What's best SOC for a cell pair voltage test?

I am hauling my LEAF to Palmdale on Friday for second annual mandatory battery usage report. After failing to make the 84 mile trip (both ways) last year, I'm going to haul the car on a flatbed trailer this year. According to my SOC meter, my car's battery has 80% of original capacity. Based on Stoaty's model, it has probably lost more like 15%.

Since I'm not driving it down, I can present the car for service at any SOC I want. Last year they did the cell pair voltage test without my asking after I had the car towed in, so I can probably get it done again this time. Unfortunately, they charged the battery to 70% before they did the test, and CPVs looked fine.

I'll probably tape the charge port door shut and write 'DO NOT CHARGE' on the tape.

Is there a 'best SOC' for cell pair voltage test? If so, I will run the battery down to that level before I load the car on my trailer. I'm thinking somewhere around VLBW. Would that put the cells well below the 'knee' on voltage?

Will this provide useful information? If not, I'll just haul it down at about 25% SOC, which is where it will be after my normal commute Thursday. That's where I try to leave the SOC when I'm not driving the LEAF for a day or two.

I'll also talk to the dealer about trade-in on leased 2012 LEAF, but I don't expect much encouragement there. The new battery capacity warranty may be my only hope.

-Karl

I think you might get the most information by having them take a set of readings at VLBW; then charge to 100%; then take another set of readings. If you can compare the lists of cell pair voltages from both measurements, you may be able to see some trends which might point to weak cells.

Gerry

When testing at low SOC, you want to test at VLBW at a minimum. The lower the SOC, the bigger the difference between cell-pair voltages if any are imbalanced.

As Gary suggests, testing again immediately after recharging to 100% would also be informative if all cell-pair voltages are read.

Yes, get the results of all 96 cell-pairs printed out
for both the low and high SOC conditions, if possible.

If Consult cannot print it, the Windows Print Screen
should be able to capture an image of the screen,
and they should be able to copy that image to
your thumb (flash) drive to show us.

Or,if they cannot print it out, possibly ask to take a
picture of the Consult screen.

I dont think you want to test it below the knee, almost every battery would fail I would think.

The service manual (EVB-65) states the minimum cell voltage must be less then 3712mV before taking the CVLI measurements (says to discharge the battery until this condition is met before proceeding). This corresponds to right at the knee (356 pack volts) so you are supposed to make sure it is at the knee *or lower*

Thanks, guys! It's hard to imagine the value of a CVLI at 100% charge; all cells will read 394.5 / 96, more or less, balanced by BMS. I'm much more interested in how the cells behave when nearly discharged.

I'll be doing well to get one test from the dealer, so will attempt to comply with TickTock's recommendation. Not willing to hang around Palmdale, 84 miles from home, an extra five or six hours to get the 100% data. One of my goals is to get truck, trailer and LEAF home before dark...

Last year, they gave me a copy of the test sheet, with all 96 Voltages filled in... by hand. Hopefully there's a better way now.

-Karl

Good idea with the sign&tape. At my dealer, first thing they do is hook it up to the charger while waiting for the technician to get to it.

TickTock said:

Good idea with the sign&tape. At my dealer, first thing they do is hook it up to the charger while waiting for the technician to get to it.

Click to expand...

+1. That's apparently what they did to my car this time too. I didn't even think you could do the battery test on a fully charged car. That's certainly not what the service manual calls for.

A test at the low end only is useful, but not as much
as also doing the same test at the high end, to CHECK
the equalization. Otherwise, the high-end equalization
remains only an ASSUMPTION.

LEAF's on the trailer and tied down. I charged to 80% last night, preheated and used some heat this morning; it was around 34 degrees. Ran the car down to VLBW, which happened at about 8.5% charge and 55.6 miles. SOCmeter read 353.5 Volts when I stopped; it was up to 354.5 after an hour of recovery (car off). SOC had dropped to 8.1%, however. If necessary, I can burn off a little more charge tomorrow using the heater, but I don't expect to need to do so.

I taped the charge port closed, and I'll ask the dealer to do cell pair voltage report first. Then he can plug the car in and do all the firmware updates and battery usage report. I hate to leave the battery discharged this far overnight; it'll probably earn me a demerit, but it's all in the name of science. I think the dealer will cooperate when I explain that the battery appears to have lost 20% of its capacity.

It was great having one of garygid's SOC meters to use when I was discharging the battery. How else are you going to measure the Voltage? And it is nice to be able to read SOC even after the GOM reads '---'.

I'll report back whatever I learn...

-Karl

Good luck, Karl. Rooting for you.

I was able to get the 96 CVP's printed from screen prints on Wednesday. Note that Min/Max Cell Voltage is also available as a parameter, and unless you actually have a "bad" pair, having all 96 is not really useful. So the question is, what constitutes a "bad" pair ? I don't recall whether Ingineer or someone else gave some guidance on this or not, but for some reason more than a 40mV Min/Max difference seems to ring a bell with me. Anyone remember any better ?

My test was not done at VLBW. Total voltage was 373.64V with a Min/Max of 3870mV/3893mV respectively. SOC (per Consult III) was 43% and reporting Battery Level of 7920Wh. Battery Temp sensors were 63/63/61/61 F.

For comparison Gid Meter read: 373.5V, C43.1, 103 Gids (36.6%) just prior to moving the car to the service area.

I wish I had had enough time to run to VLBW, but due to time constraints (re-charging in order to get home) this was not possible.

Just knowing the min/max is some indication of balance, but
having the same cell-pair be low at the low test, and high in
the high-charge test, would likely mean that the pair is a
low-capacity pair. Just one test does not give enough data
to distinguish between low/high capacity and out-of-balance.

LEAFer said:

I don't recall whether Ingineer or someone else gave some guidance on this or not, but for some reason more than a 40mV Min/Max difference seems to ring a bell with me. Anyone remember any better ?

Click to expand...

TickTock had a 65mV min/max when he had his test done. I don't remember anyone else posting actuals. And I don't know where the break point is for acceptable vs. not.

It isn't a fixed value. A cellpair is replaced if it is lower than the Cell voltage loss judgment value which is computed as:
Cell voltage loss judgment value = Average cell voltage - (maximum cell voltage - average cell voltage) x 1.5
where Average cell voltage = total battery voltage ÷ 96

So a cellpair has to be more than 1.5X lower than the average than the max cellpair is higher for it to be replaced.

According to the manual you can take the readings anytime, and the maximum allowed variation is based on the current total pack voltage:


so at 373 volts you would have had to had a variation more than 140mV in order to be out of spec. 23mV is pretty well balanced.

Service manual EVB page 129.

palmermd said:

According to the manual you can take the readings anytime, and the maximum allowed variation is based on the current total pack voltage:

Code:

Total voltage (V)	        Approx. 360    Approx. 380    Approx. 390
Max - Min voltage (mV)	        230           140             135

so at 373 volts you would have had to had a variation more than 140mV in order to be out of spec. 23mV is pretty well balanced.

Service manual EVB page 129.

Click to expand...

Which version of the SM do you have? My copy (2011 April) doesn't have that text (just the formula I listed from EVB-67).

According to the cover page it is:

Edition: November 2010
Publication No. SM1E-1ZE0U0

downloaded Dec 21, 2010

Thanks, Michael, for that reference. I found that page EVB129, it's dated "Revision 2010 November ". And the EVB-67 "formula" TickTock appears to reference, in the 2010Nov Revision seems to be on page EVB-56. However, it uses a 2.0 factor, not 1.5, in that version.

TickTock said:

palmermd said:

According to the manual you can take the readings anytime, and the maximum allowed variation is based on the current total pack voltage:

Code:

Total voltage (V)	        Approx. 360    Approx. 380    Approx. 390
Max - Min voltage (mV)	        230           140             135

so at 373 volts you would have had to had a variation more than 140mV in order to be out of spec. 23mV is pretty well balanced.

Service manual EVB page 129.

Click to expand...

Which version of the SM do you have? My copy (2011 April) doesn't have that text (just the formula I listed from EVB-67).

Click to expand...