Pack rebuild best practices

[drspeakman]

After adding another 14 packs, connected in parallel to my other 14 packs, I have effectively doubled my storage. Using Batrium BMS, and have all packs essentially the same mix of mAh used laptop cells, as best I could. Do not have an internal resistance tester, just Opus and Foxnovo testers. Over the last 2 years, I have removed and replaced underperforming packs 2-3 times, and each time it has been quite time consuming. Mostly ended up completely disassembling the pack, and retested all the cells.

After adding the additional 14 packs, 14s100p, I noticed the first pack was underperforming after the first 3 days, so I have replaced it with another pack that I had ready to go just for this purpose.

Question is...what is the best way to refurbish an underperforming pack? Is there any way other than completely disassembling, and retesting all the cells? Can you remove all the fuses and negative connections, then let it sit for a week or two, and test each cell individually for self discharging cells? Should I charge the pack up completely first?

Essentially what are the best practices and easiest and least time consuming methods of doing this?

 

[barry]

First buy a IR meter it's only 5060 dollars (vapacell YR1030) and it saves you a lot of trouble don't trust the IR readings from opus or any charger the ar way off.
what do you mean by underperforming? lower ah,balancing problems.

 

[OffGridInTheCity]

Don't know if this will help but I can share my own pack rebuild issues. I've had to rebuild 5 of 70 packs and fiddle with 10 others during my journey so far.

My batteries are 14s100-130p. There are5 of them (70 packs)so far. I've 2 basic issues....

Mismatched capacity -In 10of my 70 packs I've had to 'touch things up'by addingcells. Each pack is targeted to be 260-263ah (@3.7v). In 8 of the 10 cases, I had to add 3-5% capacity - e.g. just solder in some additionalcells. In 1 case I had to add 30cells (have no idea why) which is aprox 25% of original capacity. This seems to have worked and these packs have now been stable - no balance required - for 2years of daily charge/discharge. Note: I've gotten better at this by employing 1) matched or compatible, evenly distributed cells in each battery, 2) watching/matching IR, and 3) letting them sit for a month before soldering up - to make sure there are no self-discharging cells in the pack.

Bad IR -In another case I had 5 packs that just would not stay in sync with the others. They would charge up a 150mv higher and100mv lower than the rest - just not sustainable as this greatly reduces the max and min voltage range for the overall battery. It turned out that these 5 packs were the green Sony G5/G7 cells... and the only thing wrong that I could find after dissembledand restestwas that they all had 200+ mOhm IR (definitely abnormal)whereas the rest were in the 50-100mOhm range (more normal).

And to your overall point - battery maintenance is definitely something to consider for DIY powerwalls. On the one hand - good healthy packs are possible but even so, eventually, they will start to degrade and unless all the cells degrade smoothly 'together' its likely a fair amount of disassembly and reassembly will be needed thru the coming years.

 

[Wolf]

To build a pack that is going to perform to expectations with the least amount of trouble requires some forethought.
There are a couple of ways to test cells that will go into the packs.
The (what I call) brute force method which is charge and discharge each cell at 1A to get your mAh results. You will need an OPUS or SKYRC charger tester to be able to do that. Most all the others will charge at 1A but discharge is limited to 500mA.
Anything that gets hot or falls below 80% SOH gets discarded into the recycle bin. The cells that pass get sorted into bins by 100mAh differences and after that an equal amount are taken out of the bins and installed into packs.
It is a rather rudimentary way of doing it but it works so who am I to say any different.

The other (more delicate) way is to take each cell, spend maybe 10 seconds on it and measure the IR and V of the harvested cell to determine if it has any chance of being a good cell. As you learn each of the manufacturer's chemistries and IR readings for each part number you will quickly find good cells, mediocre cells, and bad cells. Here is a sheet that will guide you https://drive.google.com/open?id=169QnjLDBmFyfOY6RybH0VGBR4bV2NJDT
Yes and all IR measurements are done with a YR1030 or equivalent 4 wire 1kHz tester. See treadhttps://secondlifestorage.com/showthread.php?tid=7742&pid=53191#pid53191 I Personally use aRC3563 it is the easiest and the best I have found.
I personally made an Excel spreadsheet to analyze each cell. Yes this does take some time but the rewards are worth it.
So that after you have harvested your lot you can then sort by whatever criteria you want to get the best cells for your build.
The sheet with a cell database and all formulas already plugged in can be found here. https://drive.google.com/open?id=1W8CQa6CIGP6cYvkoTokpdCYUU1V-6Nsj

My personal sheet with lots of extras is here https://drive.google.com/open?id=1dvcgpoKa-9nBJUTvL9B1vcMSRtIQnzkm
In a nutshell my procedure for testing cells is.
1. Once the cell is liberated out of the pack it is checked for V and IR, anything below 2.5V orabove the recommended IR is set aside.
The cell is given a number (cell is labeled with a sharpie), initial V and IR are recorded in the spreadsheet.
2. Cells are tested with my OPUS, SKYRC, and Foxnovo. I know the Foxnovo only does 500mA discharge. But due to much experimentation and testing I find that if the IR is acceptable, the cell will do just fine whether C/D/C at 1A or 500mA.
3. Results are recorded on the cell via sharpie and again recorded in the spreadsheet which due to the formulas gives a SOH percentage.
4. Once the required number of cells have been collected and the data analyzed the build can begin.
5 I use Nemos Repackr which gives you very well balanced packs with the cell number that you previously marked on the cell with a sharpie.
All that is necessary is to copy and paste the cell number the mAh capacity and the IR into the sheet and give it the criteria for your pack build.
Nemo's Repackr can be found here. https://drive.google.com/open?id=1l8RNN64bn0wJ9KciT_6XJ75npcFRUl_f
6 Assemble all packs and charge to 4.2 V discharge with iCharger x8 regenerative discharge at 30A note the Ah of the battery and do the next one.
I have 1 more pack to test and so far all have been extremely close in Ah results.
You can view the results here. https://drive.google.com/open?id=1XuBHDyqCGIjQYYqDJ8WIscmYJ3EAdOAv
or here:



Disregard the 15th pack as I needed to raid it because as I was being overly cautious with cells that were close to the margin of acceptance or not.

So that is my procedure
There is an old saying that goes "There is never enough time to do it right the first time but always enough time to do it over"
I have also added the Excel sheetsas attachments
Wolf

 

[ajw22]

A few months ago I wrote a post on how best (AFAIK) to find a self discharging cell. Whole process takes just 30min or so.
https://secondlifestorage.com/showthread.php?tid=8759

As for capacity re-balancing: In each battery pack, I place several 1200mAh powertool cells in specific, easy to access locations. I replace those cells with higher capacity cells as required.

 

[Crimp Daddy]

Really good information in this thread...

I wanted to share this general observation from my experience testing "lower grade" or higher IR cells and using them. What I think might be happening in a pack after assembly has much to do with your actual load requirements.

Many people recommend to test a cell at 1 amp and just slap that number on the wrapper and take that as a final result and build a pack.

Then the pack ends up testing 20% high or lower in capacity than the individual cell results because the load requirements are totally different than your test.

I try to take into considering what I am trying to do with a pack even before it is build. If I have a low current application, I tend to test those cells at a lower current to get a more accurate capacity figure.

Testing that same cell at 1.5a or 500ma would yield a different capacity significant enough to note when compared to a 1a test. The higher the IR is, the more significant the impact.

 

[Wolf]

Question is...what is the best way to refurbish an underperforming pack?
Can you remove all the fuses and negative connections, then let it sit for a week or two, and test each cell individually for self discharging cells? Should I charge the pack up completely first?

The best way is to build a pack that is balanced internally to itself. In the hopes of not having to refurbish it to begin with.
This is by no means being critical here to your pack build as we are all learning how to avoid these issues and maybe find a solution.
In your case I would charge the pack to 4.2V disconnect either the pos or neg but not both and after a week of sitting check each cell for V.
If nothing shows up maybe another week. As some have suggested you can disconnect half the pack and check the disconnected cells one at a time and the rest of the half as a whole. Whichever side has a V drop is your problem spot.If you still have no V drop on some cells then it is definitely an IR issue.

.............. It turned out that these 5 packs were the green Sony G5/G7 cells... and the only thing wrong that I could find after dissembledand restestwas that they all had 200+ mOhm IR (definitely abnormal)whereas the rest were in the 50-100mOhm range (more normal).

Very true the Sony US18650GR G* series of batteries dohave a very high IR yet perform just fine as far as their mAh results are concerned.
I have not advised to use them in a powerwall build unless they are theonly cells in the whole pack and string. Seehttps://secondlifestorage.com/showthread.php?tid=5132&pid=48096#pid48096

As for capacity re-balancing: In each battery pack, I place several 1200mAh powertool cells in specific, easy to access locations. I replace those cells with higher capacity cells as required.

Interesting concept and in your case easily done because of your open pack design. It is very clever and well thought out.

................ The higher the IR is, the more significant the impact............

There you hit the nail on the head. But there is another factor and that is capacity.
We all like to crack open battery casesand finding NCR18650, or NCR18650A Panasonic cells (both rated at 2900mAh) or better yet the NCR18650B (3200mAh monster)and hoping that they test well.
The problem is that if you use these cells along with 2200mAh say aICR18650-22* Samsungcell there is an inherent interaction between these cells in parallel that does not bode well within the pack. There is much crosstalk electrically speaking between these cells. I know this has little to do with refurbishing a bad pack but it will help with building a pack that is balanced. My suggestion is that all cells be within a mAh range of ~400mAh (as an example 2250mAh to 2650mAh or 1850mAh to 2250mAh)......Yea all my NCR 18650 cells are out of the picture....and an IR range of ~40m? to ~55m? a 15m? spread at max. Optimally all cells should be at least 80% SOH. This does not mean you cannot use the NCR18650s just build a pack at their tested mAh and put batteries in parallel with them that fit the 400mAh window also considering the IR and following the 15m? window.

Best
Wolf

 

[Crimp Daddy]

Most certainly capacity matters as well, as the higher capacity cell is going to end up "picking up the slack" for the lesser capacity cells. Meaning the higher capacity cells will end with a higher discharge rate. In this scenario, the higher the discharge rate, the worse the problem becomes. High IR only compounds the issue as load increases.

I have built packs with my B and C grade cells that have a 500-900mAh delta , but its for extremely low drain uses, like a small battery for some LED strip light that draws <0.80 amps across a 15 foot section.

The higher the load, the more IR and capacity matches matters. In an ideal world, we would want to build with matching cells, as they generally do with new batteries.

As much as I like to get into the details and the technical side of batteries, for practical applications and to maintain my own sanity, I limit the scope in how I use old recycled 18650 cells.

I don't really make any high performance projects with recycled 18650 cells. Most all are low drain or over-sized packs for smaller projects. For large capacity builds, like a powerwall, I always default to large format cells like those found in EV cars.

I find these projects fun, but I always want to maintain realistic expectations about service life and required maintenance. If you expect that you can "set it and forget it", you should not be building with what is essentially considered recyclable trash.

 

[Wolf]

As much as I like to get into the details and the technical side of batteries, for practical applications and to maintain my own sanity, I limit the scope in how I use old recycled 18650 cells.

........... If you expect that you can "set it and forget it", you should not be building with what is essentially considered recyclable trash.

Ha Ha "sanity" I have lost that a long time ago when I first joined this forum.
That being said I have learned a lot!
As far as recycled trash well that is somewhat debatable but mostly accurate. Considering I have tested over 6000 cells and recorded them, after filtering by my criteria I came up with just over 1200 cells that were usable and some of them on the very edge. So yea I dug through the trash. But even a blind squirrel finds a nut every now and then.
My Inverters will draw a maximum of 50A and that is with everything on in the house highly unlikely.The load on the pack would be ~ 625 mA per cell quite reasonable I think.
The set and forget option is as you say EV or brand new. The rest there will always be some tinkering but if we can limit it................

That's all we can ask for.

Wolf

 

[Crimp Daddy]

First off, I would like to appreciate your efforts. I know what you did was hard work, time consuming, and it has been a large contribution to the community.

I'm not trying to make an argument for "my way" of doing it... i'm just look at it for what it is. Getting laptop packs from a recycling center is the battery equivalent of dumpster diving. Those packs are usually there for a reason, but sometimes you get lucky. Sometimes you find trash, sometimes you find gold, sometimes you find an oscilloscope for $20 (as I did in my own search).

I'm here doing the same thing, I do process packs and get 18650 cells, but I have also had an opportunity to get some EV cells at a great price.

In my journey I have come across many different types of batteries: laptop, medical, modem, prismatic, EV... Laptop packs have by far been the worst of the worst batteries. I have very low expectations from them, and often times find its not even worth my time/efforts to process and test. Medical packs on the other hand generally yield wonderful cells.

It takes a considerable amount of time and money to package cells into a battery, which is why I started to enjoy EV cells. In addition to that, you can generally find a large quantity that all came from the same vehical, meaning everything is matched, drastically reducing the issues we are trying to address in this thread.

Its just something to consider because I have seen many people bite off more than they can chew trying to go big without the knowledge and background to address these problems up front as you have through your process. Its generally more advanced than many builders are capable of doing.

 

[OffGridInTheCity]

First off, I would like to appreciate your efforts. I know what you did was hard work, time consuming, and it has been a large contribution to the community.

I'm not trying to make an argument for "my way" of doing it... i'm just look at it for what it is. Getting laptop packs from a recycling center is the battery equivalent of dumpster diving. Those packs are usually there for a reason, but sometimes you get lucky. Sometimes you find trash, sometimes you find gold, sometimes you find an oscilloscope for $20 (as I did in my own search).

I'm here doing the same thing, I do process packs and get 18650 cells, but I have also had an opportunity to get some EV cells at a great price.

In my journey I have come across many different types of batteries: laptop, medical, modem, prismatic, EV... Laptop packs have by far been the worst of the worst batteries. I have very low expectations from them, and often times find its not even worth my time/efforts to process and test. Medical packs on the other hand generally yield wonderful cells.

It takes a considerable amount of time and money to package cells into a battery, which is why I started to enjoy EV cells. In addition to that, you can generally find a large quantity that all came from the same vehical, meaning everything is matched, drastically reducing the issues we are trying to address in this thread.

Its just something to consider because I have seen many people bite off more than they can chew trying to go big without the knowledge and background to address these problems up front as you have through your process. Its generally more advanced than many builders are capable of doing.

To +1 on top of this...

As many of us have, I have employed several different '18650 cell purchase' strategies over the last 3 years- but lately (last 5000 cells) I've settled on buying pretested DIY cells from Battery Hookupand Power2Spare as the prices have continued to drop and the service has been fantastic and these sources have remained steady/reliable over several years now. I find that I can get large quantities (with some patience) so my later packs arematched, the work is much less, the testing is much less volatile, andthe undesirable rate is very low. The fact that the pretested cells are all pretty stable/similar in capacity and manufacturelead me to believe that my packs will have less maintenance than a mixed collection offree/raw usedlaptop cells.

I don't mean to come across asnegative against raw cell scavenging - especially if you get them free or almost free. I just don'thave any sources for free cells and new (genuine) cells are expensive or impossible find truth in advertising - so I went the middleground of paying some$ toget high quality used cells. It's what I would recommend if you go 18650.

If I could do it all over again... and I could find trusted sources over the years.... I would go EV (Nissa Leaf or Tesla or Chevy etc)as the shear volume of work to process 1,000(s)and 1,000(s) of18650shas worn me down.But for me, its too late as I'm nearly done with my battery bank

 

[Wolf]

I'm not trying to make an argument for "my way" of doing it... i'm just look at it for what it is. Getting laptop packs from a recycling center is the battery equivalent of dumpster diving.

Its just something to consider because I have seen many people bite off more than they can chew trying to go big without the knowledge and background to address these problems up front as you have through your process. Its generally more advanced than many builders are capable of doing.

Totally agree! You will not get an argument out of me!
I have somewhat refined my "diving" though and as you say medical packs are usually very good and I scout those out.
The recycler I deal with is very accommodating and will usually set those aside for me.
I will still go through the laptop pile though and seek out the OEM packs that meet my fancy. Once you get rather selective and "picky"
with those you can have a much better yield rate than with just pulling any pack out. But that comes with experience and can not be formulated into a youtube video.
I also agree that "most" will bite off more than they can chew esp. with all the videos out there of how "easy it is".
No its not. I started collecting batteries around August 2018 and am just now finishing my packs. It is a long time commitment and also an ongoing one. The rewards on the other hand are great. I have learneda lot of things. Excel spreadsheet formulas,writing arduino sketches, setting up a linux box with influx, grafana, telegraf mosquito, node red etc. (some of it is still a mysteryto me)come to mind, stuff I have never done before.
Additionally I have joined a great community to learn from and share with and also made some very good friends

The fact that the pretested cells are all pretty stable/similar in capacity and manufacturelead me to believe that my packs will have less maintenance than a mixed collection offree/raw usedlaptop cells.
If I could do it all over again... and I could find trusted sources over the years.... I would go EV (Nissa Leaf or Tesla or Chevy etc)as the shear volume of work to process 1,000(s)and 1,000(s) of18650shas worn me down.But for me, its too late as I'm nearly done with my battery bank

Another agree!

When we all started, 18650 was the way to go, they were relatively easy to get in some circumstances (took a little leg work) but all in all it wasn't that hard. Now with EV batteries coming to the recycle mainstream I totally agree it is far easier to set these up compared to breaking apart packs and going through the testing process building packs with bus bars fuses and such. The redeeming value of doing this though is that we learned a lot and we have a solid base to build upon.

Wolf

 

[not2bme]

I like how hbpowerwall does his spot by using a thermal camera for the heaters/self dischargers.

Agree with having 1 or a few source to start with. My oldest pack is over 3 years old now and they came from the hoverboard batteries. The rest of them I was lucky enough to get from a single medical pack source. I'm up to over 20kwh now and have another 10kwh to process. No rebuilds at all so far. I have a crate of laptop batteries I will not touch except for a rainy day!

So here's how I approach it, if it makes any sense:

1. Group together like cells. So if you have a 14s80p, make sure you don't have too much variety, such as keep each 1s80p with the same mixture, 20%lgs/80%samsung or 30%lg/30%panasonic/40%samsung. Don't mix 10 different brands or models (different capacities) if you can afford to avoid it.

2. Group together like capacities. It's better to have all 14s80p with mostly 2600mah batteries, then another 14s80p with mostly 2000mah batteries. Don't mix low capacity with higher capacity as that will just make one cell work harder next to the other cell, causing premature wear.

3. Be aware of rated capacities vs tested capacities. Having a 3000mah battery that tested 2000mah doesn't mean it's equivalent to a new 2000mah battery. In fact it's already less than 70% of its rated capacity so it means that the battery has gone through hundreds of cycles and may not have the lifespan left, while a new 2000mah just started it's journey. Putting the two together, you might find after 500cycles later that the 3000mah battery suddenly dies. I was very careful about picking my batteries and I only picked those that had 90% of rated capacity. My 2600mah batteries had tested 2300mah or better to make it into the pack.

 

[Crimp Daddy]

I am glad you mentioned the thermal camera because it bring up another relevant topic.

Tools, or lack of them, can significantly change the quality of the process as well. I see such a wide variance in testing capabilities depending on the person and budget.

Some people have a goal of making it for or with as little money as possible, and the other end of the spectrum where the tools are often times the star of the show. Spot welders, thermal cameras, high end chargers, quality amp meters and multimeters, not to mention all the custom tools and electronics that have been developed in the pursuit of this hobby.

I still see some people that dont meet what I would consider the minimum baseline for tooling and they are deep into a Powerwall project. Those are usually the same people that are more likely to have issues as well.

I just think its another factor for consideration.

 

[camthecam]

The last time I pulled my packs apart which I have good at one end and shit at the other so I know which ones to start verbally abusing first, I noticed that a lot of my solder joints were quite crap and the whole thing was not going so well. BUILD well. (I am going Leaf)