2016-2017 model year 30 kWh bar losers and capacity losses

My Nissan Leaf Forum

Help Support My Nissan Leaf Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
jbuntz said:
dwl said:
jbuntz said:
VLB is at the same number of GIDs but SOC in LeafSpy is much higher. At turtle SOC still reported 5.7%. Prior to update SOC at turtle was 0.0%
Useful info. Prior to the update did you notice when SOC first reached 0% (how many GIDs)?
Prior to update up to 6 GID = 0% 9 GID =1.3% 41 GID = 11.5%. Now 5 GID = 5.7 and 5.4 %
These results are similar to what has been seen on the dyno prior to update. In one case the car reached very low SoC and about 9 GIDs and kept going for over 15 mins at a constant C/3 load (27A, probably enough for 45mph). On average the car was dwelling for about 30-40 seconds per GID but at 9 and 8 GIDs dwelt for about 7 mins on each of those GIDs before moving again. Calculation wrong.

Your earlier pre-update post of many miles from LBW to turtle is also very useful. The update appears to have fixed this issue and your experience over time will be of great interest - thanks for sharing.

An update has just been released in New Zealand on our perspective of the update https://flipthefleet.org/2018/30-kwh-nissan-leaf-firmware-update-to-correct-capacity-reporting/
 
dwl said:
jbuntz said:
dwl said:
Useful info. Prior to the update did you notice when SOC first reached 0% (how many GIDs)?
Prior to update up to 6 GID = 0% 9 GID =1.3% 41 GID = 11.5%. Now 5 GID = 5.7 and 5.4 %
These results are similar to what has been seen on the dyno prior to update. In one case the car reached very low SoC and about 9 GIDs and kept going for over 15 mins at a constant C/3 load (27A, probably enough for 45mph). On average the car was dwelling for about 30-40 seconds per GID but at 9 and 8 GIDs dwelt for about 7 mins on each of those GIDs before moving again. Calculation wrong.

Your earlier pre-update post of many miles from LBW to turtle is also very useful. The update appears to have fixed this issue and your experience over time will be of great interest - thanks for sharing.

An update has just been released in New Zealand on our perspective of the update https://flipthefleet.org/2018/30-kwh-nissan-leaf-firmware-update-to-correct-capacity-reporting/

Thanks for the update.

My 2017 Leaf S just got the software update. I've had it since September 2017. It went from SOH 86% in the winter at 2500km, when I first got LeafSpy, to SOH 89% in the spring after recalibrating (running down to turtle, then charging to 100%), to SOH 95% after getting the update at about 7000 km.

I haven't tried any range tests yet, but in the winter it did seem like I got a lot more km out of 10% of the battery at the top end then I did at the bottom end. Eg. from 90%-80% I would get 12km, but from 40%-30% I would get 8km. It felt like the GOM and the % remaining would start going down faster as the battery got lower even if the driving speed was constant.
 
WetEV said:
lorenfb said:
Amazing what one can do by changing a few binary numbers, right?

Why not do a range test and see what the reality is, rather than assuming Nissan is wrong?

I don't have a 30 kWh Leaf. We'll see over time what the actual/true result of the firmware update is.
But isn't strange that the BMS firmware for my 24 kWh Leaf hasn't "over estimated" SOH degradation?
Did Nissan use another engineering team to write the BMS code for the 30 kWh versus the 24 kWh Leaf,
which recently required an update that's coincidental with recent increased battery warranty claims?
 
lorenfb said:
But isn't strange that the BMS firmware for my 24 kWh Leaf hasn't "over estimated" SOH degradation?
Did Nissan use another engineering team to write the BMS code for the 30 kWh versus the 24 kWh Leaf,
which recently required an update that's coincidental with recent increased battery warranty claims?

Nissan's claim is that they tested the batteries replaced under warranty, and these packs were not degraded enough to require replacement.

Have you any measurements that show otherwise?

If not, why not?

If not, why the allegations?

I've written a lot of code in my life, which means I written a lot of bugs. I can think of a dozen ways to write a bug that would make the BMS code might overestimate degradation of a larger pack. Why? Because I've seen such bugs, both other peoples and my own.
 
WetEV said:
lorenfb said:
Amazing what one can do by changing a few binary numbers, right?

Why not do a range test and see what the reality is,
Range tests are too susceptible to driving and environmental variables.

Charging up is a much better test, albeit with perhaps a 5% variation in charging losses even if the test is consistent with L1 or L2 charging. I'm coming around to the idea that the update changed the LB and VLB thresholds. Since people use VLB as a practical range limit, the update has merit so long as it does not drag the battery down into SoC levels at VLB that will accelerate degradation.

Personally, I like the idea of minimum cell 3.3 V as a VLB threshold, and some user experiences report that trigger is now much lower ... in the 2.5 - 2.6 V range if memory serves. If true, this update is not going to end well. A Tesla owner with a 85 kWh Model S monitored cell voltages via canbus and found that the car shut down when a cell hit ~ 2.97 V.
 
dwl said:
jbuntz said:
dwl said:
Useful info. Prior to the update did you notice when SOC first reached 0% (how many GIDs)?
Prior to update up to 6 GID = 0% 9 GID =1.3% 41 GID = 11.5%. Now 5 GID = 5.7 and 5.4 %
These results are similar to what has been seen on the dyno prior to update. In one case the car reached very low SoC and about 9 GIDs and kept going for over 15 mins at a constant C/3 load (27A, probably enough for 45mph). On average the car was dwelling for about 30-40 seconds per GID but at 9 and 8 GIDs dwelt for about 7 mins on each of those GIDs before moving again. Calculation wrong.

Your earlier pre-update post of many miles from LBW to turtle is also very useful. The update appears to have fixed this issue and your experience over time will be of great interest - thanks for sharing.

An update has just been released in New Zealand on our perspective of the update https://flipthefleet.org/2018/30-kwh-nissan-leaf-firmware-update-to-correct-capacity-reporting/
This report is the best explanation of the update I have read and correlates with my experience. Unfortunately after the update I am still at 82% SOH and falling. No longer 8 bars but still heading in that direction. I should be due a new battery by this time next year.
 
Awesome report from New Zealand.
I'd like to recommend the update to a 30 kwh LEAF owning friend if I can confirm that the lowest cell voltage threshold for VLB has not been lowered into accelerated degradation territory.
 
SageBrush said:
WetEV said:
lorenfb said:
Amazing what one can do by changing a few binary numbers, right?
Why not do a range test and see what the reality is,
Range tests are too susceptible to driving and environmental variables.

Charging up is a much better test, albeit with perhaps a 5% variation in charging losses even if the test is consistent with L1 or L2 charging. I'm coming around to the idea that the update changed the LB and VLB thresholds. Since people use VLB as a practical range limit, the update has merit so long as it does not drag the battery down into SoC levels at VLB that will accelerate degradation.
I agree range tests can be variable which is why the tests here were dyno based with external current metering. Significant efforts were also made to temperature stabilise the battery (needed heating, cold here in New Zealand at the moment). The final answer is how much energy can the battery deliver when driving.

One of the dealers here uses the charge method to prove actual capacity and this result doesn’t change with the update. One friend who had dropped to 80% SOH on Leaf Spy had the check done and charge method showed around 88% SOH so a dealer could claim no fault. However as this issue affected LBW and VLBW levels it meant the effective range was reduced. This has been a useful wake up from Nissan that EVs have their own complexities.

Another issue is mentioned in a local forum - especially relevant if people don’t like letting their car go below LBW:
“I think something important has been left out of the conversation here. The patch doesn't just fix the perception of range, it also results in more of the battery being used between charges, and thus the battery being at a lower state of charge on average. If I can charge my car every 3 days instead of every 2 days, my average state of charge will be lower and my (real) battery degradation should be lower over time.

So not applying the patch could cause it to decline in capacity sooner than if you had the patch installed.

Some of the rapid degradation we have seen, even after the patch is applied is thus very real, and caused by people keeping their batteries at a higher state of charge than they knew.

This is probably a small effect, but noteworthy for those of us who have a choice around charging and want to keep the battery in top shape.“
 
^^ Regarding that dyno test, I'd like to know if the results change when a higher discharge rate is used, say 1 C.

Second, do you have an opinion or data on what minimum cell voltage is reached at VLB in these post-update packs ?
 
SageBrush said:
^^ Regarding that dyno test, I'd like to know if the results change when a higher discharge rate is used, say 1 C.

Second, do you have an opinion or data on what minimum cell voltage is reached at VLB in these post-update packs ?
I managed to get a brief chat with the specialist and a test was done at 1C which did create a significant rise in temperature. I would expect slightly less capacity but didn’t get any data. By staying down at C/3, which is also a reference used by AESC, the temperature rise was only a few degrees and removes that as a variable.

Cell voltages were usually not taken down very low to turtle as the amount of energy on that last slope (e..g once under 2.9V per cell) is very little. These cars are owned by others and don’t want to stress the battery.
 
dwl said:
Cell voltages were usually not taken down very low to turtle as the amount of energy on that last slope (e..g once under 2.9V per cell) is very little. These cars are owned by others and don’t want to stress the battery.
I'm asking about minimum cell voltage at VLB
 
SageBrush said:
dwl said:
Cell voltages were usually not taken down very low to turtle as the amount of energy on that last slope (e..g once under 2.9V per cell) is very little. These cars are owned by others and don’t want to stress the battery.
I'm asking about minimum cell voltage at VLB

Sage is trying to find some reason to complain about the Leaf's battery.
 
WetEV said:
SageBrush said:
dwl said:
Cell voltages were usually not taken down very low to turtle as the amount of energy on that last slope (e..g once under 2.9V per cell) is very little. These cars are owned by others and don’t want to stress the battery.
I'm asking about minimum cell voltage at VLB

Sage is trying to find some reason to complain about the Leaf's battery.
OK, I ran my battery down to 13 GID's today. I couldn't get the value for VLB because I was going up a long grade and couldn't pull over to record it but I do have a value for VLBW (25 GIDS, 6%) 334 VDC (3.48 V/cell). 9% (32 GID's) was 336 VDC (3.50 V/cell) and 3%(13 GID's) was 330 VDC (3.44 V/cell). I didn't have any desire to find turtle so that's not included. Referring back to my notes for the original battery when new, I can see that the new battery is holding a higher voltage at low charge ( old battery was 317VDC at 3%) but that;s the only difference I see.

LeafSpy was reporting 9.5% SOC at the 13 GID mark. It normally reports 97.5% SOC at full charge (100% on the dash gauge and 363 GID's in LeafSpy). I honestly don't know how much further I could have driven it so It could gone another couple or three miles (13 GID's=1 KWH) or perhaps as much as another 10 miles (9.5%=2.85 KWH). I'm not that adventurous. I would be interested in what others come with in regards to the cell voltages.
 
johnlocke said:
OK, I ran my battery down to 13 GID's today. I couldn't get the value for VLB because I was going up a long grade and couldn't pull over to record it but I do have a value for VLBW (25 GIDS, 6%) 334 VDC (3.48 V/cell). 9% (32 GID's) was 336 VDC (3.50 V/cell) and 3%(13 GID's) was 330 VDC (3.44 V/cell). I didn't have any desire to find turtle so that's not included. Referring back to my notes for the original battery when new, I can see that the new battery is holding a higher voltage at low charge ( old battery was 317VDC at 3%) but that;s the only difference I see.
I'm most interested in the lowest cell pair voltage, since that is the trigger for the VLB warning.

Thanks for sharing.
 
SageBrush said:
johnlocke said:
OK, I ran my battery down to 13 GID's today. I couldn't get the value for VLB because I was going up a long grade and couldn't pull over to record it but I do have a value for VLBW (25 GIDS, 6%) 334 VDC (3.48 V/cell). 9% (32 GID's) was 336 VDC (3.50 V/cell) and 3%(13 GID's) was 330 VDC (3.44 V/cell). I didn't have any desire to find turtle so that's not included. Referring back to my notes for the original battery when new, I can see that the new battery is holding a higher voltage at low charge ( old battery was 317VDC at 3%) but that;s the only difference I see.
I'm most interested in the lowest cell pair voltage, since that is the trigger for the VLB warning.

Thanks for sharing.
I believe the trigger for the warnings (except turtle) isn't voltage but GIDs with the change to 49 and 24 GIDs for LBW and VLBW respectively. These are samples from part of one test where coincidentally SOC dropped to 0% at VLBW - note the timestamps in h:mm and extended period around 8 and 9 GIDs - the current was set at 27A on an external meter:

dyno-table-lower-res.jpg


As the battery ages, the voltage at the knee may not change a lot but the Ah curve reduces the location of the knee on the time axis. Voltage could be used but using GIDs may be a key reason why the calculation is causing more range anxiety - the VLB and VLBW thresholds are higher than expected above available capacity.

Using the example of 336V, the dyno test result had VLBW (2kWh) coming on while the measured capacity at that point down to cutoff (bit before turtle) was actually 3.74kWh.
 
SageBrush said:
@dwl,
Very informative, and reassuring that the update is not going to harm the battery.

Thanks
To clarify, the data above is pre update. The update seems a genuine fix for the reporting issues most obvious where the car kept running with GIDs stuck at 9 then 8. The update may be beneficial to the battery as on average it can probably be left with less actual charge to achieve the same reported range.
 
Back
Top