Best way to reconfigure old battery pack or is this a bad idea?

[vanturion]

Hey ES-

I'd like to get your opinion on some potential battery surgery.

So I spot welded a split pack of 20S8P (10S + 10S) back in 2014 using Samsung 25Rs. Despite not having the accuracy of the modern Kwelder and only using an old 12V car battery, the welds were good and the packs continue to function well. The problem is I used the convenient "H" nickel strip roll that is only 7mm x .15 thick across the series connections, and when I eventually get a Nucular 12F controller, I'd like to turn up the juice a bit. The ampacity just isn't there with my battery construction sadly. I've been at a Max 48 battery amps since I built the pack, but I'd like to take it to 75-80A.

Battery Pack Health
Things are still pretty good despite it's age. I'd say I'm still under 200 charge cycles all things considered. I've never once charged 100%, charged to 4.1V only a handful of times, and regularly charge to 4.05 to 4.07V. I've also never stored them while charged, and always use the pack/bike right after charging. Low-end cutoff has always been at 3.3V. Overall pack resistance has increased some, but, like I said, things are still pretty decent range-wise. All in all, I'd like to get more life out of these batteries and more performance after the controller upgrade.

Depending on the pack ampacity remedy, I also wanted to throw another possibility I've been thinking about. I'd like to get your thoughts on adding another parallel group of 25Rs to bring each pack up to 11S. I realize the pack performance would be constrained by the older parallel group capacities, but it would be nice to bump top speed and the overall torque curve up too.


Here's a pic of what I'm working with prior to removal of painters tape for pack heat shrink sleeve (hot spots). I don't have a good pic of the H-strip exposed.

Potential Solutions

• Tin copper bars and use high-wattage soldering iron to "weld" over existing nickel strip series connects.
  • Question: would this really alleviate the ampacity problem as the current still needs to travel through the spot-welded nickel strip first?

• Completely disassemble pack, rip off "H" strip from all cells, carefully dremel spot welds flat-ish, and use modern spot-weld technique with sandwiched .15mm thick copper strips + nickel/nickel-plated steel possibly for 11S configuration.
  • Question: is this even viable?

Any other options besides constructing a new pack with new cells? Advice?

 

[john61ct]

construct a new pack with **top quality** new cells as your core primary, gives you all the zoom zoom performance you need all on its own.

Say that's at 20S.

Now, "atomize" your old packs down to single cells.

Get an accurate cap test rig going, CC load, precisely timed, ideally export data for analysis on your PC.

Put the good ones into matched groups of 20, and use those strings to **extend the range** by wiring in parallel with your primary pack.

Carefully, matching voltages (in this case within a volt or two).

Do NOT allow unmatching cells or packs to be connected **in series**, just murdering the better ones.

 

[goatman]

i soldered copper wire last year, no problems anymore

https://endless-sphere.com/forums/viewtopic.php?f=6&t=104027&hilit=marin#p1526264

and ive put over 6000kms on them since then

originally i bought 2 of those packs used 2 years ago as 14s5p, then made them 2- 15s4p, 63v blew a controller i had so went to 14s4p then went to 17s4p

 

[vanturion]

construct a new pack with **top quality** new cells as your core primary, gives you all the zoom zoom performance you need all on its own.

Say that's at 20S.

Now, "atomize" your old packs down to single cells.

Get an accurate cap test rig going, CC load, precisely timed, ideally export data for analysis on your PC.

Put the good ones into matched groups of 20, and use those strings to **extend the range** by wiring in parallel with your primary pack.

Carefully, matching voltages (in this case within a volt or two).

Do NOT allow unmatching cells or packs to be connected **in series**, just murdering the better ones.

I mean, yes, ideally I'd buy a new cells, and test and match cell groups for ideal construction like you said, if I was in the market for a new pack. Definitely trying to not do that since the 25Rs still has life in them yet (and untapped current capability)! That said, I'm not sure all that is necessary with new cells from a good supplier presumably from the same manufacturing batch. Is that standard practice these days to do all that?

The most I did was measure the resting voltage when they arrived before paralleling just to confirm before forming the parallel groups. Groups generally run within .01V of each other even after almost 6 years now.

As for adding a series group, I still use an icharger3010b to balance when charging. What I was thinking is if I were to add a new 1 additional series group, I'd make one of the balance leads swappable with the new group to keep the old a new cell groups relatively in line since the charger only does 10S. Just thinking about this, not decided.

 

[vanturion]

i soldered copper wire last year, no problems anymore

https://endless-sphere.com/forums/viewtopic.php?f=6&t=104027&hilit=marin#p1526264

and ive put over 6000kms on them since then

originally i bought 2 of those packs used 2 years ago as 14s5p, then made them 2- 15s4p, 63v blew a controller i had so went to 14s4p then went to 17s4p

Nice goatman, that's good to know that's been viable and working for you, thanks! Since this is an older pack, I'm not as considered about a little degradation due to externally applied heat too so it would be an acceptable risk IMO. Did you use anything special regarding soldering tools?

I've got a sht soldering iron, so if I end up going this route I'll need to upgrade the iron anyways. Thinking something high wattage and fat tip to minimize contact time.

 

[goatman]

yes , i used a 60 watt iron with a chisel tip that was about 8mm wide, you need heat, 40 watts wont do it.

copper strand wire solders quick but solid copper acts like a heat sink

pre tin the nickelstrip and pre tin the wire before soldering to the pack so your in and out and done the joint in 3 seconds

if you dont get a good solder joint on of your solder joints in 3 or 4 seconds just leave it to cool and move onto the next one and come back and fix it once it cools

 

[john61ct]

6 years with non-LFP cells is a long time

These packs are a consumable item, if you can't afford a new one now, start saving now, just a question of how soon.

My suggestion ensures usability while you're futzing around

you will still be eking out every last drop if value out of your likely past EoL cells

but by testing and matching them per cell tossing the bad ones

getting say 100 more cycles out of the reconfigured pack(s) not 20 or 40

If you think not worth the trouble then try and see

but it is not the case that cells wear evenly over six years