Battery Capacity Not Adding Up

When I use a public charger and get a report of kWh delivered, I often feel like the amount of SOC increase doesn't jive with the energy delivered. It charges "too easily": for the amount of energy delivered, I shouldn't see so much SOC increase. I also see this with home L1/L2 charging.

The most recent example of this is a 14.38 kWh L3 charge, where SOC increased from 22% to 100% (a 78% increase). Next to this entry in my log, I noted: "How does 14.4 kWh fill nearly 80% of a 24 kWh battery?".

Setting aside onboard charger inefficiencies, 14.38kWh / 78% yields a battery capacity of 18.44 kWh. I know there is reserved capacity that brings the 24 number down to closer to 21. With onboard charger inefficiencies that 18.44 number drops to 16.6 (using 90% efficiency).

Today I saw this in a nearby thread:

SageBrush said:

A brand new 24 kWh battery pulls about 0.25 kWh from the meter for every 1% increase in battery SoC. The 0.25 kWh/% takes into account charging losses.

Click to expand...

Running the numbers from the above charging situation gives 0.184 kWh/%, yielding only 74% capacity compared to brand new. Yet I still have 12 bars and LeafSpy is giving me SOH number around 90% (the full range I've seen over the last 100+ days of using LeafSpy is 87%-94%). I checked some L3 session results from earlier in the year and the 0.18 number holds pretty well for those. These are 5-6-temperature-bar charges so can't blame too cold or hot.

Any idea what's going on? Are charge sessions being underreported? Seems unlikely. Is that 0.25 kWh/% number too high? Not seeing where else the discrepancy could be coming from (aside from actual capacity loss).

bobkart said:

The most recent example of this is a 14.38 kWh L3 charge, where SOC increased from 22% to 100% (a 78% increase). Next to this entry in my log, I noted: "How does 14.4 kWh fill nearly 80% of a 24 kWh battery?".

Click to expand...

It doesn't. While the battery is purported to have capacity of 24kWh nominal, the actual user-accessible capacity is more like 20kWH. And that is when new. The reported SOC% is based upon user-accessible capacity. So with some degradation your numbers don't seem unreasonable.

bobkart said:

Setting aside onboard charger inefficiencies, 14.38kWh / 78% yields a battery capacity of 18.44 kWh. I know there is reserved capacity that brings the 24 number down to closer to 21. With onboard charger inefficiencies that 18.44 number drops to 16.6 (using 90% efficiency).
...

Yet I still have 12 bars and LeafSpy is giving me SOH number around 90% (the full range I've seen over the last 100+ days of using LeafSpy is 87%-94%).

Click to expand...

21 kWh * 0.9 = 18.9 kWh

It's hard to know how closely SOH tracks capacity.

Nubo said:

While the battery is purported to have capacity of 24kWh nominal, the actual user-accessible capacity is more like 20kWH. And that is when new. The reported SOC% is based upon user-accessible capacity. So with some degradation your numbers don't seem unreasonable.

Click to expand...

That's reassuring. (I did mention the capacity held in reserve just after what you quoted.)

P.S. I used to live in Vallejo.

cwerdna said:

21 kWh * 0.9 = 18.9 kWh

It's hard to know how closely SOH tracks capacity.

Click to expand...

That's an interesting point. So those not tracking evenly could be the culprit. I.e. capacity drops much more quickly than SOH at least for the first parts of the curves.

Still not liking the prospect of only 3/4 capacity. Would be interested to see kWh-charge-per-SOC% from other similar-aged Leafs, to help me believe I don't have some outlier situation.

bobkart said:

Nubo said:

While the battery is purported to have capacity of 24kWh nominal, the actual user-accessible capacity is more like 20kWH. And that is when new. The reported SOC% is based upon user-accessible capacity. So with some degradation your numbers don't seem unreasonable.

Click to expand...

That's reassuring. (I did mention the capacity held in reserve just after what you quoted.)

P.S. I used to live in Vallejo.

Click to expand...

Sorry; I did rush past some of that math.

Be careful of combining multiple values of unknown precision, and rules-of-thumb, into an equation to draw conclusions; drift can accumulate quickly. Even LeafSpy values are based upon best-guesses about the vehicle's internal data representation.

The actual capacity can be tricky to determine. Iirc, there are threads from the peak of the LEAF "battery-gate" fiasco that give recommendations on how best to empirically determine remaining capacity; involving lots of driving and measuring of wall-power. Tony Williams coordinated some of the detailed testing that was done in Arizona, I think. You could look for those if you're really into it. Probably easier to just put your faith in LeafSpy's full-charge "Gids" reading and call it a day.

How long ago since Vallejo?

OP: has your car had the Nissan software patch applied ?
The essence of the patch was to correct the SoC meter

SageBrush said:

OP: has your car had the Nissan software patch applied ?
The essence of the patch was to correct the SoC meter

Click to expand...

No such patch for a 24kWh battery since 2012...

Nubo said:

Be careful of combining multiple values of unknown precision, and rules-of-thumb, into an equation to draw conclusions; drift can accumulate quickly.

Click to expand...

Yeah I was going to make a similar point in one of my earlier replies. But more towards questionable accuracy rather than precision.

Nubo said:

Probably easier to just put your faith in LeafSpy's full-charge "Gids" reading and call it a day.

Click to expand...

Good suggestion. I'll check that the next time I fully charge (not very often).

Nubo said:

How long ago since Vallejo?

Click to expand...

It was quite a while ago (~1970). River Meadows had just been built and we had nothing but cow pastures all the way to the Napa river beyond our back yard.

SageBrush said:

OP: has your car had the Nissan software patch applied ?
The essence of the patch was to correct the SoC meter

Click to expand...

Looks like that doesn't apply to my model year.

Given that you're the source of the 0.25 kWh / SOC% number (for a brand new 24 kWh battery), I'd be curious to see some of your kWh / SOC% numbers. Still wanting to believe my battery capacity is close to typical and not having some problem.

bobkart said:

SageBrush said:

OP: has your car had the Nissan software patch applied ?
The essence of the patch was to correct the SoC meter

Click to expand...

Looks like that doesn't apply to my model year.

Given that you're the source of the 0.25 kWh / SOC% number (for a brand new 24 kWh battery), I'd be curious to see some of your kWh / SOC% numbers. Still wanting to believe my battery capacity is close to typical and not having some problem.

Click to expand...

IIRC I posted the results in this forum. At the time I performed the test on a DC charger I was not sure if the reported kWh delivered was before or after conversion losses. I was subsequently told that it was metered electricity but I cannot recall if that information required a recalculation.

Fwiw the 0.25 kWh/% value is a calculation based on 22 kWh usable with a new battery and a presumed 12% charging loss with L2 charging so
22/0.88

A DCFC that reports kWh to the car (after inverter and rectifier conversions) would have a ratio of 0.22 kWh/% for a new battery

---
Re: the software update, I think you and WetEV are correct that it applied to the early 30 kWh models.

Re: the software update, I think you and WetEV are correct that it applied to the early 30 kWh models.

Click to expand...

I think that it has been applied to (or suggested for) all but maybe the last batches of 30kwh packs. Given the mixed results of the patch, I remain convinced that this was more Nissan trying to reduce warranty pack replacements than it was an effort to fix a widespread BMS calibration issue. There apparently was a problem with some BMS programming, but for many if not most of the packs to which it was applied it seems to have done little more than make the BMS a little more optimistic, and to temporarily reset the capacity bars to 12.

SageBrush said:

A DCFC that reports kWh to the car (after inverter and rectifier conversions) would have a ratio of 0.22 kWh/% for a new battery.

Click to expand...

That number fits my observations much better, yielding a capacity of nearly 84%, much closer to the ~90% SOH reported by LeafSpy. With a downward adjustment of that 22 kWh number to 21.5 kWh, it moves up to nearly 86%.

I may give it a full L3 charge tomorrow and note the GIDs . . . 281 is the baseline number for the 24 kWh battery, right?

I'd still be interested in seeing other people's kWh-per-SOC% numbers.

I know this thread talks more about the smaller packs, but I am starting to wonder whether Nissan BMS is managing down the battery packs to hit the 70% capacity almost exactly at the 8 year mark. the .01 a day, translates to 3.5% per year, which is 27% in 8 years.

DougWantsALeaf said:

I know this thread talks more about the smaller packs, but I am starting to wonder whether Nissan BMS is managing down the battery packs to hit the 70% capacity almost exactly at the 8 year mark. the .01 a day, translates to 3.5% per year, which is 27% in 8 years.

Click to expand...

Intereresting question. I ask something similar when I see battery-replacement procedures that cover resetting the so-called 'gradual capacity loss data'. If a battery's health metric is strictly derived from its measurable characteristics, then such data wouldn't be needed. But if they added some kind of 'clock' that ticks away at the health over time (in combination with measurables), then your theory starts to track better.

Back to the question of how well kWh-per-SOC% can be used to determine battery capacity, one thing I now realize is that the 'baseline' number (for a new battery) will be different depending on the charging rate. This is due to the relatively-fixed overhead that charging incurs (circulating coolant for example). And it's not even as easy as a different number for each of L1/L2/L3, because there are differing charging rates even within those groups.

When this occurred to me was as I was planning a full L3 charge, to at least read off the GIDs once there. But also the kWh delivered versus SOC% increase could be examined. But due to the tapering off of L3 charge past 80% SOC, that fixed overhead would consume a larger fraction of the total energy delivered in that tapering-off segment (leaving less to make it to the battery).

Seems like a tru(er) formula for knowing energy added to the battery from energy coming out of the EVSE would need to consider the amount of *time* the charge took, in addition to onboard-charger efficiency. I suspect the 0.25 kWh / SOC% number is for the maximum DC rate the Leaf can accept, *and* with no tapering off considered (otherwise the number needs to go up).

To follow up on this: I got a full (L2) charge from home this morning and LeafSpy Pro told me 263 GIDs. Also that's its 93.6% of the full number (281 then, as I thought I had heard).

That tracks well with the ~90% SOH I see. Agree though that SOH need not be closely reflective of battery capacity.

Thanks for eveyone's help.