Thoughts on this battery harvesting scenario?

So, I was the lucky purchaser of a lot of 4s1p Panasonic 25R battery packs that should have received little use. I would like to use these to build a large ebike pack - 14 or 16s and 6-8P. From what I have read, creating parallel groups first and then putting them in series is the way to go. Unfortunately, that requires disassembling all of the packs. Is this the only way?

If I were to break apart all of the packs and test the cells individually, I can cut the nickel strips so that all of the cells retain a bit of spot welded strip.

--Am I better off trying to use the small amount of connected strip for soldering or welding, or am I better off taking all of the strip off and spot welding a new strip down once the pack is laid out?

I just purchased the maletctrics spot welder and have a high power soldering iron, so I have all of the tools.


In my head, the ideal, yet labor intensive, workflow is to:

-Separate all of the individual cells, removing the previously welded strip.
-Test each one for capacity and IR.
-Group the cells in 6-8P groups, standardizing capacity across each group, then spot weld.
-Connect the groups in Series with a spot welder.


Is this the way to go?

Do not separate cells!
Test the 4s1p as individual cells.
Charge as 12s or 16s(?), clip or tack together (I use Neodymium magnets wrapped with adhesive copper foil), make sure all cells at same start voltage and monitor as charging.
1st - Charge all cells to full, precise 4.200V(?) or label each cell with exact voltage 1 hour after charged.
2nd - Let set a full week, then check voltage for self discharge. Separate and discard any with notable or abnormal voltage loss.
3rd - confirm all cells at identical voltage. connect in series, 28s or 32s. Discharge at 1 Amp, (2 x 60w light bulbs @ 120>>100VDC = 1 Ah per hour) for 2 hours.
Check cell voltage at 10min(?) for voltage sag (bad IR) mark any poor-bad IR cells for removal.
Monitor cells and can stop discharge at 1 hour and note individual cell voltage as preliminary capacity measure.
2500mAh "good" cells should tolerate another full hour of 1 Amp discharge, monitor and make sure no cell falls below 3.0V.
Can stop discharge early, possibly note voltages at 1.5 hours?
2 hour discharge, label cells with precise resultant voltage several minutes after discharge, will give a good comparative capacity between all cells for battery building guide.

Thanks for the input DrkAngel.

The packs already have a BMS installed on each to keep them balanced during charging. Would this change your strategy?


If I am reading this correctly, you say to keep them connected for testing. What about when it's time to build the pack? Should I split them up then?

E.S. Thank you for DrAngle.

If I were to break apart all of the packs and test the cells individually, I can cut the nickel strips so that all of the cells retain a bit of spot welded strip.
--Am I better off trying to use the small amount of connected strip for soldering or welding, or am I better off taking all of the strip off and spot welding a new strip down once the pack is laid out?

Click to expand...

When you separate the cells. Just cut the strips and leave sections attached to the cells.
Whilst removing all the strips may be neater, you run the risk of tearing tiny. Puncture holes in the can base where the strips are welded. ( top is not an issue, as its not part of the sealed can).

transposon said:

Thanks for the input DrkAngel.

The packs already have a BMS installed on each to keep them balanced during charging. Would this change your strategy?


If I am reading this correctly, you say to keep them connected for testing. What about when it's time to build the pack? Should I split them up then?

Click to expand...

Preliminary charge with BMS intact but remove BMS for cell testing. BMS will "hide" defective cells!
(I often refer to a BMS as a Band-Aid for a defective pack, and build - use - monitor manually ... But can not recommend doing without for the average user)
Only good reason to "split them up" would be to remove defective cells.

After completing build, (before installing BMS), I would recommend running some carefully monitored cycles, to confirm successful - solid build.

Good point on the BMSs. Thanks.

I have a hobby charger that can measure total IR for a pack as well as individual cell IR within a pack. I think I will use that to check all of the resistances while the packs are intact.

I don't understand how I can build the pack correctly without separating the packs. Isn't the only good way to build a pack to make the parallel groups first, and then link the parallel groups in series? If my batteries come in a 4S1P configuration, how would I make the parallel groups for my build?

After, if, determining all cells of similar condition, doesn't matter if built in parallel or serial 1st.
Personally, I usually lay out in series, then parallel together.

If BMS put extra stress on #1 cells and you want to use them, stagger them to equalize banks.
12s
1st series 4-4-4
2nd 1-4-4-3
3rd 2-4-4-2
4th 3-4-4-1
5th 4-4-4
etc.
Keeping batteries largely intact ...

I am having a hard time visualizing how the layout would work. It seems like you would need a separate BMS for each series string to make sure they are balanced.

+1 on do not separate the existing 4s strings. I like slightly used batts better than new, because they're worlds cheaper and the connections and good matching are already proven. Check them all for good balance first, and set aside and to use those that are perfectly balanced first. Then charge them up in parallel, leaving them connected to each other for at least a day to ensure they are all equal in voltage. Next connect them in a single series string to run a discharge test on all of them at once using incandescent light bulbs unless you have an appropriate resistor. Be extremely careful due to the quite lethal voltage and have your resistors well away from the string to avoid heating the packs.

Run the discharge through an ammeter and stop the discharge at 60-70% DOD of the nominal capacity, and disconnect your series connections. Check them again for balance to weed out packs with weak cells. Rate all of the packs based on voltage (those at highest voltage have the most capacity) and use packs of the same rating to make your series strings, and then make your parallel cell level connections to attach to the BMS.

I processed literally thousands of Sony cells from Makita tool packs in 4s2p blocks using this method back in 2008 and 2009, and never used a BMS due to the self balancing nature of the VTC1's so I skipped the cell level paralleling. A few of those packs are still in almost daily use, though with quite diminished capacity. Not bad for cells originally rated with a 500 cycle life.

I ran discharge strings at such high voltage that I had to put as many as 8 or 10 100W bulbs in series, and it gets scary. Be sure that nothing with a big voltage differential gets anywhere close to each other for safety reasons. High DC voltage will arc across what seems like a significant distance which you'll see when you make the final connection to start the discharge. Again BE CAREFUL and be very cautious and deliberate with that high voltage string.

transposon said:

I am having a hard time visualizing how the layout would work. It seems like you would need a separate BMS for each series string to make sure they are balanced.

Click to expand...

6 x 7s2p bricks paralleled together with minimal leads for parallel-balance. Except the end "rails" for power output, with heavy tinned copper braid.

Thin black wires to balance connector or BMS. With properly (IMO) constructed batteries , parallel connections are subjected to minimal to 0 current, except for the end rails - battery output connections.

Ah, I think I understand now. Thank you.

Is there any guideline on whether or not the nickel strips used in the already assembled packs are adequate for a specific current?

I plan on pulling ~30A peak / ~15A continuous and am not positive the welded strip is up to the task.

im running 17s4p of 25r at 25/30 amp is 4 volt sag? continuous. and 50 amp max burst is 7 volt sag. I opened my pack up and its only 1 layer of nickel strip. nickelstrip is 7amp max(4strips=28amps). so now im going to solder 2 additional series copper jumper wires between 1and2 p then between 3and4 p from 1s to 2s etc til all my series connections are done. im hoping that get rids of the sag.

a quick example, went for a 20km ride today. average speed was 37km/h, the last half km to my house is a medium uphill so I went about 50km/h and by the time I hit the top of the hill the ca was showing 51.3v under throttle that was 3 hours ago and now my battery is at 59v

I am thinking of making a 14s6p 25r pack. My controller is a 26A KT controller. I am assuming 30A max on starts and steep hills.

If my math is correct, 30A / 6p = 5A per series string. I took out my calipers and measured the nickel strip. 6x0.15mm. Should I be concerned about the current capacity of this strip?

After opening up a few packs, for the series connections, some of the packs use 6x0.15 strip while others use 8x0.15 strip.

Would this make it necessary to pull off and reweld all of the strip so that they all match in cross section/current capability?

transposon said:

After opening up a few packs, for the series connections, some of the packs use 6x0.15 strip while others use 8x0.15 strip.

Would this make it necessary to pull off and reweld all of the strip so that they all match in cross section/current capability?

Click to expand...

Typically, cell "strips" are standardized, 6x might be 0.20?
Keep same "strip" type in series! and use heavier parallel wires?
If nearing cell output capability, might want to add "2" strip, but removing "6" strip - not recommended.

DrkAngel said:

Typically, cell "strips" are standardized, 6x might be 0.20?
Keep same "strip" type in series! and use heavier parallel wires?
If nearing cell output capability, might want to add "2" strip, but removing "6" strip - not recommended.

Click to expand...

According to my calipers, they are all 0.15mm in thickness.

Why would thicker parallel wires help? Shouldn't the current be pretty low across parallel cells either way?