BQ34Z100-G1: Getting CHEMID for bqstudio

David Pfaltzgraff

Part Number: BQ34Z100-G1
Other Parts Discussed in Thread: GPCCHEM, BQSTUDIO, , BQ40Z50-R2

I am picking this up from about a year ago so if the procedure has changed, please let me know...

I obtained a log of the discharge for my battery which is four NiMH cells in series. I uploaded this and the GPC report merely said:

"Please contact your TI representative to send the cell to Dallas for characterization."

OK, what am I supposed to do now that there are no local reps?

over 7 years ago

0 Bryan Kahler over 7 years ago

TI__Mastermind 25955 points

Hi David,

You came to the right place! Please send me your logs so I can run the test internally.

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

Unfortunately, I am not in the office right now and it may be a couple days until I can get the original. In the meantime, I have some questions. Would it be better to wait a few more days and get a more current set of data? Since I did the original run, I have built a better load and believe I can get more accurate data. The other question is: Should this specifically be a new set of cells, or doesn't that matter? The cells I have available have been in use during this year. To get a new set might be a couple more days.

Thanks,
Dave

0 Bryan Kahler over 7 years ago in reply to David Pfaltzgraff

TI__Mastermind 25955 points

Hi David,

It would definitely be best to wait a few days and get a current set of data with brand new cells.

The older cells will have aged and will not give nearly as good of a match. We can start configuring the gauge with the older cells, but may have to redo all the work as the newer cells will have a different Qmax.

I'm glad to hear that you have a build a better load - a constant current load is important in determining the proper ChemID when using the GPCCHEM tool.

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

Finally back in the office with new cells, etc. The problem now is, how do I attach the log file?

I sent it off to the chem id program and that came back with essentially nothing. The GPC_report file had just the number '0' and the MiscInfo only repeated the processing type. I figure I must have done something basic wrong.

Thanks.

0 Bryan Kahler over 7 years ago in reply to David Pfaltzgraff

TI__Mastermind 25955 points

Hi David,

To send the information to the tool, a config.txt file and log file need to be zipped. Here is a link to the instructions: www.ti.com/.../slva725.pdf

The GPCPacker tool in bqStudio is not currently associated with the bq34z100-G1, however, if you (with device disconnected) open up bqStudio and select the most recent bq40z50-R2 gauage, you will be able to select the GPC Packager tool. You may use this tool to create the package to send the GPCCHEM tool.

With the package created, please upload the package at the upload link found here: http://www.ti.com/tool/GPCCHEM

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

Apparently you didn't understand my message... I did all that and got back nothing. I'd like to send the file to someone for review. I THINK I'm doing everything right, but obviously the tool doesn't agree.

Thanks, Dave

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

First a comment: The reason that my original file didn't work was a typo in the configuration file. It would be helpful of the GPCHEM program would flag when something it expects is not there.

I followed your procedure using the bqStudio as well as finally found my typo and ran both sets of data through GPCHEM. The results are the same - and w-a-a-y wrong. It suggests a LiCoO2 chemistry with 100 to 200% error!

These cells are NiMH and I just want to be able to use the fuel gauge functionality. If I knew more about the bq34z100-g1's workings, I'd try fudging it as right now I feel as though I'm moving backwards.

All help appreciated.
Thanks,
Dave

0 Bryan Kahler over 7 years ago in reply to David Pfaltzgraff

TI__Mastermind 25955 points

Hi David,

I appreciate the feedback with respect to the GPCCHEM tool, that would be a very useful feature.

The GPCPACKER tool was recommended because it seemed there was an error with your config.txt file. This tool would have created the file for you. I glad to hear that you were able to solve the formatting issue.

With respect to the results, LiCo02 as the result is a bit of an oddity, especially for a NiMH cell. With your 4S configuration, was that detail included in the config.txt or was it set to 1? If it was set to a 1S configuration, this could explain the poor match. If so, please set the value to 4 and rerun with the GPCCHEM tool.

To push through this hump and get your device operational ASAP, please send me your learning logs and gg.csv file so I can analyze them. The battery datasheet would also be helpful.

Also to streamline this process and allow me to deliver results sooner, please let me know:

Is this a custom board or the EVM?
If EVM:
- Where are your jumpers set on J5 and J2?
- J5: Shorting Pins ___ and ___, and pins ___ and ___
- J2: Shorting pins ___ and ___
If Custom Board:
- Voltage Divider Values on BAT:
- High Resistor: __________ Ohm
- Low Resistor: __________ Ohm
- Current Sense Resistor: _________ mOhm

Battery Cell Information:
- Chemistry: NiMH
- Max Voltage: ______ V
- Nominal Voltage: ______ V
- Capacity: _____ mAH

Battery Stack/Application Information:
- Configuration: __S__P
- Max Load: ________ mA
- Nominal Load : ________ mA
- Normal Operating Temperature: _______ degC
- Operational Temperature Range: ______ degC through _____ degC
- External Thermistor on Pack: Y / N
- Pack side or System side Gauge: Pack / System

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

I'm not in the office until Monday (the 29th), but I'll answer as much as I can.

First, I have not been able to find out how to attach a file. I have found "inster" a file, but to me that's different.

Second, I had so many problems using bqStudio with bad communications and bogus numbers, that I gave up trying to use it to capture the data and wrote my own application on our board that gathered the necessary data. So, I don't have the gg.csv and log files that you expect. I can send the csv and config files that I gathered.

Yes, I did verify that the config file specified 4 cells in series.

To answer your questions:

- Voltage Divider Values on BAT:
- High Resistor: 100K Ohm
- Low Resistor: 165K Ohm
- Current Sense Resistor: 100 mOhm

Battery Cell Information:
- Chemistry: NiMH
- Max Voltage: 1.4 V
- Nominal Voltage: 1.2 V
- Capacity: 3200 mAH

Specifications can be found at: data.energizer.com/.../nh15-2300.pdf

Battery Stack/Application Information:
- Configuration: 4S1P
- Max Load: 20 mA
- Nominal Load : 16 mA
- Normal Operating Temperature: 22 degC
- Operational Temperature Range: 0 degC through 30 degC
- External Thermistor on Pack: N
- Pack side or System side Gauge: System

If you need more, just let me know.

Thanks, Dave

0 Bryan Kahler over 7 years ago in reply to David Pfaltzgraff

TI__Mastermind 25955 points

Hi David,

Thank you for the details above. With bqStudio, what communication adapter were you using? If it was the EV2300, please make sure it is not version 3.1C. This version of the EV2300 firmware is known to cause communication issues and must be updated. Also, the EV3200 includes pull up resistors that may interfere with the pull up resistors on your board.

bqStudio will help simplify the process of support so not everything must be converted and we are working with a 'known good' setup. I understand this hasn't been the case thus far and do apologize for that. Also, bqStudio is required to program the chemID (We're not quite there yet).

In that vein, let's try to troubleshoot bqStudio. To attach files, please click on the Insert Code, Attach Files and more... link below your response box.

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Looking at the time stamps on your post, I'm presuming that you're somewhere west of the Eastern time zone. I won't be in the office for a bit, but I can answer some of our questions and make some progress.

First, I am using the EV2300, so... (The specific model is the gray case with the correct labeling of the HDQ lines per the errata.)

a. How can I verify and update the firmware? I didn't see anything in the User's Guide, etc.

b. Looking at the schematic provided, I didn't find any pull-ups on the SDA and SCL lines. Maybe it's a newer schematic... So, how can I identify them? If necessary, I do have the tools to remove them. At this point, I don't feel that this has been the issue, but am willing to try.

Just a comment: I have been able to program the part with bqStudio. At least it went through the calibration process and it was necessary to restore the original firmware to get back to what I had before. The hangup has been the CHEMID and getting to the point of training.

As far as attaching files... I have clicked on the link, but do not see any specific path onthe subsequent screen. I'm probably missing something.

Thanks

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

More information... I pulled up bqStudio and saw that the version for the EV2300 is shown. I do have 3.1c. So, I will need instructions on how to update that.

Thanks,
Dave

0 Bryan Kahler over 7 years ago in reply to David Pfaltzgraff

TI__Mastermind 25955 points

Hi David,

The updater for the EV2300 v3.1C firmware can be found here: e2e.ti.com/.../3225.EV2300FWUpdaterUnprogDev.zip

The instructions are included in the archive.

a. When you plug the EV2300 in to the computer and open bqStudio, it will display the firmware version on top of the Device in the left pane.

b. David, I apologize, you're absolutely right. SMBus on the EV2300 is pulled up, but there are no pull ups on the I2C and HDQ lines.

I have sent you a friend request so you can e-mail the files.

Thanks!

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

In looking at the procedure, it appears that I need two EV2300's to perform the update. Is this correct?

In my case I have only one. What to do now?

Thanks,
Dave

0 Bryan Kahler over 7 years ago in reply to David Pfaltzgraff

TI__Mastermind 25955 points

Hi David,

I followed up with an e-mail. Thanks!

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

OK. Got the second EV2300 and updated the first. It is now version 3.1m and things are certainly going better. I have been able to calibrate the voltage and current. (May need to refine that a bit...) However, whenever I try to calibrate the temperature to 22C, internal. It continues to report back -39.2C.

For my application, I am not concerned with this as we use a different sensor for temperature. But, bqStudio is going to report/log whatever it reads. So, what am I doing wrong?

One step at a time...

Later,
Dave

0 David Pfaltzgraff over 7 years ago in reply to David Pfaltzgraff

Expert 2445 points

Hi Bryan,

OK. Figured out how to get the TEMPS bit set/cleared - that was round-about. So, right now everything seems to be working. My question now is: What next? I presume that you want me to capture a log, but please point me to a specific procedure (document or otherwise). I seem to have gotten lost in chasing my own tail!

One note that should be just FYI. We have a 0.10 Ohm sense resistor in our circuit. Since out application is low current, I am leaving bqStudio think that it's a 0.01 Ohm value. This means that we read the current in tenths of a mA and that our battery has a 23000 mAHr rating. If this presents a specific problem. let me know.

Thanks,
Dave

0 Bryan Kahler over 7 years ago in reply to David Pfaltzgraff

TI__Mastermind 25955 points

Hi Dave,

At this point, it would be best to switch back to the steps listened in section 8.2.2.1 of the Datasheet.

In section 8.2.2.1.4, Step 4, we determine the Sense resistor value. The parameters to be modified are:
- Data Flash:Calibration:Data:CC Gain
- Data Flash:Calibration:Data:CC Delta

These values will be modified when calibrating the device (in the calibration tab). Please apply a 1 A load when calibrating.

When calibrating the device, please perform the Board offset calibration prior to calibrating voltage or current.

It is recommended to use bqStudio for modifying these parameters. If EVSW is used, the value shown will differ because there is a required DF to EVSW conversion, as described in more detail in Table 12 of the Datasheet (page 29).

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

First, I need some clarification. You refer to section 8.2.2.1 of the datasheet. Is this for the BQ34Z100-G1? As Section 8.2 only has the block diagram. I just want to make sure we're on the same page. (My document is: SLUSBZ5A –JANUARY 2015–REVISED JULY 2015)

Anyway, what I did this morning in bqStudio was to:
- Set up a zero current condition and calibrated the CC Gain and CC Delta
- Calibrated the temperature
- Set a current to 100mA and calibrated to 1000 (remember I want a 10x scale)
I didn't recalibrate the voltage as I had done that yesterday and it was still valid.

So, at this point my unit is calibrated. What's the next step in getting what I need for the GPCHEM?

Note: bqStudio has been operating so much better since the firmware upgrade!

Thanks,
Dave

0 Bryan Kahler over 7 years ago in reply to David Pfaltzgraff

TI__Mastermind 25955 points

Hi David,

The document I'm referring to is the next revision, SLUSBZ5B, found here: www.ti.com/.../bq34z100-g1

With your unit calibrated, the chemID may now be determined by using the built in logging features of bqStudio, a charger, and an electronic load.

To create the log, please follow the steps found in the GPCCHEM User's Guide, found here: www.ti.com/.../slva725

The important parameters to log are:
- Time (in s, logging should be in 1s intervals)
- cell voltage (in mV)
- cell current (in mA)
- cell temperature (in degC)

Several cells may be in series for this log, but it is more accurate to us the voltage of one of the cells.

Please note: The resting time values for your battery may be longer than suggested by the guide (as it was written with Lithium Ion cells in mind). Ensure the cell is able to rest after charge and before discharge, and after discharge.

After the log is created (and saved to a file), you can either create the archive by hand for upload to the gpcchem tool, by following the instructions in the User's Guide or automatically with the GPCPackager tool. If you would like to use the GPCPackager tool to create the archive, disconnect the EV2300 from your system and restart bqStudio. Select the latest version of the bq40z50-R2 gauge. Click on the GPCPackager tool. Fill out the information on this tool and it will create the archive for you.

The archive may be uploaded to the GPCCHEM tool, found at this link: http://www.ti.com/tool/gpcchem

If you have any questions, please let me know!

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

OK. Things seem to be going better. However, due to the time required, I probably won't capture the log file until Thursday. (I have to be present to control the charge/discharge sections.) My question right now is: Is there a way to control what's going to the log file? Right now it seems as though it's recording all registers. I think it would be good if we could just eliminate all the unrelated data from the start.

Thanks,
Dave

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

To answer my question about limiting the logging, I found mention of the "Advanced View" in Kevin Main's video. So, that one is answered. (It would help if the manual mentioned this capability.)

Now my next question is: Which should be recorded straight or averaged voltages and currents? If it doesn't matter, I'll go with the average because of smoothing the data.

Thanks,
Dave

0 David Pfaltzgraff over 7 years ago in reply to David Pfaltzgraff

Expert 2445 points

Hi Bryan,

You're probably getting tired of me by now, but... Following the video, Kevin Main talks about the "Learning Cycle" tab. I am unable to find that. On top of that a search in the "help" menu bring up nothing about it. I'm using version 1.3.80 and was interested in the "Manual Mode". Is this covered differently now?

Thanks - again,
Dave

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

OK Bryan, one more question... maybe...

What is the preferred discharge rate for GPCHEM? My previous data was based on Q/10 meaning that the discharge cycle was about 10 hours. When I add the charge and rest periods to this, I end up with a minimum of 17 hours and that's where I run into the scheduling problem. (My program did the timing for me, using bqStudio, I need to be present.) If Q/20 is acceptable, then I could time it so that the discharge time of about 20 hours was during the night.

One place where I'm going with this is: What is the difference if I use bqStudio in manual mode to gather the data or use my program to gather the same data? I've already sent you files: One that I manually manipulated, and one that i packaged as per the GPCPackager using your instructions. Both gave me the same result. Why should I expect anything different now? ("Insanity Is Doing the Same Thing Over and Over Again and Expecting Different Results.")

Thanks,
Dave

0 Bryan Kahler over 7 years ago in reply to David Pfaltzgraff

TI__Mastermind 25955 points

Hi David,

I have copied and pasted your questions and have attempted to answer in line with the [BK] tag. Please let me know if there is anything else missed or that needs clarification.

1. OK. Things seem to be going better. However, due to the time required, I probably won't capture the log file until Thursday. (I have to be present to control the charge/discharge sections.) My question right now is: Is there a way to control what's going to the log file? Right now it seems as though it's recording all registers. I think it would be good if we could just eliminate all the unrelated data from the start.

[BK] You can reduce what is going into the log file, but it is not recommended. The extra columns are extremely helpful in debug. Please use the Advanced View to modify what is logged.

2. To answer my question about limiting the logging, I found mention of the "Advanced View" in Kevin Main's video. So, that one is answered. (It would help if the manual mentioned this capability.)

Now my next question is: Which should be recorded straight or averaged voltages and currents? If it doesn't matter, I'll go with the average because of smoothing the data.

[BK] Excellent! We have new collateral coming out that includes this. Thank you for the feedback! I usually record all but in the config.txt specify the straight voltages and straight currents to supply to the GPCCHEM tool.

3. You're probably getting tired of me by now, but... Following the video, Kevin Main talks about the "Learning Cycle" tab. I am unable to find that. On top of that a search in the "help" menu bring up nothing about it. I'm using version 1.3.80 and was interested in the "Manual Mode". Is this covered differently now?

[BK] The learning cycle tab is a great feature that wasn't added to the tool spec for bqStudio and the bq34z100-G1. I have passed this feedback along to the tools team - this would be a great feature to include! Unfortunately, with the bq34z100-G1 it is by hand for the time being. You could attempt with opening bqStudio to a bq275xx device and then connecting your bq34z100-G1, but I have not had the time to test this - yet.

4. OK Bryan, one more question... maybe...

What is the preferred discharge rate for GPCHEM? My previous data was based on Q/10 meaning that the discharge cycle was about 10 hours. When I add the charge and rest periods to this, I end up with a minimum of 17 hours and that's where I run into the scheduling problem. (My program did the timing for me, using bqStudio, I need to be present.) If Q/20 is acceptable, then I could time it so that the discharge time of about 20 hours was during the night.

One place where I'm going with this is: What is the difference if I use bqStudio in manual mode to gather the data or use my program to gather the same data? I've already sent you files: One that I manually manipulated, and one that i packaged as per the GPCPackager using your instructions. Both gave me the same result. Why should I expect anything different now? ("Insanity Is Doing the Same Thing Over and Over Again and Expecting Different Results.")

[BK] Great question! The preferred discharge rate changes based upon the chemistry of the device. Li-ion devices use a C/10 rate. Lead-acid and NiMH cells use a C/20 rate. Do you have voltage limiting on your electronic load? If so, just set it and forget it, even in manual mode. When not using automated test equipment, I will let the cells run over the weekend. The algorithm is perfectly happy to have the nice flat ocv reading after the cell is discharged.

[BK] With respect to GPCPackager, it simply 'automates' the process of creating the config.file, renaming your log file, and creating an archive with those two files therein. I find it easier to just crank it out by hand, like you did the first time around. Did the GPCCHEM tool return the same result for both log files?

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

Thanks for the thorough reply. I'll work on getting this all in place and start the test this evening. To answer the few questions I can right now...

My intent on limiting the amount of logging is to eliminate those parameters that have no value until we get past this hurdle. For example, why record remaining charge (and related parameters) when the unit has not yet been programmed?

I'll adjust to Q/20 and run as best as I can. My "electronic load" isn't smart enough to disconnect at the end of the cycle. (Hmmm, maybe I can add that yet today.) I'll do the best I can.

The results returned by GPCHEM on the manually formatted data versus the GPCPackager are essentially the same. I'll send both files via the external e-mail just for your perusal. The interesting thing is that even though it lists a 4S configuration, it's as though the program ignores that. If the chemistry is a known factor, why can't the GPCHEM program accept that as a parameter and thus limit the work it needs to do? For example, if we know that it's NiMH, why search the database for a match on the Li technologies?

This question is just "food for thought" from my end... Aren't most NiMH cells essentially the same in their charge/discharge characteristics and merely differ in their total capacity? Shouldn't it be sufficient for me to select an existing chemistry and adjust the total capacity and then run with that?

Later,
Dave

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

I believe I got a good run last night. I tried your trick of disconnecting the EV2300, pulling up bqStudio, and specifying a bz40q50-r2 (I believe I got that right). But bqStudio just said that the packager could not be loaded. There were two versions and I tried both. I ended up doing the packaging by hand. As I typed this, the result came back - nothing! I double checked the config file and it looks OK to me.

I'll send the zip file to you via the external e-mail. Hopefully you'll see what I did wrong...

Have a great weekend!
Dave

0 Bryan Kahler over 7 years ago in reply to David Pfaltzgraff

TI__Mastermind 25955 points

Hi Dave,

Well, the good news is - you're not alone. I ran the file internally and received a 0 as the output. To fix the config file, please update ElapsedTimeColumn=3 and CurrentColumn=6.

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

Well, it's good to know that I'm not alone, but...

Noting that the only real change was in the CurrentColumn, going from average to present, I just changed that line. Same result!

So, I did a copy-n-paste of the text that you included in your reply and resubmitted. I also made sure that the last line was terminated with a <CR> just to be clean. Same result!

As they say: Insanity is doing the same thing over and expecting a different result.

If you got a response other than just the big '0', what did GPCHEM tell you?

Thanks, Dave

0 Bryan Kahler over 7 years ago in reply to David Pfaltzgraff

TI__Mastermind 25955 points

Hi David,

Please also modify: ElapsedTimeColumn=3

That should clear things up.

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Hi Bryan,

Apparently you didn't understand what I said in the second part of my reply. I did a copy and paste of *both* sections and got zip response. (If you want I can e-mail the zip file I submitted for your review.)

If you got something different, I'd be interested in a copy of the files you submitted as well as the results.

Thanks, Dave

0 Bryan Kahler over 7 years ago in reply to David Pfaltzgraff

TI__Mastermind 25955 points

Hi Dave,

I did get a result, I'll send it via e-mail when i'm out of the lab.

Sincerely,
Bryan Kahler

0 David Pfaltzgraff over 7 years ago in reply to Bryan Kahler

Expert 2445 points

Although the thread is marked as resolved (it probably happened by clicking on a link in error and there seems to be no way to change that status), I would like to make a point that TI has failed to correctly address my issues and answer questions related to my application. As a result, we are abandoning all attempts to utilize the fuel gauge functionality as "promised" in the data sheet. Over a year has elapsed since we started this effort and I have worked with two supposedly knowledgeable support personnel and yet the primary questions have not been addressed.

From my viewpoint some of the issues arise because TI does not understand a requirement where the fuel gauge is used to measure small currents and that the chemistry is not Li based. The other is that they refuse to accept data unless it originates from within bqStudio. I have found that the interface, both machine and user, are not friendly. The data produced by bqStudio was noisier than that produced by an outside application. Yet I am expected to believe the result of the GPCHEM program based on the noisy data.

Yes, this issue is closed, but it is not resolved to the customer's satisfaction.

Regards,
David J. Pfaltzgraff

Power management

Power management forum

BQ34Z100-G1: Getting CHEMID for bqstudio