In the last few days I reset the BMS in my car, as described in the following thread:viewtopic.php?f=23&t=3943&start=60#p38994
In hindsight not a clever idea because it resets the Ah capacity back to the factory default of 45.8Ah, and the car is very slow/reluctant to re-learn the true capacity when it differs so much from the factory new capacity!
(On a positive note, we now know that the factory new capacity of these packs is 45.8Ah! Thus we can confidently use this figure to calculate % SoH)
Due to this, the BMS grossly over estimates the range of the car and will try to drive the battery right down to the 3.0 volt cutoff point at which point you are very suddenly restricted to 7mph. Not good! So to solve this in a timely fashion I had to do the battery calibration procedure as detailed in that thread. For anyone wanting to force the BMS to re-measure the Ah capacity of the battery in a safe way - just do the battery calibration procedure under maintenance, not the full BMS reset! (battery replacement procedure)
I decided to analyse some of the new data I've found from the latest calibration and recent Canion voltage graphs regarding the state of my cells. Earlier in this thread I noted that I have two cells that are particularly weak and a third that is well on the way to being weak, and that my suspicion for a while has been that these two or three cells are responsible for the majority of the seemingly accelerated capacity loss I've had over the last year, and I think I have the data to prove that now.
So first some Canion voltage graphs. This first graph I've posted before and was taken at 10% SoC in April 2018 at about 41k miles, which was when I first started to notice abnormally fast drops in Ah:
You can see there is a 50mV spread at 10% SoC - not terrible, but not good either. Fast forward to October 2018 at about 47k miles and it is now a 125mV spread for the same SoC. Ouch:
About a week ago I recorded this voltage graph at a slightly higher 14% SoC, and at 52k miles. The spread is now 135mV at 14% SoC - as the spread increases at a lower SoC it's worse than it looks at about 10% SoC it's more like 150mV now. I've highlighted in red the cells that I have particular concerns about and propose replacing:
Also of note is that some cells have high internal resistance. This can be tested by discharging the car to a relatively low SoC like 20% then Chademo rapid charging the car. Cells with high internal resistance will try to go over voltage very quickly. Here is a voltage graph taken only a few minutes into a Chademo charging session where the charge rate is already throttled back due to Cell 25 hitting the maximum allowed voltage. I've highlighted three cells that I have concerns about high internal resistance, not surprisingly 3 of the same 4 cells from above:
The difference between 4.075v and 4.105v might not seem like much, but when repaid charging this can mean the difference between pushing the good cells up to 43kW (before reaching 4.105 volts) and having to throttle way back to 21kW to keep the high resistance cells from going over voltage.
In short, the weak cells are seriously limiting rapid charge speeds as well. In fact no matter what SoC I start at my charge rate only remains at 43kW for less than a minute, as shown in the following graph where I started charging from about 30% SoC - note that it doesn't even last 30 seconds at the full 43kW before throttling back, and it is painfully slow by even 55%:
So I've known for a while the difference between best and worst cells is getting progressively worse but I couldn't put a quantitative figure on it as I can't directly relate the relative cell voltage near discharge to the capacity of the cells.
After running the battery calibration routine I've realised that it actually measures the individual cell capacities and reports the minimum and maximum cell capacities of the pack! And I also ran this calibration and kept the results about a year ago so now I have two data points. So a year ago it reported this:
And now today:
So in the last 12 months and 12k miles the best cell has degraded only 0.5Ah (better than I would have expected!) while the worst has degraded 3.3Ah. So the spread has increased from 2.9Ah to 5.7Ah in the last year.
I've added these new data points to my excel spreadsheet and graph and have come up with the following interesting display of the data:
The dotted blue line is actual Ah measurements taken by Canion or Diagbox, with the yellow line being a linear interpolation of the degradation rate. The red line shows the degradation rate of the weakest cells as reported by the battery calibration routine which pretty much exactly follows the yellow line.
The green line shows the degradation rate of the best cells in the pack - which is dramatically slower than the worst cells, and you can clearly see the gap between the green and red lines opening up...
The question is, if I replaced those 4 cells, what would the new minimum capacity be ? Since the weak cells seem to be significant outliers I think I could recover at least half the lost capacity if I replaced them with better cells - I'm estimating that it would go back up to about 38Ah, and more importantly and hopefully arrest the rapid degradation that I'm seeing at the moment. If I could get it back up to about 38Ah and reduce the degradation rate back to what it was before 40k miles it would add an extra several years of useful life to the car for me.
The way the weakest cells are degrading at the moment I will not be able to make my 35 mile daily commute next winter without a rapid charge on the way home every day - and at the reduced rapid charge rates I'm now seeing, painfully slowly...