12% capacity loss in 9 months is "normal"

[TickTock]

After reading through this thread I am having a thought… a charger problem??? I am wondering if this could be an overlooked cause of your lower state of charge.

Do you have the QC port? Could you do a QC and then read your gids? I believe a QC completely bypasses the onboard charger.

Would someone with a bit more experience or knowledge with the communication between the BCM and charger comment about the possibility of a charging issue?

I don't think so. My final pack voltage (394V) is consistent with everyone else... Unless the voltage sensor is out of cal. Wouldn't that be nice if it was as simple as that. However, I don't think so since my pack creeped back up a bit during the first 6 months. I think the most probable explanation is the pack saw some abuse prior to delivery (stored at high temp, full charge for a long time) and I am observing it recover a little, but the damage was still done.

But I will try a QC (have not yet). Be one more datapoint.

 

[DaveEV]

Regarding the CVLI, I think the 0.06V difference between the highest and lowest cell could explain some of the low capacity if:
- the pack is balanced
- the lowest voltage cell is an outlier
- the highest voltage cell is not an outlier

I think this is an important point. If all of these conditions are true, which is the worst case for cell 25, then you can expect that cell 25 will receive quite a severe overcharge whenever the pack is charged to 100%, which should greatly shorten its life. (This assumes that the Leaf does not actively shunt it during charge in order to protect it. I seriously doubt they are doing something like that.).

The LEAF definitely takes measures to avoid over-charging any single cell-pair and over-discharging any single cell. Each cell-pair has a shunt which can be activated to drain energy from high voltage cells. (Remember that the LEAF's battery is made up of 48 modules, each module containing 4 cells in a series/parallel setup).

 

[TickTock]

Regarding the CVLI, I think the 0.06V difference between the highest and lowest cell could explain some of the low capacity if:
- the pack is balanced
- the lowest voltage cell is an outlier
- the highest voltage cell is not an outlier

I think this is an important point. If all of these conditions are true, which is the worst case for cell 25, then you can expect that cell 25 will receive quite a severe overcharge whenever the pack is charged to 100%, which should greatly shorten its life. (This assumes that the Leaf does not actively shunt it during charge in order to protect it. I seriously doubt they are doing something like that.).

The LEAF definitely takes measures to avoid over-charging any single cell-pair and over-discharging any single cell. Each cell-pair has a shunt which can be activated to drain energy from high voltage cells. (Remember that the LEAF's battery is made up of 48 modules, each module containing 4 cells in a series/parallel setup).

Yeah, but as I pointed out earlier. 60mV per cell near dead only buys my about 80wH (1 gid). So It appears all the cells are pretty well matched.

 

[IBELEAF]

Aside what the GIDs tell you did you actually notice that you are getting 20% less range (2-3 bars more?) for you daily commute?

 

[vegastar]

Yeah, but as I pointed out earlier. 60mV per cell near dead only buys my about 80wH (1 gid). So It appears all the cells are pretty well matched.

AFAIK the CVLI is done at the LBW point (around 356,5V pack voltage) and not at the VLBW. The slope there is a lot more flatter than at VLBW.

Edit: I checked your voltages and they are indeed near the LBW point (3,715V).

 

[FLBoy]

If a cell-pair is low (high) when the pack is low AND low (high) when the pack is high, the c-p is out of balance.
However, if it is low when the pack is low, and high when the pack is high, the c-p has LOW capacity.

One needs two cell voltage tests to detect this condition (one at high charge and one at low change).

Gary, I don't follow your second sentence. If a cell-pair tracks with the rest of the pack, how can you conclude from that alone that the cell-pair has low capacity, as opposed to high capacity or normal capacity? :?

 

[GaslessInSeattle]

yeah, I agree 100%. Whether this turns out to be a charger issues or not, Nissan is now on record as to how they plan to respond to unprecedented battery capacity loss... clearly this is PR suicide! Surely they realize: "the whole world is watching!"

Carlos, go smack some sense into your people and get to the bottom of this, replace the whole pack FOC if you have to!!!

Nissan needs to fix this particular car (replace battery / cells etc). Nobody would want to buy a car where 20% capacity loss in the first year is "normal".

 

[smkettner]

yeah, I agree 100%. Whether this turns out to be a charger issues or not, Nissan is now on record as to how they plan to respond to unprecedented battery capacity loss... clearly this is PR suicide! Surely they realize: "the whole world is watching!"

Carlos, go smack some sense into your people and get to the bottom of this, replace the whole pack FOC if you have to!!!

Nissan needs to fix this particular car (replace battery / cells etc). Nobody would want to buy a car where 20% capacity loss in the first year is "normal".

I think Nissan may have more actual data on this car than we do. I am not convinced the issue is that big.

 

[Stoaty]

I think Nissan may have more actual data on this car than we do. I am not convinced the issue is that big.

I would be pretty unhappy if I had a 9.3% loss of capacity in less than one year. What makes you think that the Gid-o-meter isn't giving an accurate picture of capacity? Sure, that isn't 20%, but it is pretty significant, especially since the data indicates that this problem was present pretty much from the beginning. In other words, the buyer never had the capacity that Nissan claims.

 

[TickTock]

Yeah, but as I pointed out earlier. 60mV per cell near dead only buys my about 80wH (1 gid). So It appears all the cells are pretty well matched.

AFAIK the CVLI is done at the LBW point (around 356,5V pack voltage) and not at the VLBW. The slope there is a lot more flatter than at VLBW.

Edit: I checked your voltages and they are indeed near the LBW point (3,715V).

Good call. You are correct that the test was run at LBW. So the slope there is 10V/35gids or ~3mV/gid. Therefore the most 60mV max-min can account for is 1.6kWh or ~6% degradation from a perfectly matched battery. It would be really helpful if we knew how a typical battery which still takes a full charge scores on the CLVI test. I guess I'll double my efforts to figure out how to read all 96 cells...

The slope observation is a good point. In the SM, it just says to get down to 1 or 0 bars. The results of the test will vary dramatically between LBW and VLBW. Anyone else who goes in for this test may want to ensure they are at VLBW when they arrive to get the best indicator (and record your gids before handing off the car).

 

[aqn]

What makes you think that the Gid-o-meter isn't giving an accurate picture of capacity?

Quotes from Gary:

"No, it appears that the "SOC" number is not linear, or even dependable, especially at the low end."

"We call the LEAF's value "SOC" (in quotes), for lack of a better name. Maybe "Fuel" is better, since the LEAF appears to derive its Fuel-Bars from this number. But, whatever the name, it is not a perfect indicator of ... anything.

We have no inside information about the meaning of the CAN-buss values ... it is all (slightly) educated GUESSES.

But, the "SOC %" and "SOC raw" (that is from the EV-CAN buss) are equivalent."


"The SOC Meter is just repeating a value that it reads from the LEAF.

But, the LEAF probably gets things wrong ("lies") occasionally, so please
BEWARE of possibly (occasionally) misleading (lower) percentage values
that are too optomistic about the useful energy left in the battery."

"Since this new "ScanGauge-for-the-LEAF" will not work with "normal" cars (due to different pins being used for the EV-CAN buss, if nothing else), I suggested that they call it the "LeafGauge". Or, possibly the "ScanGauge-EV", but that might be too general, since this would be specific to the LEAF, and it MIGHT even be specific to the 2011 LEAF.

IF the LeafGauge uses a different divisor (not 281, but 300), that would explain their readings of 76.8% instead of 82.2% after a Charge-to-80% session.

So, for now, I will assume that the 300 "gids" equals 100% is what they are doing, and thus LEAFfan's somewhat low "soc%" readings are explained.

Next week I will talk with their "technical" guy, and probably find out more, perhaps even give them some useful advice.

 

[Stoaty]

From Ingineer:

Keep in mind "Gids" are actually the Battery ECU's Watt-hours number (times 80), so as the battery ages, expect this number to drop, while SOC will still rise to ~95%. This means the watt-hours/SOC will slowly lessen.

 

[Ingineer]

Keep in mind the Watt-hours remaining figure, ("Gids") is derived from the total calculated capacity of the battery times SoC. The SoC is derived mainly by Coulomb counting, (watching amps in/out of the pack real-time) However, the device Nissan uses to do this is called a Hall-Effect current transducer, which works off the magnetic field generated around a wire when passing current. Because it's using magnetic fields and the device used to detect them is a semiconductor, it's sensitive to external factors, such as the earth's magnetic field, and temperature. They try to compensate for these factors, but it's not perfect.

Because of these slight variable errors, the measurement has drift, and thus will become more and more inaccurate over time, so Nissan solved this problem by correcting the SOC periodically when it doesn't line up with known battery conditions. A side-effect of this is a sometimes artificially high SoC reading when the battery is near full or near empty. This means the SoC (and thus "Gids") will deplete a little faster initially and when nearing turtle. Just a "seat-of-the-pants" estimate is that it seems to have no more than 5% error at any time. I expect Nissan to fix this in future versions of the Leaf as they gain more knowledge.

This error, and the subsequently needed correction is why Hidetoshi Kadota, (Chief Vehicle Engineer for the Leaf) told us they decided to not give us the true SoC with reasonable accuracy. In other words, it's not accurate enough and they felt it might confuse/annoy customers. While I understand that logic, As far as I'm concerned, it's worse! They base the GOM calculations on this, so we get error on top of error! I guess they assumed that we'd get used to a GOM jumping all over the place. (Masking error via way of noise)

Hope this clears up understanding all this!

-Phil

 

[RegGuheert]

Reg,

Congrats on your new car.

Thanks!

I'm not sure how much you've followed along on here prior to your purchase, but nobody puts much thought into what the GoM says.

Yes, I'm aware of the issues, but I think you use the word 'nobody' pretty loosely! I will venture to guess that the vast majority of The 20,000 Leaf owners worldwide put a lot of thought into what it says. I promise you that my wife does. Why? Because that is what Nissan tells everyone to use.

Personally, I cover the thing, since bad data (to me) is worse than no data.

If I were to cover the GOM, I suppose I could use the bars and/or the miles/kWh reading, but I don't think those are overly useful without first knowing the capacity of my battery. I was planning on not purchasing a GID meter because I would prefer to wait and get the device that Phil is putting together.

Snipped excellent, detailed description of how to determine range of my Leaf!

Thanks for thes detailed description! I will definitely try to figure this out, but I simply won't have a chance until at least June. Until that time, we will just have to see how far this thing will take us. Fortunately we rarely will need to drive the leaf to the limits of its range. The excursions that will be the most challenging to predict involve crossing a mountain and back. I expect that no amount of calculating will tell us how many miles we can travel on the other side before we will be unable to make it back home. Around here, there is no such concept as making it to a charging station. They simply do not exist except at the Nissan dealers (the 2 or three we could reach).

Thanks again!

 

[RegGuheert]

The LEAF definitely takes measures to avoid over-charging any single cell-pair and over-discharging any single cell. Each cell-pair has a shunt which can be activated to drain energy from high voltage cells. (Remember that the LEAF's battery is made up of 48 modules, each module containing 4 cells in a series/parallel setup).

Thanks! I saw in the service manual that there are shunts, but I'm interested to learn more about the theory of operation of the shunts. What is the resistance (or current or power) of the shunts and how are they used? During what phases of charging are they used? How many can be on at once? If there is some detail in the service manuals or elsewhere, I would be greatful if you could point me to it.

TIA!

P.S. I think I've seen you on the NAWS forums previously!

 

[RegGuheert]

So the slope there is 10V/35gids or ~3mV/gid. Therefore the most 60mV max-min can account for is 1.6kWh or ~6% degradation from a perfectly matched battery.

I'm not convinced unless I know for sure the exact operation of the cell shunts. If those shunts operate during the *discharge* of the battery, then there would be no way to know what the capacity might be in the absence of the low-capacity cell. Do we know that the shunts do not operate during discharge? The words I see in the service manual make me wonder if they do.

I guess I'll double my efforts to figure out how to read all 96 cells...

That would be awesome! It would be cool to see histograms of cell voltages from our various cars at particular operating conditions.

 

[surfingslovak]

Because of these slight variable errors, the measurement has drift, and thus will become more and more inaccurate over time, so Nissan solved this problem by correcting the SOC periodically when it doesn't line up with known battery conditions. A side-effect of this is a sometimes artificially high SoC reading when the battery is near full or near empty. This means the SoC (and thus "Gids") will deplete a little faster initially and when nearing turtle. Just a "seat-of-the-pants" estimate is that it seems to have no more than 5% error at any time. I expect Nissan to fix this in future versions of the Leaf as they gain more knowledge.

Phil, we had this topic earlier, and I feel compelled to remind everyone that while the instruments and sensors in the car are not perfectly accurate, it will be very hard, if impossible, to explain the 8% discrepancy we see between energy in and out of the battery with instrument error alone. Aside from coulombic inefficiency, the only other source of losses that appear to be unaccounted for, is battery heat. I would expect this to be in the 1-2% range, and it does not explain the discrepancy fully. It would be good to keep that in mind. I think the Leaf is still good for a surprise or two.

 

[Ingineer]

we had this topic earlier, and I feel compelled to remind everyone that while the instruments and sensors in the car are not perfectly accurate, it will be very hard, if impossible, to explain the 8% discrepancy we see between energy in and out of the battery with instrument error alone. Aside from coulombic inefficiency, the only other source of losses that appear to be unaccounted for, is battery heat. I would expect this to be in the 1-2% range, and it does not explain the discrepancy fully. It would be good to keep that in mind. I think the Leaf is still good for a surprise or two.

Well don't forget the balancing system! The balancing shunts being discussed just before your post suck down about 1 amp each while on, and the come on throughout the charging process to keep the cells aligned.

-Phil

 

[surfingslovak]

Well don't forget the balancing system! The balancing shunts being discussed just before your post suck down about 1 amp each while on, and the come on throughout the charging process to keep the cells aligned.

Right, excellent point!

 

[RegGuheert]

Well don't forget the balancing system! The balancing shunts being discussed just before your post suck down about 1 amp each while on, and the come on throughout the charging process to keep the cells aligned.

-Phil

Since a level 2 charge is on the order of 10 amps DC it seems a 1-A shunt is capable of reducing the charge going into a given cell pair by about 10%. Since there are 96 cell pairs in series, that only comes to about 0.1% of overall charge power per shunt. Do you know how many shunts can be turned on at any one time?