BQ27542 State of Charge

Adam Johnson

Other Parts Discussed in Thread: BQ24296, BQSTUDIO, BQ27542-G1, BQ27510-G3, EV2400, GPCCHEM, BQ25890

I am working on a project using the BQ27542 Battery Gas Gauge paired with a BQ24296 Charge Controller and a user-replacable NCR18650B Li-ion MH12210 battery. I have evaluation modules for both the charge controller and gas gauge, and a PCB of our own design with both chips on it.

Here's my puzzle: when I connect the battery to the development modules, the state of charge reads correctly. When I connect the battery to our PCB, the state of charge is always very low.

Other features seem to be working fine, so I don't think it's a hardware problem. I've made sure the Design Capacity is set correctly on both sets of hardware. Are there other settings on the gas gauge I should be verifying?

Thank you.

over 9 years ago

0 Fernando Lopez over 9 years ago

TI__Expert 8230 points

Hello Adam,

Couple questions that will help me understand better your scenario:

Can you please provide a schematic?
Have you performed a learning cycle on your PCB?
Have you calibrated the gauge on your PCB?
Did you import your golden image from the EVM to the PCB and performed a cycle?
What ChemID are you using? How did you select it?
Is IT enabled?

Thank you.

0 Adam Johnson over 9 years ago in reply to Fernando Lopez

Expert 1310 points

Fernando,

I have not performed a learning cycle on my PCB.
I have not calibrated the gauge on my PCB.
I have not created a golden image.
I haven't changed the ChemID on the chip.

I was hoping that the default values would work, since they appeared to work on the evaluation module. But it sounds like there's lots more to do. Is there a document describing how to do these things? Is there a way to use the gas gauge without using a golden image?

Thank you,

Adam

0 Fernando Lopez over 9 years ago in reply to Adam Johnson

TI__Expert 8230 points

Hello Adam,

For our Impedance Track gauges, it is crucial to set them up appropriately for them to function correctly.

The step may seem a little long at first, but it's simple to follow and the more practice you do, the easier it gets. The gist of the the process is as follows:

Calibrate the gauge. This is perform by going on bqStudio and follow the prompts on the "Calibration" tab.
Obtain and load a chemID. Best way to do this is to perform a relaxation-discharge-relaxation and use our online GPC tool to obtain a match to one of our chemID from our database.
Set your data flash parameters based on your system (charging voltage, taper current, design capacity, etc.)
Perform a learning cycle on the gauge.
Generate the golden image.

We provide a User's Guide for the bq27542-G1 that walks you through the process of setting up the software and the hardware to begin evaluation.

A more detailed process is as follows:

Connecting the cell to the EVM and the EVM to the EV2300/2400 and start communication using bqStudio.
Calibrate the gauge through bqStudio. There is a calibration tab on bqStudio that helps you calibrate the CC offset, board offset, current, voltage and temperature. Information on calibration can also be found here. The gist of calibration consists of the following:
- CC Offset and Board Offset: ensure that no current is flowing through the board and simply check the boxes for these 2 parameters and click "Calibrate Gas Gauge".
- Current: Apply at least a 1A load and enter the value (negative value since it's discharging) on the field. Check the box and click "Calibrate Gas Gauge".
- Voltage: Apply a known voltage to the pack terminals with no current flowing through the board and type the value on the Applied Voltage field and click "Calibrate Gas Gauge".
Choose a chemID for your cell. bqStudio has another plug-in that allows you to load a chemID to the gg. To choose your chemID we recommend you perform a rel-dis-rel cycle and use our online GPC tool to obtain a chemID from our database. Once a chemID has been selected, go to the Chemistry tab on bqStudio, look for the chemID and select Update from Database.
Perform a learning cycle. We have an app note that discusses how to achieve a successful learning cycle. A learning cycle is necessary to obtain the correct QMAX value and update the impedance tables to ensure the gauge can properly calculate it's values.
After this, a golden image can be generated. See this app note that describes the process.

Regarding GDK, it is an excellent tool to help you streamline the cycles during the evaluation process and the plug-in is built into bqStudio. You can also use another setup that can source/sink current to the cell to perform the cycles, but you would need to monitor your setup manually to ensure the charging/discharge is performed correctly.

Hope this helps.

0 dMax over 9 years ago in reply to Fernando Lopez

TI__Genius 14985 points

Adam,

Fernando correctly explained the flow for working with the bq27542, but after re-reading your original post it sounds like you might be better off using a different gauge. The bq27542 is considered a "pack-side" gauge that is typically embedded inside a smart battery pack with the cells. If you want the gauge to be on the main system board while the user swaps out "dumb" batteries then you want a "system-side" gauge, also known as "host-side". The bq27510-G3 would be an example of a flash-based gauge similar to bq27542 but with specific features designed to be used with removable packs and working on the host-side. You also implied that you are hoping for something that you can use with mostly default values while only changing Design Capacity. Along those lines we would recommend something from this family: bq27421 (chip-scale package with integrated sense resistor), bq27441 (plastic package requiring an external sense resistor), or bq27621 (chip-scale package using no sense resistor at all). Read the quick start guides in those corresponding product folders to see the basic parameters which you can quickly configure. Note that they are ROM gauges, not flash, so you will need to re-configure them every time they lose power, but this is easy to do. We also have sample code and drivers for working with any of these single-cell gauges which you can find starting here:

http://processors.wiki.ti.com/index.php/Linux/Android_Software_Solutions_for_TI_Single-cell_Gas_Gauges

There is also code in there for bqTool, which allows your host to import and write a full or partial gg.csv file for configuration, in the same manner you might do with bqStudio. Just make sure that the gg.csv file has all columns when you first export it from bqStudio, by selecting Export All Columns in the Data Memory section in the Preferences (found in the Window menu).

0 dMax over 9 years ago in reply to dMax

TI__Genius 14985 points

One more comment: The reason you probably get reasonable SOC values when connecting the battery to the EVM but not to your system board is that the system is probably drawing current from the battery and distorting the voltage when you first connect it. The gauge is trying to use the voltage to get an initial guess about the SOC, but if there is significant current flowing in or out of the battery it can cause significant error. That's one of the problems which a pack-side gauge will avoid, but if you use a system-side gauge then you need to try to minimize current in or out for the first few seconds while the gauge initializes.

0 Adam Johnson over 9 years ago in reply to dMax

Expert 1310 points

Fernando, dMax,

Thanks, both of you, for your timely responses. That helps a lot!

Adam J.

0 Ganesan Manikandan7 over 8 years ago in reply to Adam Johnson

Intellectual 285 points

Hello, I am working with a very similar setup as Adam J. I am using the BQ27542-G1 as a system side battery fuel gauge. Right now I am at the point where when I connect the I2C lines from my PC board to the EVM, the fuel gauge IC is detected but state of charge is always reported as 0. I am yet to do the calibration and create the golden file. Also, I am trying to obtain a Chem ID for my battery using the online GPC tool, but I always seem to get an error message when I submit the zip file containing the config file and data.
The battery is a custom made 1S2P, 7.2Ah minimum, Li Ion Polymer from Ascent Batteries. I have the setup including bqStudio software and the EV2400 interface module. Can someone please help me find the right Chem ID?

Thanks in advance!

Best,
Ganesan

0 Onyx Ahiakwo over 8 years ago in reply to Ganesan Manikandan7

TI__Guru* 76230 points

Hi Ganesan,

Are you following the instructions listed in the user guide for gpcchem? More often than not, customers have issues with the tool because they didn't follow the instructions. If you send the files you are sending to the tool, perhaps i could help check to see where the problem is coming from. Btw what error message is returned by the tool?

thanks

Onyx

0 Ganesan Manikandan7 over 8 years ago in reply to Onyx Ahiakwo

Intellectual 285 points

Hi Onyx,

Thanks for the quick reply. Yes I can send you the zip file. Please provide an email address. But here is the error reported by the tool.

Your Gauging Parameter Calculator report is here.

Texas Instruments
Thank you for your interest in the Gauging Parameter Calculator from TI.
There was an error in processing your report. The specific error encountered is:

There was a problem with the content of the zip file.
Invalid Zip File Member - TestData.txt
Required file 'roomtemp_rel_dis_rel.csv' is missing from the input zip file.

Please see the Gauging Parameter Calculator for information about the file format and the procedure for creating the input zip file. Should you want to resubmit or need assistance with the Gauging Parameter Calculator, please reference the sites below:

• Go to Gauging Parameter Calculator to re-submit.
• Visit TI E2E™ Community on Battery Fuel Gauges or the Gauging Parameter Calculator.
• TI Battery Management Homepage

Thank you for using TI's Gauging Parameter Calculator for Battery Fuel Gauges.

Thanks,
Ganesan

0 Onyx Ahiakwo over 8 years ago in reply to Ganesan Manikandan7

TI__Guru* 76230 points

Ganesan,
ps read the instruction for using the tool. Your zipped folder ought to contain two files named accordingly:
1. config.txt
2. roomtemp_rel_dis_rel.csv

thanks
Onyx

0 Ganesan Manikandan7 over 8 years ago in reply to Ganesan Manikandan7

Intellectual 285 points

Onyx,

I submitted the files with the proper naming convention. I get back a report in my email with a GPC_Report.txt but it shows an error like:

Error: Wrong data format, make sure file has coma or tab separated values. Name: roomtemp_rel_dis_rel.csv

Thanks,
Ganesan

0 Onyx Ahiakwo over 8 years ago in reply to Ganesan Manikandan7

TI__Guru* 76230 points

Ganesan,

ensure that your file follows the column format listed in the instruction manual. Like the error reports, your columns should either be tab or comma separated. Did you use bqstudio to extract your log file?
thanks
Onyx

0 Ganesan Manikandan7 over 8 years ago in reply to Onyx Ahiakwo

Intellectual 285 points

Hey Onyx,

Thanks for the reply. No I am not using the bqStudio to extract the values, I am using a battery simulator and its included software. However I ensure that a csv with the required name is submitted. I get back the zip folder with the following text in GPC_report.txt

0
Warning: current in A detected, divider changed to 1

I don't see the chem ID.

Thanks,
Gan

0 Onyx Ahiakwo over 8 years ago in reply to Ganesan Manikandan7

TI__Guru* 76230 points

Gan,

Can you attach your log file so i see what is wrong with it

thanks

Onyx

0 Ganesan Manikandan7 over 8 years ago in reply to Onyx Ahiakwo

Intellectual 285 points

Onyx,

Here are my config file and roomtemp_rel_dis_rel.csv files

https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/196/bq25890.7z

Power management

Power management forum

BQ27542 State of Charge