BQ40Z60: CompRes (Compensated Resistance)

What else can caused compensated resistance to be inaccurate, given the battery is only used at room temperature?

e.g. We have a battery pack which experienced continuous increase in compres (up to 2000 mOhm), but once the pack is allowed to REST, the compress fell back down again (Under 200 mOhm).

Hello Asfan,

I would need to see the logged data and the .gg file to analyze what the gauge is reporting. It sounds like the Ra table was increasing during your tests if the temperature was always at room temp. You would need to pull a gg file from before and after to compare Ra tables. If the Ra table wasn't increasing then there may have been a large degree of accumulated error before we took the OCV to correct for it and that was making the gauge use the wrong Ra grid point for CompRes.

Sincerely,

Wyatt Keller

Here's the data including the log and data memory for the cycle before and after Qmax update cycle. This is the case when we tried running the Qmax update cycle and found that the CompRes was reduced significantly afterwards. Can you please explain what is causing such a massive drop in CompRes? This pack has always been running in room temperature so we don't think it has anything to do with temperature. Please ignore CompRes cell 4 as there is issue with logging.

7026.TI2.zip

Hello Asfan,

It looks like it may just be an issue with your Ra table, the .gg file pulled after seems to have a lot lower values than the ones before. What is your normal system load? Impedance track will not be able to accurate update the Ra table for pulsed loads and it can lead to corruption of the Ra table over time.

Sincerely,

Wyatt Keller

Hello Wyatt,

I do apologize for late respond.

The latest update of our discussion is, You have another understanding that the Ra Table issues are because of fluctuate or dynamic load in our application, hence the BMS is quiet difficult to update the Ra accurately. For that topic I'll respond that on the other thread.

I still have some questions regarding CompRes because I have some cases where the relationship between R_a and CompRes is completely different. From my understanding, given the condition where temperature is similar, higher R_a should result in higher CompRes as well. However, this isn’t what I’m getting in my data. Here are the cases I’ve found:
1. Power pack A : The highest R_a14 is approx. 2000 mOhm with maximum CompRes nearing 2000 mOhm as well.
2. Power pack B : The highest R_a14 is approx. 6000 mOhm with maximum CompRes only around 300 mOhm.
Comparing those 2 packs, it seems like the pack with lower R_a ended up with higher CompRes. Both of them are running in room temperature, so temperature shouldn’t be the issue here.

I understand that previously you’ve explained that CompRes calculation is proprietary and you can’t disclose the details. However, can you please help me understand if what I’m seeing in these cases is normal or is there any error? If this condition should not have happened, what could possibly cause the error? We need to be able to prevent such error from happening again in the future considering that CompRes is very important in gauging.

Thankyou,
Asfan.

CompRes & R_a comparison.zip

Hello Asfan,

The compres should be the temperature adjusted resistance only, from what I see in the logs I don't think it is a normal readout since cell 4 seems to have a much lower value. Also it looks like you are not cycling in room temp, your temperature reading is almost 40degC. It also looks like your grid points are not incrementing while discharging, something seems to be incorrect with the tracking of the present DOD. This could be due to the pulsed load as well. When you cycle with constant current do you see the same behavior?

Can you share what chem ID you are using? I'd like to check it to make sure the temperature adjustments look okay.

Sincerely,

Wyatt Keller

Hello Wyatt,

Wyatt Keller said:

The compres should be the temperature adjusted resistance only

Do you mean that Compres is temperature adjusted from Ra value without considering other parameters?

Wyatt Keller said:

from what I see in the logs I don't think it is a normal readout since cell 4 seems to have a much lower value

I do apologize, I forgot to mention that please ignore the Compres 4, it has some error on logging. We can take a look at compres 1 - 3 (Except 4).

Wyatt Keller said:

This could be due to the pulsed load as well.

Both of the packs are running with the same load profile (Fluctuate load). So, I don't think this what causing the ratio difference between compres and Ra value. 

Wyatt Keller said:

Can you share what chem ID you are using?

The chemical ID is 2141 (Sony/Murata VTC6).

Thankyou,
Asfan.

Hello Asfan,

Yes only the Ra table that is temperature compensated. The the temperatures you are testing at (high temp) should make the temperature compensation lower than the Ra table.

Okay no problem.

The load could cause a difference since it is unpredictable how the gauge will learn the Ra table under a pulsed load and it could increase it much higher or lower than reality. We would need to see the .gg file before and after, and ideally during the testing. I'm also not sure why the grid is locked at 1 or 2, this will also affect the compres since the grid is always pointed to the same location for Ra lookup.

Sincerely,

Wyatt Keller

Hello Wyatt,

Okay, thankyou for the explanation. 

Wyatt Keller said:

We would need to see the .gg file before and after

I've attached the "before" .gg file Compres_Before_condition.zip

Another question is,

Wyatt Keller said:

I'm also not sure why the grid is locked at 1 or 2, this will also affect the compres since the grid is always pointed to the same location for Ra lookup.

1. From your previous explanation, If the cell grid isn't updated, let's say stuck at grid 1 and 2. In the condition where the pack is discharged near to empty (where the grid should be a lot higher), does this means the CompRes will still be calculated based on R_a1 and R_a2?

2. I've learned a lot from this SLUA, the way BMS predict the Remaining capacity. The document explains very well about the formula to predict DOD final, in order to achieve the remaining capacity value. The following formula is: 

• V(DODx,T) = OCV(DODx,T)+ I×R(DODx,T)

You have said earlier on our discussion, 

Wyatt Keller said:

The gauge always uses the temperature compensated resistance for gauging

Does this means that the R(DODx,T) in the aforementioned formula is actually the CompRes?

Thankyou, 
Asfan

Hello Asfan,

It looks like there's no change between the two .gg files. This means the gauge may not be learning the Ra table during all discharges due to the negative scale factor flag which is common for pulsed loads.

Yes if the gauge is not tracking the discharge capacity properly and the grid is in the wrong location the compres will be calculated based on where the gauge thinks the present DOD is.

Yes that is a very helpful app note for impedance track, yes R(DODx,T) would essentially be the compres.

Sincerely,

Wyatt Keller

Hello Wyatt,

Thankyou for the respond. 

Wyatt Keller said:

Yes if the gauge is not tracking the discharge capacity properly and the grid is in the wrong location the compres will be calculated based on where the gauge thinks the present DOD is.

From your previous explanation, the compres value will be calculated based on the last updated cell grid. It means, when the cell grid had gotten stuck (e.g. last updated grid point 2, when it's supposed to be at much higher grid based on its RSOC), the compres will be calculated based on R_a value on grid point 2. Then, is the real-time (present) DOD where the gauge think it is currently at, would also be stuck at grid point 2? This is quite confusing as I thought the present DOD is calculated based on DOD0 + DOD0 Passed Q.

Hello Asfan,

Yes the grid being stuck would seem to indicate the gauge may not be in discharge mode since otherwise it should be updating the present DOD based on the equation.

Sincerely,

Wyatt Keller

Hello Wyatt,

Wyatt Keller said:

the gauge may not be in discharge mode

So do you mean that even if the current is negative, which indicate we're currently discharging, but with the grid stuck at certain point, It means that the gauge would assume that we're not in discharge mode?

If so, then how could the voltage simulation on Remaining capacity calculation possible? Since this simulation would require the present DOD to be constantly moving. 

Thankyou,
Asfan.

Hello Asfan,

Yes it's really strange, I just mentioned the discharge mode because that's the only thing I can think of that would stop the grid from moving.

Are the logs pulled with BQStudio? I see the names are similar but the Pack Grid is not correct, just want to make sure the data was captured correctly since I haven't really seen this behavior before.

Sincerely,

Wyatt Keller