How it is possible to find a dead cell without disassembling?

That is a pretty bold claim. One method would be a sophisticated BMS that tracks each cells performance. This was done with earlier model Tesla's if I remember correctly (as I recall mentioning that a single cell could be deactivated to keep from further failure; might also be a different manufacturer).
The other option would be that they put a load on the packs and use a very sensitive Non Contact Current Probe. It might not get them the "exact" cell, but might get them within the ballpark of it though. For instance, it might detect a decline in current flow in a group of cells. Then narrow further by removing only that part of the cover.

Other than that, I have no idea. Would be interesting to know how they actually do it, if it's legit.
 
Thinking out load....

How much would one declining cell affect the overall Tesla battery? I presume it would show as a balance issue? but I would think the overall system could handle one cell that was self-discharging. If it was shorting, the cell level fusing should take care of it. How bad can one self-discharging cell be?.

I'm thinking of the 11,000 cells in my powerwall in the 112 packs of 1s100p. I'm pretty sure I've had one cell with the fuse wire not soldered correctly or broken thru mis-handling etc. Surely I have some self-discharging cells here and there? But I don't see anything Batrium wise (balance wise) to indicate an issue.

I'm imagining that eventually Batrium will show me a pack that won't keep up with the rest. Then I'll take if offline and diagnose. First level would be to unsolder one side let it sit for a few weeks and measure the voltage of each cell. This should show any self-dischargers. If this doesn't reveal anything then I presume it would be a capacity decline? - maybe completely disassemble and retest all the cells?.
 
Last edited:
OffGridInTheCity said:
How much would one declining cell affect the overall Tesla battery?
Click to expand...
This isn't the issue, though. Remember when Tesla's were first being put on the road, they were using 18650 Lithium Ion cells. Certain types were known to explode fantastically. So to help keep a cell from going nuclear, you fully disconnect it when detecting an abnormal surge through it. So if all the cells around it were taking 2A, and this one cell was pulling 4A, and they are all connected in parallel, something must be wrong with that one cell. So disconnect it before it becomes an issue.

I think the first Teslas on the market didn't have that fuse sheet like a lot of people use now. They were basically over-engineering the detecting system to minimize any issue of cells going short circuit.

Now Tesla is moving away from Li-Ion and going with LiFePO4's now that the tech for them has gained performance in capacity and life cycles.
 
I've seen this before in an early video, they apparently use an extremely good multimeter to measure the voltage of each cell.
 
For a live pack while doing sustained charging, etc, a thermal camera might help see a hot or cold cell?
Like said above, watching the BMS for differences & unusual balancing should be a thing.
 
hbpowerwall said:
I've seen this before in an early video, they apparently use an extremely good multimeter to measure the voltage of each cell.
Click to expand...
I thought that the quality of contact while testing will give more difference and voltage of the cells while all cells connected in parallel, no?

Can you give the example of such multimeter? Wondering how much does it cost? :)
 
Bold claim, but not impossible.
A very precise thermal camera can show you which cells warm up the most (internal short) or which dont warm up at all (totally dead) when charging or using very precise voltage meters when charging and probing various points inside the battery assembly.

Both are expensive to buy but can possible save a lot of time large scale.
 
I wonder if it's possible to test with the amp function on the DMM, if it was accurate enough you might be able to see a different amp draw on a bad cell? * Just thinking aloud
 
hbpowerwall said:
I wonder if it's possible to test with the amp function on the DMM, if it was accurate enough you might be able to see a different amp draw on a bad cell? * Just thinking aloud
Click to expand...
This is likely, sort of. But you'd have to take the pack apart and intersect the connection.

It would be more likely that you'd use the 20 setting to read milliVolts and measure from the cell to the buss connection and look at the voltage drop. I'm guessing this would give you a different reading based on cell condition. But not really sure, though.
* Just thinking out loud :p
 
overthinking it, even more, you'd have to have a machine to connect the probes at exactly the same angle/pressure, aslo clean both electrode & cell perfectly clean with such small voltages any difference in how the probe was connected would give a different result..
 
You must log in or register to reply here.
Back
Top