's voltage stability. A used cell with high capacity will not bleed down significantly because the cells chemistry is in good shape.

Not true!
Plenty of low capacity cells have excellent voltage stability.
and
You even reported that some cells with self discharge still had good capacity! ... How sad is it that you don't even listen to yourself?

Sadly, e-beach seems willing to put obviously defective cells into his builds?

Con respecto, I haven't missed that point at all. And from my study it is also not 100% accurate as I have found a small number of cells that self discharge and still have a very useful capacity. I use my bike on a daily basis and their-fore charge my bike on a daily basis so one option for me could include cells like that,

This is why new cells don't bleed down significantly.

You must only deal with top of the line new cells?
Dependent on manufacturer and quality control level, most all cell productions output a percentage of self-discharging (defective) cells, capable of full rated capacity.

As for the several 18650 several pack I have built, I have not published any info on them because I am not done with what I am doing with them. As I have stated in the past, it could take a while. When I am satisfied with my work with them I may publish or not depending on my time constraints and mood.

Last you reported, you were 4 months into preliminary tests on your 1st batch of recycled cells. With no pack ever produced! ... ?

 

[DrkAngel]

There is a direct correlation between a cell with high capacity and it's voltage stability. A used cell with high capacity will not bleed down significantly because the cells chemistry is in good shape.

Maybe that is your problem, ... you do not understand cell construction?
You have been told again and again and again that capacity and self discharge are 2 separate, largely unrelated, factors.
Excessive self discharge is not typically a result of cell chemistry!
Excessive self discharge is due to defect, contamination, damage, corrosion, abrasion etc. in the separator ... responsible for or creating various degrees of electrical short circuit!

Any type of short circuit is dangerous, especially with Lithium cells, and liable to get progressively worse at any moment and rate!

This is why I continually insist that self discharging cells be eliminated!!!

Advocating that self charging cells not be removed, or even tested for! ... is sabotaging battery reliability and safety.

When you mention and advocate(?) using known self discharging cells in your battery build ... it might be bordering on criminal? ... !!!

Sadly, e-beach seems willing to put obviously defective cells into his builds?

Con respecto, I haven't missed that point at all. And from my study it is also not 100% accurate as I have found a small number of cells that self discharge and still have a very useful capacity. I use my bike on a daily basis and their-fore charge my bike on a daily basis so one option for me could include cells like that,