Cell voltages during discharge - Log data wanted

[MadTechNutter]

Thanks for your feedback guys.

I live on the coast, when I say hills, they are fairy short and not steep inclines, maybe 30 seconds 5˚.
Former owner lived in Brisbane, known for it's mud flats, no proper hills there either.
It is fairly hot there. Many days exceed 30˚C but from the near new looking condition this car is in, it even still smells like a new car, I am very sure it was kept in a garage mostly away from the heat of the very harsh sun there.

I have evidence that my car was in use for 4 years and 8 months. So at 76% SoH right now, in possibly 1 year it will not meet my daily requirements in EV mode, which is very disappointing.

I have been reading that Mitsubishi expects now 80,000km out of these batteries, no idea if this information is reliable.
But from what I have been researching, like 10% drop of SoH for the LEV50 after 700 FULL 100% cycles from an official technical report, this makes a lot of sense.

There is a theory that the LEV40 is a rated down LEV50 and should last longer but that does not make sense to me.
Both have 4.1V-2.75V for a full cycle. I would expect if you wanted to derate a larger capacity battery for longer service life you would reduce the
'usable' DoD. However the discharge graphs of the LEV40 seem to be kept a secret or I am just too stupid to find them.
Could somebody please post a link to the official report of the LEV40 with graphs?

 

[Darkflow]

Brisbane, known for it's mud flats, no proper hills there either.

You should see the area I regularly visit around Ascot / Hamilton - short and very steep.

 

[MadTechNutter]

Brisbane, known for it's mud flats, no proper hills there either.

You should see the area I regularly visit around Ascot / Hamilton - short and very steep.

30m above sea level ...yeah rrright
https://commons.wikimedia.org/wiki/Category:Ascot,_Queensland

Those ramps to get on and off the highway can be pretty steep too.

Very sad if the car is not designed for that.

 

[greendwarf]

But surely its a matter of physics - lifting a 2 ton car 30m uses more energy than pushing along on the flat and it will be the frequency that is more important than the occasional climb into mountains! And not many opportunities to climb even 30m in Holland other than in a multi-storey car park. :lol:

 

[ScottyDont]

Does anyone have some log data they would like to share? I found the quality was best with in town driving where it passed through different discharge rates often. My temperatures have been 15°C to 25°C so comparable temperatures would be most useful but I have the discharge curves for colder temperatures ready.

 

[ScottyDont]

So I have tried a couple of times to log data while charging and both my OBD adaptors disconnect, so no luck there.

EDITED with latest data

Elm70 mentioned voltage at rest, mine was sitting at 305.8V after 2 hours at 25.4%.

My battery SoH dropped recently, it is now 76.6%, and this is reflected in charging with only 7.8kWh at the wall from this 25.4% SoC.

I noticed a few people are getting their batteries down below 3V after these resets, that would be great data to compare as you would be able to see the knee point to scale correctly.

 

[obiuquido144]

I have a couple of DBCAM data points, this is with an immediately preceding Control Information and Estimated Capacity reset (i.e. 38SOH baseline). With just an Estimated Capacity reset (40Ah SOH) the SOC % values would be around 4-8pct points higher in this specific case.

T-12min: 18.4% 7Ah 292.3V 3.642Vmin 3.657Vmax 5kW load
T-8min: 16.1% 6.1Ah/1.8kWh 290V 3.591Vmin 3.616Vmax
T-5min: 12.6% 4.8Ah 280.4V 3.462 3.531 0.069Vdelta
T-2min: 10.3% 3.9Ah/1kWh 267.7V 3.240 3.391 0.151Vdelta
T-30sec: 2.937 3.276 0.339Vdelta
T-10sec: 8.7% 0.8kWh 250.4V 2.854 3.274 0.420Vdelta
charging started
T+5min: 3.438 3.534 0.096Vdelta

Notice the huge voltage drop on some cells during the last two minutes of discharge.

 

[ScottyDont]

Thanks for the data points obiuquido144, that drops off really quickly at the end. If it still thought it had 0.8kWh left when it was empty, looks like there's closer to 35Ah left in your battery. What value did the DBCAM calculate after it charged?

The voltages you have are higher than the LEV50 curve at 10% SoC so would be very interested if you have any PHEVWatchdog logs after getting this procedure done.

 

[obiuquido144]

34Ah natural degradation before everything.
DBCAM wrote 31.8Ah, but on first powerup after the DBCAM charge (after couple hours rest) the SOH was already bumped up to 32.3Ah. In the evening the same day after another charge the SOH jumped to 34.4Ah (this was after BMU Manual Write of around 60kAh and the usual ~1600 days, likely played a role).
Don't have logs yet, but this car now does around 30-35km typical EV range with 34soh. ICE with SOH reset to 38Ah first kicks in around 36km-40km at 3.7V/300V/31soc when driven gently around 60km/h.
I believe this car has some below-average cells whose voltage under load gets depressed quicker than the rest (the extreme of which was seen in the DBCAM numbers above), causing premature ICE kick ins e.g. when normally pulling off lights. And I believe this is also a reason why it doesn't respond much to Lindqvist/BMU resets - the voltage floor causes the ICE to kick in no matter what estimate SOC/SOH values are in the system.
If I get soc-voltage data from another DBCAM on this same car I will post them up for comparison.

 

[ScottyDont]

So I was able to log a fairly constant discharge using the heater and the fuel pump fuse removed again, but still not getting down to the voltages that some owners are seeing after resets. If anyone has any logs where the car goes into turtle mode or similar after a reset it would help target the low area of the discharge curve where it starts to drop off.

 

[ScottyDont]

This is what I have been trying to create, attached are three graphs using a compilation of multiple data logs from a PHEV, creating a full discharge curve that lines up reasonably well to an iMiev discharge.

I included both minimum voltage and maximum voltage curves to show the variation between cells in this pack, though the weakest cell will determine the usable pack capacity. In this car the variation between the best and worst cells is only 0.5Ah, which is ~1%. I also included a comparison to simulate the BMU at 38Ah to give an idea of the actual performance straight after one of the many resets. This car is still fairly healthy so it still has a reasonable buffer when the car states 30% SoC, but cars with worse SoH will drop off sooner which is where users are seeing turtle mode after a reset.

The important thing to note is the measured voltage is completely independent of the BMU reported SoH, so this may be a way to get a better idea of a cars actual SoH.

 

[Felix]

Here's a test I did a while back on a used cell, discharge at 9A 20.5°C from 4.2V down to 2.75V
4.1V equates to 95.7% SOC and 3.8V equates to 29.5% SOC
Cell capacity measured at 39.3Ah

 

[ScottyDont]

Here's a test I did a while back on a used cell, discharge at 9A 20.5°C from 4.2V down to 2.75V

Thanks for the data Felix! It matches the PHEV data much better than the other discharge graphs I was using. Here is a discharge from above with your discharge included.

 

[elm70]

Here's a test I did a while back on a used cell, discharge at 9A 20.5°C from 4.2V down to 2.75V
4.1V equates to 95.7% SOC and 3.8V equates to 29.5% SOC
Cell capacity measured at 39.3Ah

Very nice.

So this confirm that these cells are capable of 40Ah once charged up to 4.2v

I don't agree with 4.1V equates to 95.7% SOC, since the voltage depression while discharging is too much relevant at this level
It should have been discharged to 4A or even less .. 2A ... and it will be visible that from 4.2v to 4.1v there is much more then just 4,3% SOC in this range (possible even more then 5% ... still the 38Ah assumption from BMU is relative correct, but potentially a bit optimistic ... this might explain why BMU quickly downgrade the SOH when the battery is new)

Having not a correct number for SOC at 4.1V ... it is hard to verify the Mitsubishi assumptions that 30% SOC equate to around 3.81v (value that change with battery age, older the battery and higher voltage is required for 30% SOC for the BMU) ... as well .. 30% SOC compared to 4,1V at 100% SOC

It is interesting to see that from 4.1v down to 4.0v the line is more flat, then if get more steep from 4.0v to ~3.92v .. and then it get again more flat, but not as flat as initially ..

This explain while SOC after 2h rest ... inside then first 5/10km from fully charged most of the time cause the car to restart with an increase value of SOC per the BMU (extra EV range sine on the Dash)

Anyhow ... excellent data above .. thanks a lot Felix