What are some practical techniques for sorting cells in bulk quickly?

[Redpacket]

You might like to look at some testing flow charts in this thread:

18650 Harvesting Flow Charts

As a community we all have our different ways of processing cells and I wanted to take the opportunity by starting this thread to find some common ground on how we all should harvest cells. This is not meant to be a "be all end all" but more of a guide for new members and maybe some senior...

secondlifestorage.com

Looking for cells to weed out early (before charge/discharge testing) is likely to save a lot of time.
Like Wolf says above, Eg testing for low voltage & poor IR first.

 

[Overmind]

ok, that is so obvious! BUT grabbing a different color all the time would be a pain.. but still a neat idea

Red means 1k, green 500, blue 100. There's no need for higher accuracy than 100 since that's values under are usually tester's tolerances.

 

[jdeadman]

100% Agree Wolf. I have batteries that I have tested good Let sit a Month and are still good. After sitting for 6 months some are showing some self discharge. I realize that is kinda going overboard but all my built packs using the above method are still 97% the same capacity 2 years later. and All in balance. I balance my packs once a year in the summer but they are always monitored

 

[Wolf]

Red means 1k, green 500, blue 100. There's no need for higher accuracy than 100 since that's values under are usually tester's tolerances.

So it would be like a resistor band identifying the cells capacity.
In essence a cell with a 2561 mAh result would look like this?

Wolf

 

[OffGridInTheCity]

100% Agree Wolf. I have batteries that I have tested good Let sit a Month and are still good. After sitting for 6 months some are showing some self discharge. I realize that is kinda going overboard but all my built packs using the above method are still 97% the same capacity 2 years later. and All in balance. I balance my packs once a year in the summer but they are always monitored

>the above method are still 97% the same capacity 2 years later.
@daromer reports X % loss each year from just sitting around. I'm curious - do you store them at 3.6/3.7v? What's your experience with regard to capacity loss from sitting?

 

[daromer]

Papers state around 3-10% loss per year in total capacity just sitting. Note that it depends on cell type.
I have 6 years new-old stock cells. They had Lost around 5-8% in total over this time. Bur they were Stored basically empty and had a max shelf life of 3 year from that manufacturer.

So it all depends.

I dont think its an issue leaving Them for some time but based on papers they loose alot more when fully charged in terms of life compared to 4.0 or even 3.6 or below. That is a fact. The amount varies

 

[jdeadman]

the Capacity tests are my active packs capacity. For these packs in service they are held to 4.05V to 3.6V per cell but for the most part their cycle depth is really shallow.

The cells tested were sitting for 6 months held a set voltage of 3.8V set by a digital voltage PS that I have connected to a large parallel cell holders to make sure the cells don't sag or recover to a different voltage. once stable they get shevled My voltage drop limits are within the 5-8% capacity

 

[Overmind]

@Wolf yes, but the dots I put on the bottom and are very small.
Example (pic is from the net). Dots can be distributed not to look ugly also.

 

[anton_voltx]

Everyone else gave good advice and decent procedures. If you plan on processing 3k cells often, there are also automated machines for cell sorting by voltage/capacity that we use.

 

[OffGridInTheCity]

Everyone else gave good advice and decent procedures. If you plan on processing 3k cells often, there are also automated machines for cell sorting by voltage/capacity that we use.

I just finished processing 2500 cells over the last 7 months - so this sounds fantastic. What is the machine / how does it work? Does it do a charge/discharge test or read lables to get capacity? Does it report IR? How much?

 

[anton_voltx]

I just finished processing 2500 cells over the last 7 months - so this sounds fantastic. What is the machine / how does it work? Does it do a charge/discharge test or read lables to get capacity? Does it report IR? How much?

For speed it's usually sorted by voltage/resistance. For example we have used a machine similar to this:

Automatic Cylindrical Battery Sorter Sorting Machine For Lithium-ion Battery Pack Assembly Production Line - Buy Automatic Cylindrical Battery Sorter,Automatic Cylindrical Battery Sorting Machine,Lithium-ion Battery Pack Assembly Production Line Prod

Automatic Cylindrical Battery Sorter Sorting Machine For Lithium-ion Battery Pack Assembly Production Line - Buy Automatic Cylindrical Battery Sorter,Automatic Cylindrical Battery Sorting Machine,Lithium-ion Battery Pack Assembly Production Line Product on Alibaba.com

www.alibaba.com

But there are more mechanical / simpler solutions that I think can get closer to $2000. So I'm not sure 2500 cell in 7 month justifies industrial equipment but I suppose it depends how much you value your time

For capacity testing we use big racks that will measure the cells while they charge/discharge. You can get these anywhere from small 16 channel to big 256+ channel cabinets. There is no automatic sorting though, just run-of-the-mill capacity testing (but that makes sense when you consider the time needed for capacity tests and how small a proportion the actual sorting process is comparatively).

 

[deswong]

Expanding on this topic, I am curious to see what others are doing for storage of these after sorting them.
I've followed Wolf's idea of tagging the cells, and putting them into a spreadsheet.

I started to put all the cells of the same type into a box, so a box of INR18650-15M, INR18650-15Q, LGDBHE21865, etc, but feel I will have heaps of boxes, and then when picking out the cells from the spreadsheet I will have to hunt through boxes to find the cell number to extract.

I see others are sorting them in boxes by the capacity - this seems to be a bit easier to sort through, and means that I would have various different cells in the box, but around the same capacity making it easier to locate.

I'm trying to work out what would be easier, as if I do this by capacity, I would then have boxes of 1000mAh, 1100mAh -> 3000mAh, and that then would be a lot of boxes to work through, esp when going back to do a voltage drop test after a few weeks.

I'm currently at 300 cells in my spreadsheet, and didn't want to progress too far and have to re-sort boxes. I have at least another 500 cells to go that I have sorted into boxes of brand and type.

 

[floydR]

I store all my cells in cell holders with a piece of construction paper between layers. I would be putting the cells into a spreadsheet, but I am rather disorganized. i just write on the cell the mAh and the IR.

later floyd

 

[daromer]

100mAh bins. Nothing else and it need to be easy and quick. If i spend more time than that its cheaper to just buy new cells.

 

[Wolf]

100mAh bins

I completely agree with @daromer if you want to build a bunch of packs and complete a battery in record time. Your choice, and it seems to work for some. I for one don't consider building a battery a race. Some do. If you do then buy new and skip the whole process. Matter of fact just buy a ready made battery, why go through all this pain and suffering, buying equipment, wasting time. My first battery build took almost a year. We call it "secondlifestorage" for a reason.
Now I would say it takes me about 4 months from initial test to completing a fully balanced and functional 14s80p battery.
I consider it a labor of love and a hobby that has financial benefits with a lower power bill and the satisfaction of knowing that if I need to be independent from the grid I can be. I also consider that I have a very small, infinitesimal really, contribution to the environment we all live in.

As far as time spent on a cell I'm not in a rush. I take as much time as is necessary. I want to be absolutely sure that the cells I am committing to my powerwall are the best. I am after all recovering and re-using cells that would have otherwise gone to the recycler and would have been scrap. Trying to do my very small part of reusing rather than buying new.

I sort my cells by the number sticker. I have 2 megacell chargers and have a batch of 32 at a time which I put into cell holders. Once I have 160 cells I put them into a box with the highest number cells at the bottom, The boxes just happen to hold 16 cells in a row, when the excel sheet tells me that it's been ≥25 days since the test. I will pull the cells out in numerical order and enter the SD voltage test and IR numbers into the spreadsheet. Returning the cells in numerical order again, so that I can find them after sorting the best of the best from the spreadsheet and then of course the final choice by the excel repackr.

I know I spend a lot of time on each cell and some may think it's foolish and that's ok, everyone is entitled to their opinion. I don't care what others think I just want to know the quality of EVERY cell that goes into my system. I also want my system to be trouble free for as long as possible. I believe in the old adage... "The quality goes in before the name goes on".... You choose.
Wolf

 

[OffGridInTheCity]

If I may add to @Wolf's comment. I'm over 3yrs in now and understand things much better. I (pretty much) followed the @daromer method except.....
1) I gravitated to making sure all the 14s packs in a 'battery' have the same cell type (from same batch of used cells) and the widest difference was 15%. For 2500mah cells - that's 2500, 2400, 2300, 2200 - evenly distributed among the packs.
2) About 10% of my 84, 260ah packs had to be tweaked by adding 3 - 9 cells (avg 6% ah) after they had been in operation for a few months and I could see them sagging a bit.

This gave me a functional system of 84 healthy packs that will go 5 months with no balance - and even then its a drift of a few packs by 30mv. Some packs are 100 cycles (newer) and some are 958 cycles as of today.

However... I cannot (in practical terms) balance the 84 packs any closer than 40mv - 70mv (min / max) thru a 60% DOD cycle. After 5 months this can verge on 90mv. The charge/discharge cycles of some packs simply don't 'lock-step' with the others. This means...
1) I had to tweak 10% of the packs after putting them online and watching them in Batrium
2) Operationally I need to allow for a 90mv max difference - which means that 4.11v is the max I can charge to be safe.

You might say my process is somewhere inbetween @daromer and @Wolf's. I'm not as patient as @Wolf but patience has it's rewards... from what I can see - @Wolf's batteries won't suffer from the need for tweaks (that I had to do) or the wider variance in min/max mv ranges that I have. On the other hand, @daromer's less detailed advice will get you off the dime and lead you to an operational powerwall which is more fun than never getting there.

 

[daromer]

Its also important to understand that the smaller the DOD is the less tweaks are needed and also the larger packs the less tweaks are needed.
I have still not done any tweaks to my packs due to this. Yes they drift when drained out but its not a problem since i dont want to be there anyways

 

[Wolf]

from what I can see - @Wolf's batteries won't suffer from the need for tweaks (that I had to do) or the wider variance in min/max mv ranges that I have

This is a trace of the last 24 hour charge / discharge /charge cycle of my 2 batteries. As you can see by the purple single trace which is the voltage difference between min and max of each pack. My packs are all 80p and all of you know that it is rather difficult to tweak them. I also don't expect to have to. The min is ≈0.02 and the max is ≈0.04. So yea patience and meticulous cell selection has its rewards.
Wolf

 

[Doc3G]

Yes sir - its classic DIY. Takes significant effort but you can create batteries for a lot less. I've been thru - ebike packs (very difficult, sore hands), modem packs (not so bad), medical packs (had to use a table saw - horrible), and most recently the RING batteries (not bad - just crack open in the vice). After each battery created I say "I'm done!". Then 4 months go by and I see a great deal.....

You mention cracking the ring batteries open in a vice, can you expound upon this a bit more? How are you placing it in the vice in order to get it to break open cleanly?

 

[OffGridInTheCity]

You mention cracking the ring batteries open in a vice, can you expound upon this a bit more? How are you placing it in the vice in order to get it to break open cleanly?

Put the humped side with crack 1/8" below the edge of the vice and 1/2" to 1" into the vice and squeeze the vice pretty hard (in some cases)

The end will 'crack' and you can then pry it apart down the seams pretty easily. Sometimes they squish but won't crack - then try the other end. Experiment a bit and you'll find the 'squishy' spot that forces it to crack open. This worked on 99% of 1000+ packs processed.

In a few cases - its just didn't work and I used a hack saw to saw (along the line where it would crack) deep enough to get a screwdriver in (as in the 1st picture) to pry it open. The USB end has a circuit board and a good end to 'saw' thru the plastic as you'll hit the circuit board rather than the cells

In a few cases - always with the green NCR18650A - a small dent appeared near the bottom wall of a cell. Not sure if it was the vice (e.g. this never happened on blue cells) OR NCR18650As have 'paper-thin' metal + vice OR they were that way in the pack. Per my personal policy of anything <2mm I used them.... but that's not the general recommendation of the board.
Here's an example: