Help with first pack build

[mynameisjoe]

Hi everyone and sorry if I ramble a little here. I'm a newbie to all of this and trying to research as much as I can, something I wish I had done BEFORE going ahead with this project! But I have money and time invested in this now so there is no turning back, only different paths to take.

I fly drones and want a field charger, but they're £500 for the one I want. On another forum I frequent I came accross a guy who made their own battery pack out of reclaimed 18650 cells, great idea as I have a mate in IT who has just done a large roll-out of new laptops in his company. So I have now collected 100's of cells from the old laptop batteries and over the last wee while have been sorting them out into groups. Anything with physical damage or below 2 volts were recycled. The cell type I have the most of are Sanyo R1122, around 300, so it made sense to use these first as they're all the same type, same chemistry, same charge/discharge etc. I have tested each for voltage and sorted into piles of 2.5-3v, 3-3.5v and 3.6-4.2v. Fortunately I have enough of the 3.6-4.2v cells to build the pack. So I have now charged a little over 100 of them, monitored for heat and further discarded any that were abnormal, and then checked their internal resistance. All are 4.12 to 4.18 volts so shouldn't cause any problems as there is less than 0.1v difference. IR is between 45 and 70mOhm. All I can find on these cells so far is IR should be below 100mOhm. My plan was to group them by IR, but now I read it is better to group by capacity. OK, I can do that as my charger/IR tester can also test for capacity. The problem is it can take 5 hours or so to test 4 cells, and I will not charge/discharge cells unatended or overnight so it is going to be a lengthy process. While I am thinking on this I have these cells stored in holders, and will check for self discharge in another 3 or 4 days. The plan is to build a 7S12P pack. I was going to go for 4S10P, but the voltage did't really quite work out (I want 12v) so decided on 7S with a 24-12 regulator. The 12P is purely because that is how many I can fit in the available space. Now the batteries for the drones I have are 3500 mAh 40.42 Wh (3S) and 2250 mAh 17.32 Wh (2S). Their respective car chargers will handle the charging so I don't need to worry about the different voltages, never more than two being charged at a time. Plus I'll be fitting some panel mount PD and QC outlets that can take the voltage down from 26v so I can charge things like phones and tablets if needed. There will also be an XT60 connector to take charge in from a 7S balance charger, and yes I will be fitting an appropriate BMS. Does all of this sound reasonable so far?
As for grouping the cells, do I REALLY need to do a capacity test on a minimum 84 cells for this type of build? I would have thought that IR is a decent indicator of battery health? I'm not fussed about getting the full (almost) 700watts out of it, I'd be happy with somewhere between 400 and 500.

Do any of you experienced members have input or can highlight any failings? Things I should definitely do/not do.
Thanks in advance for any input or thoughts

 

[mynameisjoe]

Nobody?

 

[floydr]

As for grouping the cells, do I REALLY need to do a capacity test on a minimum 84 cells for this type of build? I would have thought that IR is a decent indicator of battery health? I'm not fussed about getting the full (almost) 700watts out of it, I'd be happy with somewhere between 400 and 500.

Yes you need to capacity test all cells. Also was the IR tested with an ac 4 wire IR tester? if so depending on which meter it can be used to sort the cells before you capacity test them. see IR Meter Recommendations
Later floyd
ps if you need cheap but good testers take a look at these https://www.aliexpress.us/item/3256805258616061.html? shipping was fast the testers were reviewed on the same site as the IR meter Recommendations

 

[rafalg]

I can give some advice, but the value of it may be random, i'm not a very advanced builder
- if you can't / dont want testing, charge the cells and give them few weeks of rest, then measure voltages, find outliers with voltage dropping more than 0.1V (i dont know what is the right value here, but in my opinion it shouldn't be more than 0.1V after a month. Just do the first measurement after 1-2 h of rest after charging because there might be small initial drop after disconnecting the charger.
- dont build a big battery for a first time, learn on something smaller.
- are you soldering or spot welding? With spot welding the technique is more or less standard, you just have to make repeatable good welds, and with soldering you need some practice too to find materials and tools that work best.
- with used cells, consider a modular battery where all sections can be disconnected and replaced. It's quite common that after some time a cell develops some self-discharge and brings down the whole section, then it would be useful to be able to replace it without tearing apart the whole battery. I don't have a system for that, but in last build i did that by putting each section in a separate holder and connecting them with wago-type connectors. All enclosed in a box so the sections are fixed well and dont rattle around. Didnt have a need to replace anything yet so maybe it wasnt necessary, but feels that the thing is serviceable at least.
- when soldering, use copper wires or braid, don't solder nickel strips
- think about emergency rescue plan in case things go wrong during the build (for example, battery catches fire) - how to quickly put the burning stuff to some safe place
- 500 watts of power for 24V battery is like 20amps, this wouldn't be much for a 12P battery. Maybe better to make two smaller batteries, like 7S8P or 7S7P, then you'll have one still working after some failure, also easier to find an enclosure for that and carry it around (7S8P is like a brick, almost 3 kg)

 

[mynameisjoe]

Yes you need to capacity test all cells. Also was the IR tested with an ac 4 wire IR tester? if so depending on which meter it can be used to sort the cells before you capacity test them. see IR Meter Recommendations
Later floyd
ps if you need cheap but good testers take a look at these https://www.aliexpress.us/item/3256805258616061.html? shipping was fast the testers were reviewed on the same site as the IR meter Recommendations

Just a 2 wire tester that is on my charge/discharger. I will have a look at the tester you mentioned, thank you. And I tried to order one of those testers from aliexpress, but everytime I tried it blocked me and said there was a security concern, so I gave up :-(

 

[mynameisjoe]

I can give some advice, but the value of it may be random, i'm not a very advanced builder

Any advice is valuable and very welcome

- if you can't / dont want testing, charge the cells and give them few weeks of rest, then measure voltages, find outliers with voltage dropping more than 0.1V (i dont know what is the right value here, but in my opinion it shouldn't be more than 0.1V after a month. Just do the first measurement after 1-2 h of rest after charging because there might be small initial drop after disconnecting the charger.

They have been sitting for two weeks now, will do a test over the weekend to how they are getting on and then leave for another 2 weeks

- dont build a big battery for a first time, learn on something smaller.

I didn't think it was that big

- are you soldering or spot welding? With spot welding the technique is more or less standard, you just have to make repeatable good welds, and with soldering you need some practice too to find materials and tools that work best.

Spot welding, but I can solder too

- with used cells, consider a modular battery where all sections can be disconnected and replaced. It's quite common that after some time a cell develops some self-discharge and brings down the whole section, then it would be useful to be able to replace it without tearing apart the whole battery. I don't have a system for that, but in last build i did that by putting each section in a separate holder and connecting them with wago-type connectors. All enclosed in a box so the sections are fixed well and dont rattle around. Didnt have a need to replace anything yet so maybe it wasnt necessary, but feels that the thing is serviceable at least.

Very good idea, thanks

- when soldering, use copper wires or braid, don't solder nickel strips

Noted

- think about emergency rescue plan in case things go wrong during the build (for example, battery catches fire) - how to quickly put the burning stuff to some safe place

I am working on a bench right beside easy to open double patio doors for that very reason. Anything goes wrong and they're straight out of here and onto the poor flower bed

- 500 watts of power for 24V battery is like 20amps, this wouldn't be much for a 12P battery.

I'm using a regulator to step down to 12V, I don't need anywhere near the 20 amps. Most I'm likely to draw at one time is 6A, maybe a little more but certainly not over 10A

Maybe better to make two smaller batteries, like 7S8P or 7S7P, then you'll have one still working after some failure, also easier to find an enclosure for that and carry it around (7S8P is like a brick, almost 3 kg)

I am actually building two packs and they will be in paralel with eachother, this is just due to the way I am planning on fitting them inside the enclosure. It won't be getting carried far so the weight isn't much of an issue

Thanks for all the help

 

[mynameisjoe]

Just as I was reading and typing all the above, I realised that I may have to rethink the plastic enclosure I was planning on using!!! Something tells me that a vented metal enclosure would be a bit safer if things go wrong in use or when charging the pack up

 

[mynameisjoe]

Yes you need to capacity test all cells.

OK, I will heed your advice and listen to your knowledge. Afterall, that is why I am here.
BUT, can you help me understand why, please? I'm quite happy for things to take longer than I originaly anticipated if it's the correct thing to do. But for for my own learning I like to know why things are just the way they are. I get it from from a capacity point of view, good to know your actual capacity rather than guessing. But what are the other implications of not capacity testing? Will it effect the BMS I plan to use? Is it dangerous? If so then why? Again, I am not asking because I don't believe you, I am an engineer and like to understand the hows and whys, quite happy to learn and be better in everything I do. Especially if it results in not setting my house on fire!!!

 

[floydr]

But what are the other implications of not capacity testing? Will it effect the BMS I plan to use? Is it dangerous? If so then why? Again, I am not asking because I don't believe you, I am an engineer and like to understand the hows and whys, quite happy to learn and be better in everything I do. Especially if it results in not setting my house on fire!!!

Cells not having around the same capacity if one cell group reaches or exceeds the set voltage it will end the charge early. The same cell group will discharge faster than the rest ending your discharge early limiting range. The Cell group might reverse polarity if discharged to far. This shouldn't happen with a BMS. Testing also helps weed out heaters. Both discharge and charging would be a hazard if a cell group reverses polarity. You only have the capacity of the lowest cell group if you have 6 cell groups of 23Ah and 1 cell group of 19Ah you would have a19Ah 7s7p battery 492Wh, if you have 7 cell groups of 23Ah your battery would be 23Ah 7s7p 595.7Wh. Read wolf's posts starting with Wolf's PowIRwall and the ones I mentioned in post #3 of this thread. He explains much better than I can. Hope this helps
Later floyd

 

[docw009]

When you build ebike packs of 4P and lower, having one abnormally low capacity cell really kills the useable capacity. In addition, that low capacity cell could also have other issues, like a high self discharge rate that also kills the battery over time. For a 7S13P power wall pack with used laptop cells, it's probably less critical to test all 91 cells, especially if the IR is over 40 milliohms. They're not high current cells. It won't be a high performance pack.

WIth an otherwise unifrom group of similar cells of the same model and manufacturer, significant variations in capacity and IR are red flags.

I really hate capacity testing as I only have one 4-slot tester and it's slow, When I used recycled or almost new cells, I would test them. And my results, typically in the 1-2% range told me I had to keep testing or these cells would wind up in my battery, With new cells, I would only check voltage/IR and sample test a few, I figured that if I was paying $5/cell for Sanyo GA's, they better be good.

Now I realize that we hobbyists are not getting automotive grade cells even when we buy brand name. I bought a group of 28 new Lishen cells in January and one was down in voltage. Turned out to be way down in capacity too. The other 27 were perfect. Bummer.

 

[mynameisjoe]

Thanks all for your input, much appreciated. Especially @floydr for you explanation