BQ34Z100EVM: FC keeps turning on prematurely

I have lots of questions around the values for certain things and how they relate to this battery pack. For instance, should I make Qmax cell 0 5000 because that is the capacity for the entire pack? Then what about Cell Terminate Voltage? I set that to 2500 because it's the actual terminate voltage I want for the whole pack divided by the number of cells. Some of this was assumed by the max for those values. I've had a tough time finding values for my specific use case.

For instance:

Here it references the "Charging Voltage", but there is no variable to input the charging voltage.

I've been trying to get the chemid and perform a learning cycle, but I can not charge the battery all the way because of this problem. It just stops charging at 13344 mV.

So is that charge voltage setting correct for my battery?

What is TRM? There is no TRM listed in the data.

OK it's the technical reference manual. I have changed many of the values such as those above. I've set the design capacity, design voltage, design energy, Cell Charge Voltages, Taper Current, Number of series cell, Cell Terminate Voltage, Dsg Current Threshold, Chg Current Threshold, Quit Current,  and Voltage Divider. Can you look at these values and see if they seem sane? What else am I missing?

Battery stats:

4s LiFePo4

The other thing of note is that I am on the R2 firmware. Is this a mistake should I have stayed on G1? Is Design Capacity still for the entire pack on R2?

The nominal capacity is set for 5000 mAh and the nominal capacity of the pack listed above is  5 Ah. Is my understanding here incorrect somehow?

The reason I ask is that you called out the design capacity and implied that everything other than that  was reasonable. I do not understand why the design capacity value would not be reasonable. Or were you just pointing that out because it's a common stumbling block? I just want to be specific here because I'm don't want to spend another full two days trying to accomplish a learning cycle when I could still be misconfigured.

But isn't that what I have done?

I posted my configuration 4 messages ago and design capacity is the first value listed, that is why I thought you were reacting to the specific configuration.

https://www.ti.com/lit/an/slyt402/slyt402.pdf

Are these steps necessary for LiFePO4 battery packs? Do I need to do this before the learning cycle?

Also, I have the Design Voltage at 3200 mV. This is the voltage for only a single cell, which seemed to be correct since the field has a max of 5000. Is this actually correct? Do I need to change the voltage scale to accommodate the Design Voltage of the entire pack? This is not int he documentation for the chip or the firmware from what I can tell. This is the only documentation of the field I have found:

Another thing. In this datasheet:
https://www.ti.com/lit/ug/sluuco5a/sluuco5a.pdf

It says "Ensure that the battery is at a low SOC and relaxed when running IT_ENABLE. This is typically not a problem, but it is important." However, everything I've seen indicates that the battery should be full when starting the learning process. Does this just mean that it should be rested after charging and before enabling IT?

That is the document I have been using along with the firmware datasheet I linked above. To be clear, the learning cycle is for achieving the initial qmax update as well as for populating the Ra table, but this document is for subsequent qmax updates after the initial learning? Is that correct? It just seems counter intuitive to me that all but the initial qmax update would benefit from this process. 

Additionally, is this process required to get reliable qmax updates after the learning cycle?

This comes from this doc:

https://www.ti.com/lit/ug/sluuco5a/sluuco5a.pdf?ts=1750324220032&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FBQ34Z100-R2

In the section about things to look out for during the learning cycle:

"Ensure charger cutoff upon charge completion: Most users do not have a battery cycling automation setup, so a bench power supply is used to change a battery overnight. This is not recommended. The system does follow a CC/CV profile, but there is no cutoff. Therefore, when the gauge recognizes a full charge and tries to take an OCV measurement, it actually measures the supply voltage and disrupts the system."

So what does this mean exactly? I am using an appropriately rated power supply, not a bench power supply. Does this mean that I need to disconnect it at a certain time? Is it relative to when FC is set? If so, how long after FC (or equivalent) is set do I have to disconnect it?

OK, but does this mean that I need to disconnect the charger I have after the battery has charged and FC is set? Or can I leave it indefinitely then disconnect and rest it? If not, how long exactly do I have to disconnect it?

So I performed another learning cycle, and I noticed that my qmax updated but my Ra tables did not. Maybe this is the main issue I'm having now. I don't think they budged at all. I am going to include the entire data memory, but here is just the Ra table:

Ra Table R_a0 R_a0 Flag 55
Ra Table R_a0 R_a0 0 0
Ra Table R_a0 R_a0 1 58
Ra Table R_a0 R_a0 2 93
Ra Table R_a0 R_a0 3 100
Ra Table R_a0 R_a0 4 63
Ra Table R_a0 R_a0 5 82
Ra Table R_a0 R_a0 6 93
Ra Table R_a0 R_a0 7 100
Ra Table R_a0 R_a0 8 96
Ra Table R_a0 R_a0 9 85
Ra Table R_a0 R_a0 10 80
Ra Table R_a0 R_a0 11 111
Ra Table R_a0 R_a0 12 174
Ra Table R_a0 R_a0 13 281
Ra Table R_a0 R_a0 14 0
Ra Table R_a0x R_a0x Flag 0
Ra Table R_a0x R_a0x 0 0
Ra Table R_a0x R_a0x 1 58
Ra Table R_a0x R_a0x 2 93
Ra Table R_a0x R_a0x 3 100
Ra Table R_a0x R_a0x 4 63
Ra Table R_a0x R_a0x 5 82
Ra Table R_a0x R_a0x 6 93
Ra Table R_a0x R_a0x 7 100
Ra Table R_a0x R_a0x 8 96
Ra Table R_a0x R_a0x 9 85
Ra Table R_a0x R_a0x 10 80
Ra Table R_a0x R_a0x 11 111
Ra Table R_a0x R_a0x 12 174
Ra Table R_a0x R_a0x 13 281
Ra Table R_a0x R_a0x 14 0

Do these values actually make sense? My understanding is that the values should be more of a smooth curve and not oscillating like this. Also, if they were changing then wouldn't the values in the R_a0 table and R_a0x table be different?

datamemory3.gg.csv

Reprogramming to the same chemid, 460, does change the values a lot. It does not totally consistently go up, but it is a lot more smooth. This indicates to me that the values were changed to these over the course of the learning.

The problem I have currently is this:

The SOC goes from about 94 directly to 0.

I just discharged it again with the default 460 chemistry (for which the Ra tables are very different from the learned one above) and it went from 96 SOC directoy to 0. It looked normal between 99-96. Before I did the learning entirely the SOC seemed to decrease at a normal rate, but it would get to 0 earlier than I expected. It was a very linear dropoff though.

I suspect you are correct because now it appears to actually work while charging up. It even seems accurate! However, during the discharge the SOC drops to 0 prematurely as before. What could be causing that? Here are the files:

5074.test2.log

This is the log of me charging, discharging, then starting charging again immediately after the latest learning cycle was completed. You can see while charging it works fine, then during discharge the SOC drops to 0 prematurely.

datamemory4.gg.csv

For completeness I took the datamemory before and after this log. The above is before. Below is after:

datamemory5.gg.csv

Here are the critical values as they appear in the data memory right now:

These are the parameters I am using currently. Here are my questions:

1. Term Voltage does not appear to be settable in my firmware. There are several references to it in the datasheet, but it is not a settable value from what I can see, and it is only referenced in relation to other values:

https://www.ti.com/lit/ug/sluuco5a/sluuco5a.pdf

Is this supposed to be settable somewhere? If so, where?

2. Design voltage in the successful learning cycle is defined as such: Design Voltage = (number of series cells) x average cell voltage. However, since the max for the R2 firmware is 5000 mV, I have been treating it as the voltage of a single cell, 3200 mV. Is this correct?

3. I believe I originally left Charge term taper current at the default, 100 mA, but it got changed as part of the learning cycle to 250 mA. Does that sound plausible? Maybe I should set it to 70 as suggested in the successful learning cycle doc? Here is the charge I'm using: https://www.batteryspace.com/Smart-Charger-0.7-A-for-12.8V-4cells-LiFePO4-Battery-Pack-100-240VAC.aspx?srsltid=AfmBOorxD8UvgMaKA5Wt5FOM_TfxdbU-MWyWgdoQba2uhvdkLFltSwWD It doesn't list a termination current, but maybe I should call it like 100 mA?

4. Dsg Current Threshold is default (same for Chg Current Threshold and Quit Current) because the docs say that works for most applications. However, the learning cycle doc states: . A value below C/10 is generally reasonable. This would be 500 mA for my pack which seems very high. It doesn't seem to have any trouble detecting charge and discharge though, so maybe that's fine.

5. Quit current is default as well, and it seeeems to be resting reasonably, but maybe you can glean more from that than I can:

6. I set the voltage divider register to 13952 mV because that seemed reasonable based on the resistors on the evaluation module: https://www.ti.com/tool/BQ34Z100EVM Is this correct?

7. I am using chemistry 460 currently, which seems like it should be correct mostly. However, I have not been able to get a value out of your chemistry identifier, as is being tracked in this other thread: https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/1522360/bq34z100evm-gpc-tool-always-outputs-0/5857417#5857417

Hello Matthew,

We have received your update and a response is being looked into.

Thank you,
Alan

Thank you very much! I have got a report from the chemistry selection tool now, but please see the other thread for the update there, since I hit another wall in that direction.