Identify bad cells in pack.

Planning ahead, if you have a 100cell pack which the BMS is seeing as one unit how do identify failing or dead cells, in a timely manner.

dragonfly said:

Planning ahead, if you have a 100cell pack which the BMS is seeing as one unit how do identify failing or dead cells, in a timely manner.

Click to expand...

The expectation seems to be wait till it fails short and hopefully blows its fuse, which you'll hopefully notice ..... or over time you'll hopefully notice one pack has less capacity than others, by which time there'll likely be a good few failing or dead cells to replace.

Having a voltage graph of each pack definitely helps identifying the pack(s) havingtrouble. Assuming all packs started out with reasonably similar capacities.

Example#1, if a pack has difficulty reaching the same top voltage as other packs, despite the balancing function working overtime, then that pack has one or more cells with very high self discharge.




Example#2, if top balancing was working fine, but discharging towards ca 3.2Vonepack's voltage drops much faster than the others, then that pack as somehow lost capacity.





As to your actual question: Using my graphs,I've so far detected 8 bad cells. On 7 of them there were visual cues that something was wrong: "rust" from leaking electrolyte and/or abroken fuse.
In just one case, I could not see anything wrong. I plugged it back in and let it runfor a few days until the weekend, by which time "rust" had formed on the positive end of one cell.

When doing such maintenance, I also add some capacity to the weakest couple of packs by replacing one cell each (eg. 1700mAh -> 2300mAh). So it is conceivable that I have dead cells in my pack that have failed very gradually over time. I suppose I'll have to retest all cells eventually... perhaps in a couple of years time.

As said of above. Wait and identify a pack. Then either be lucky with failed fuse or retest all cells.

Best is to build n+1 system where you always have a spare pack you can swap out a failing with. Then you can retest without have to stress.

So I guess it's a case of general pack performance and eye balls. So I would guess that makes fuse wire better than glass fuses, i.e. easier to spot the blown fuse.

Is there anyway to test individual cells without dismantling the pack?

Voltage check for testing blown fuse. Or flit gun

Hmm, I fancy packing 1 wire temperature sensors into the cell spaces. wonder if I can fins a simple way of keeping them wired safely away from the main electrics.