Thank You so much Jaime. Looks like I'm battery shopping then. The intermittent explains a lot and the direction i'm moving next.

 

Since you've been into it pretty deep, it's very important to establish that you torqued all terminal nuts to 48 in-lb.

I'm not too worried about the outliers. It's likely a sensing error, reporting error, logging error, etc. I'm going to focus on the sensible areas of the log visually.

SoC and 20 voltages plotted. Voltage on left vertical axis; SoC % on right vertical axis. Time is the X-axis. Current was omitted as I don't believe it has a lot of value or accuracy in this case.

In the various regions of consistent SoC drop, one thing happens very regularly... the voltages diverge from a relatively tight bunch to a very wide spread. They do this at different voltage levels and different SoC levels and usually after a period of at least brief SoC rise. The charging "pumps" the cell voltages higher, resulting in fairly consistent values, but as load is applied, they diverge steadily and significantly.

Conclusion - battery is completely shot and must be replaced. 90% of the modules are likely at 10-20% of rated capacity and can't be reconditioned to a meaningful degree. I'm eager for Jaime to complete the detailed testing of his 40 modules. It will illustrate just how hammered this packs are when they finally fail and how module replacement or other half-measures are completely off the table.

 

Thank you S. Keith, Yes all terminals , busses and connections were thoroughly cleaned and reinstalled before they could start to oxidize, and all were properly torqued. I checked for bad wires also, but only at the bus, not further into the harness. If i pick up used cells like Jamie suggested, what should a voltage drop test look like @4.5 amp draw? or is this test useless?
I only took readings at 60 and 120 seconds looking for weak cells.
Thanks, Andy

 

There is no answer to that question. Too many input variables, and 4.5A is too low a load to give truly meaningful results. The more accurate question is how much voltage difference should there be? The answer is "VERY SMALL." On healthy packs/modules, at higher currents, it's rare that you have modules outside of a 0.10V wide range until you start to get to < 20-30% SoC where they start to diverge according to their variation in capacity. At 4.5A, their variation should be less than 0.05V, but the issue is that wildly different capacities could show consistent results at the 1 and 2 minute mark where very little of their capacity has been consumed - so you learn little if anything.

You should likely do EXACTLY what Jaime did - buy 2X similar mileage Prius C packs from a mild climate, measure voltages to confirm consistency and just install them. I recently did the same thing for a Lexus GS450h 40 module pack. I tested them very thoroughly, but they were all perfect.

 

Tipping my hat to you fellas!!!
Thank you for schooling me

 

Thanks for all the great information on this thread! I find it very valuable in that I can run a self assessment on my Yukon. I don't have much faith in the local dealerships providing me with accurate information. I went out to collect some data today as I suspect my battery in a 2009 Yukon might be aging. I will attach the data log file. I merged two separate drives into one file so that is why you are seeing a time gap. I had a pretty difficult time finding a place it wanted to perform an auto stop. Even then, the truck did not stay in battery only mode for very long. Not sure if the environment played a part in that, the temp was 95 degrees F. Let me know what your thoughts are on the data, I welcome everyone's feedback.

Also, has anyone considered disassembling the SUV battery and replacing each individual cell instead of searching for a Prius battery pack(s)? If my battery is due for replacement, this was the route I was considering.

Thanks!

 

That's exactly what I did to the Tahoe referenced above. I bought 40 refurbished balanced cells from Electron Automotive, and rebuilt the pack. It's been running perfect for the last 3 months.

 

Results are indicative of widespread deterioration. The attached spreadsheet has two charts representing the two captures labeled charge and discharge, which correspond to the predominant operation recorded.

SoC is state of charge
Current is current (I suspect highly inaccurate)
dV is difference between highest block and lowest block - should typically be less than 0.3 in most cases.
V01-V20 are the block voltages.

The discharge chart has two areas circled showing substantial spread in the block voltages and a spike in dV. If you only had one or two voltages as outliers from a tighter group, then one may conclude a handful of bad modules. Given that you have a large spread from a single UPPER outlier, this is likely indicating that you only have ONE block that is marginally acceptable, and the others have deteriorated substantially to 10-20% of their rated capacity. The top two lines in that region are V01 and V20 - the coolest blocks in the pack.

Your proposed solution was recently documented here. Someone purchased 40 modules (not cells, there are 240 cells in the battery) and installed them.

EDIT: And there he is

 

Attempt 2 is attached.

The first attempt didn't turn out well. This attempt I didn't make it quite to 60% but hopefully it has useful data.

Autostop didn't last very long so I waited for it to cycle again and repeated the process until the engine came back on.

Let me know if you'd like for me to attempt something else.

 

Very Dead Battery