Auto-stop charging at 80%?

[Snargleblarg]

No BMS does a perfect job of measuring SOC. On this forum, the LeafSpy reported SOC seems to be taken as the gospel truth, but the analog front end (a Maxim chip I think) will have some error (maybe 10mV or more) per channel across process, calibration, voltage, temperature, humidity and long term drift. And it's very difficult to define SOC at mid-capacity. The only well known points are at full - defined by the charge termination threshold, and at empty - defined by some low voltage operation threshold. Otherwise the BMS software has to estimate SOC based on some combination of coulomb counting, which is linear and accurate in the short term, and open-circuit resting voltage, which is a more stable estimate in the long term. I wouldn't be surprised to see some small instantaneous adjustments occurring, given all of the variables in play. I expect the raw capacity measurement actually moves quite a bit more dramatically and we're only seeing small movements after some software smoothing is applied.

I've seen reports of some other EVs shutting down before 0% reported SOC. Thankfully the Leaf does not suffer that. Nissan obviously chose to err on the conservative side when choosing their detection of 0% SOC. At 0% reported by LeafSpy it seems the software shuts down operation before the actual limit of circuit operation.

 

[SageBrush]

How do you figure ?

You can't see it on the dash, but with the help of LeafSpy, using mine as an example. The *dash* says that the battery is at 100% SOC, but checking LeafSpy shows it's only around +56 kWh out of the 62 kWh capacity. 56/62 = .90322 or roughly 90% total charge capacity out the max total.

Degradation is not unused capacity

 

[LeftieBiker]

The pack voltage doesn't drop with capacity. When the BMS (and LeafSpy reading directly from it) measures 90% it isn't basing that on capacity - that's what the capacity bars are for. (Then Nissan's display changes that reading to what they want the state of charge display to read instead, in order to make it less likely that the driver will overcharge or drain the pack.) What happens with degradation is the voltage drops faster as the same quantity of energy is pulled from the pack. The pack voltage at 100% of what Nissan allows for a full charge doesn't drop as the pack degrades.

 

[knightmb]

I agree, it's not perfect, but I don't what other tool one could use without having to pull the entire pack to do direct measurements and testing. That's why the BMS keeps a data history of voltage levels and power usage to figure out how much energy is going into the pack and how much is going out of the pack. I don't depend on the LeafSpy SOC calculation myself, that's why I would rather see the actual power calculations that the Leaf thinks the battery has remaining, since that is all LeafSpy can see. If the BMS thinks that the battery has 56 kWH of stored energy and the pack voltage is high enough to be considered 100% for what Nissan allows, I tend to believe it. Myself and others here have done long drives of draining the pack down to almost nothing with LeafSpy recording the entire way, it's pretty darn close to what the BMS thinks the battery has left at the end. Unlike LeafSpy though, the BMS is doing constant re-adjustments to it's calculations based on the data it collects. I have wondered just what pile of data the BMS is collecting all this time. My guess is just a large spreadsheet of dates and numbers with just some averaging of everything based on some preset algorithms that probably relate to the battery chemistry.

No BMS does a perfect job of measuring SOC. On this forum, the LeafSpy reported SOC seems to be taken as the gospel truth, but the analog front end (a Maxim chip I think) will have some error (maybe 10mV or more) per channel across process, calibration, voltage, temperature, humidity and long term drift. And it's very difficult to define SOC at mid-capacity. The only well known points are at full - defined by the charge termination threshold, and at empty - defined by some low voltage operation threshold. Otherwise the BMS software has to estimate SOC based on some combination of coulomb counting, which is linear and accurate in the short term, and open-circuit resting voltage, which is a more stable estimate in the long term. I wouldn't be surprised to see some small instantaneous adjustments occurring, given all of the variables in play. I expect the raw capacity measurement actually moves quite a bit more dramatically and we're only seeing small movements after some software smoothing is applied.

Yeah, I was able to get my wife's Leaf to do that when her cell balance was all over the place when she bought hers. If just one cell gets down to the 2.5 voltage range, the entire pack is shut off, regardless of the other cells. Luckily, her pack only took a few gentle deep discharge sessions to get everything lined up. I can only imagine the frustration if someone had a pack with a genuinely bad cell that causes unexpected shutdowns when the pack capacity gets near the bottom or just hitting the accelerator hard to lower the voltage enough to trigger it when the rest of the cells are fine.
That's probably why the Leaf has such a large gap near the bottom of the pack. It probably scares the driver more than anything, but at least it keeps them from literally losing power in the middle of a drive which might be in high speed traffic or some other situation that would be dangerous to just all of sudden lose speed.
I have to think back though, I've driven many ICE cars over the years that had sudden engine failures due to belts breaking, cam sensor going bad, etc. At least with my EV experience, it just takes a plug to fix the issue and not some expensive engine repairs.

I've seen reports of some other EVs shutting down before 0% reported SOC. Thankfully the Leaf does not suffer that. Nissan obviously chose to err on the conservative side when choosing their detection of 0% SOC. At 0% reported by LeafSpy it seems the software shuts down operation before the actual limit of circuit operation.