[Solved: Piev's Strange New Pack] Voltage vs SOC graph or table ?

[kiev]

i was searching for a graph, chart or table of this data--seems like someone posted this before but i can't seem to find it using the search. Just a simple table of battery pack voltage at 10% SOC increments would be helpful.

i know that 100% SOC is about 360 Volts but didn't have a feel for the rest to know if there is a problem here in piev's #3 car. [Edit: this is a problem] Here's a Canion shot, Mileage and RR are in km.

and from the MUT clone

Seems like capacity and SOC should be higher for this pack voltage? And this is a new replacement pack that is supposedly only about 17 months old.

SOLVED: This is a new pack but it wasn't acting as such. i sent Paul the FSM instructions for replacing a pack, which includes a step to "reset" the capacity. It appears that this step had not been completed properly and was causing the low capacity reading and the lack of ability to use the entire voltage range of the pack. He had to take it to the dealer service dept twice, and show them a printout of the procedure, to finally get the "reset" completed. Now the capacity reading is 45.6 Ahr and he is getting the expected range of a full pack! Details in post#13.

 

[DBMandrake]

Not wanting to rain on your idea, but the minimum cell voltage at low SoC seems to change as the cells age and lose Ah capacity.

For example here is my car at 0% and 10% SoC back in April at 41k miles:

Now look at 10% SoC in October, after there has been significant Ah capacity loss - the voltage at 10% is now almost the same as the voltage was at 0% in April:

So it looks like when the pack is new the BMS doesn't push the voltage of the cells very low (still 3.7 volts for 0% SoC, even at 38Ah) but as the cells degrade and lose capacity it is willing to push the cutoff voltage lower.

Having said that, here are some figure from my pack, for what it's worth:

0% SoC - pack 320v, cell min 3.560v, cell avg 3.640v, cell max 3.680v
10% SoC - pack 328v, cell min 3.695v, cell avg 3.729v, cell max 3.745v
33% SoC - pack 338v, cell min 3.830v, cell avg 3.845v, cell max 3.850v
35.5% SoC - pack 339v, cell min 3.835v, cell avg 3.848v, cell max 3.855v
62.5% SoC - pack 348v, cell min 3.955v, cell avg 3.957v, cell max 3.960v
71% SoC - pack 351v, cell min 3.990v, cell avg 3.995v, cell max 4.000v
100% SoC - 360v, all cells 4.100v

These figures were probably only valid at the capacity the pack had at the time, of around 38Ah.

What is the problem you think you're seeing ? If you're looking for cell imbalances you won't see them at 67% SoC unless the imbalance is really bad. You need to charge to 100% letting balancing complete then discharge to about 10-20% SoC to look at cell capacity balance.

 

[CZeroOwner]

Here is a graph of Voltage as a function of SoC and current at 25oC. It’s based on lab data for a new LEV50 cell shown in the technical report “Development of a Large Sized Lithium Ion Cell “LEV50” and its Battery Module “LEV50-4” for Electric Vehicle” by Kitano et.al.

Let me know if this fits the bill.

 

[jray3]

So it looks like when the pack is new the BMS doesn't push the voltage of the cells very low (still 3.7 volts for 0% SoC, even at 38Ah) but as the cells degrade and lose capacity it is willing to push the cutoff voltage lower.

Thank you for that observation! It is understandable that the designers would try to preserve range by 'squeezing the cells harder' near the end of their service life, and would also help explain my sense that degradation is accelerating.
I exacerbated the situation by replacing some bald Enasaves with new Continentals up front and a pair of bargain SunFull 688 tires at the end of this summer, and am about to make it worse with the transition to winter tires by Christmas. :roll:
Morning RR is usually 45 miles, and our morning temperatures have been around 45F ...
I'm looking forward to trying out my salvaged battery pack but will stay on course for a full 100k miles of service on the original, with less than 4k to go!

 

[kiev]

Some good data and graphs, which just points that something is not right with this pack's calculations. When all the cells average 4.005, pack at 352, then the SOC should be higher than 67% and RR much greater than 16 miles.

The car shut down completely showing 0 bars and the pack at 330V , that's 3.75 volts per cell, which is higher than the nominal cell voltage of 3.7. Why is the ECU preventing the use of the remaining cell capacity?

There is a ECU reset procedure for the capacity value associated with replacing a pack, but we have no visibility into what that is or what it does.

 

[CZeroOwner]

Here is some data from my car. All of these measurements were recorded when the load on the battery was less than 1 amp. The spread is probably due to varying temperatures. Apparently a combination of 352 volts and 67 % SoC is possible. However 330 volts should give much more than 10% equal to 0 bars.
The iCarsoft meter shows a full battery capacity of 21.4 and remaining capacity of 15.2. That’s 71% which is kind of close to 67%.
15.2 Ah is about equal to 5000 Wh. If we subtract 10% for turtle power that is 4500 Wh. With a rest range of 27 km the car believes that the present consumption is about 167 Wh/km. That’s high. My car gets an average of 130 Wh/km but it can easily get up to more than 167 with the heater on.

 

[DBMandrake]

Here is some data from my car. All of these measurements were recorded when the load on the battery was less than 1 amp. The spread is probably due to varying temperatures. Apparently a combination of 352 volts and 67 % SoC is possible.

Which perfectly demonstrates why voltage estimation of SoC of Lithium Ion batteries in the "middle" SoC range is not meaningful or workable.

Notice how it converges at both ends though.

However 330 volts should give much more than 10% equal to 0 bars.

When I tested my car back in February, it reached a reported SoC of 0% in Canion at a voltage of 320 volts, at which point it shut down. 330 volts is around the point where turtle mode comes on or "no" range remaining based on extrapolation of the range remaining figure after it went to dashes...

The iCarsoft meter shows a full battery capacity of 21.4 and remaining capacity of 15.2. That’s 71% which is kind of close to 67%.
15.2 Ah is about equal to 5000 Wh. If we subtract 10% for turtle power that is 4500 Wh. With a rest range of 27 km the car believes that the present consumption is about 167 Wh/km. That’s high. My car gets an average of 130 Wh/km but it can easily get up to more than 167 with the heater on.

Unless your car has a REALLY sick battery, I don't know what your figures of 21.4 and 15.2 are. They're certainly not Ah, nor kWh. Are you sure iCarsoft is giving meaningful figures here ? I think my battery is in relatively poor shape and it still has an Ah capacity of 35.5, so 21.4 is absolutely dreadful - a SoH of less than 50% of new.

 

[kiev]

Thanks for sharing your graph, that is a great set of data.

It's difficult to know what is the issue with Piev's pack since it is only showing a full capacity of 21.4 Ahr, yet it was replaced under the warranty about 19 months ago.

Assuming that a new pack is good for 40 Ahr by allowing a 20% reduction for the turtle and an upper voltage buffer, then this pack is down almost 50% after less than 2 years. Doesn't make any sense how it could be so bad.

The car was in the service shop for several months due to a defective charger--could that somehow affect capacity?

 

[DBMandrake]

It's difficult to know what is the issue with Piev's pack since it is only showing a full capacity of 21.4 Ahr, yet it was replaced under the warranty about 19 months ago.

Canion also agrees with the 21.4Ah figure ? Is the range also severely impacted ?

Assuming that a new pack is good for 40 Ahr by allowing a 20% reduction for the turtle and an upper voltage buffer, then this pack is down almost 50% after less than 2 years. Doesn't make any sense how it could be so bad.

The Ah figure quoted in diagnostic tools querying the BMS includes the capacity in turtle mode. So is the full 100% to 0% capacity. A new battery should be >46Ah - my car at 6 years old and 28k miles was still 39.9Ah but has dropped to 35.5Ah since then... I have not heard of any of these cars with a reported capacity less than 33Ah yet, so 21.4Ah is very anomalous.

The car was in the service shop for several months due to a defective charger--could that somehow affect capacity?

Hard to know. If I was trying to diagnose the car I would monitor all cell voltages with Canion and first charge it to 100% and let it balance, ensuring that all cells are +/- 5mV when fully charged.

Then discharge the battery down to a reported 20% SoC on Canion and compare the cell voltages for outliers. It would only take one sick cell to knock the usable capacity down to 21.4Ah, and it is exceptionally unlikely that this degree of capacity loss would be uniform across all cells.

It's got to be a faulty cell unless there is something wrong with the BMS system.

Is there another thread already discussion Piev's car's battery ?

 

[CZeroOwner]

If the charger charged the battery to more than 4,1 volts per cell then that would speed up battery aging.

 

[DBMandrake]

If the charger charged the battery to more than 4,1 volts per cell then that would speed up battery aging.

Do we have any reason to believe this happened though ?

Enough over charging to cause that amount of degradation in such a short time is more likely to result in the cells catching on fire... :twisted:

 

[CZeroOwner]

It is unlikely.

 

[kiev]

i updated the first post to reveal that this issue has been solved. The pack was replaced but the "reset" procedure was not completed properly. He took it to the local dealer and now it's acting like a spankin' brand new pack.

Here is link to the pack removal and installation instructions, notice the second bullet of the caution banner:

"After replacing the main battery, use the M.U.T.-III to reset the main battery capacity estimation information and control information. (Refer to ->)"

and Here is link to the MUTIII "reset capacity" procedure

snapshot of "reset"

i think the capacity will never show greater than 48 Ahr even if you were to build a pack with 100 Ahr cells, at least until someone figures out how to change the "reset" value.

 

[CZeroOwner]

good news!

 

[DBMandrake]

i think the capacity will never show greater than 48 Ahr even if you were to build a pack with 100 Ahr cells, at least until someone figures out how to change the "reset" value.

Old post I know but linked from a new thread.

The "reset" (battery replacement) procedure does indeed just reset the capacity to a factory default of about 45.8Ah and wouldn't "see" a larger battery. But there is also the battery calibration process, at least in the PSA versions of the car using Diagbox.

This actually measures the capacity of the pack by discharging the pack until the lowest cell is about 3.75v then charging it up again and integrating the total amount of current over time. Afterwards it adjusts the Ah figure and also saves a battery calibration report to the ECU's memory (which is persistent, unless you do the battery replacement procedure, which clears it) which lists both the lowest and highest Ah capacity of the cells in the pack, which is nice.

This battery calibration process almost certainly would be able to measure a battery with higher than original capacity and utilise it - the only question is how much higher would it allow before considering that the figure was unrealistically high. In other words would it be happy to measure a 100Ah battery and adjust the figure accordingly or would it be thinking "hang on a minute.....". :twisted:

The only way to know for sure short of disassembling the BMS's firmware would be to try it!

By the way the battery calibration process is supposed to be performed by the dealer during every regular major service, but on my car that was dealer serviced for the 6 years before I bought it the report had never been run! (All previous results fields were zero, until after I did my own battery calibration)

As the battery calibration process takes nearly a whole day to run and ties up their diagnostic tool for a good part of that my assumption is that the dealers are just too lazy to spend the time doing it for no return even though the service schedule says to do it.