BQ20Z95 LiPoly: compatibility and reset

1) You should do a discharge test on your battery to identify the best CHEM ID using the math cad program downloadable from the 20z95 product folder http://www.ti.com/lit/zip/sluc138.

details of the test are in attachment 6355.chemselect_cont.pdf

Log the discharge dataevery second. If you have problem running mathcad you can send me your discharge data and I will analyze it for you .

2) Once you know your CHEM ID you will have to configure your part using BQEASY software which will use a new template. Then you will have to run learning cycles to allow the gauge to learn the battery's actual capacity etc. 

Thank you very much Mr. Ewer. 

I've tried to look at the pdf file you sent me but I'm still unable to execute the test.

At first just let me tell you my equipment: BQ24133EVM-15V, BQ20Z95EVM, three LiPoly 4.2V 2000mAh cells connected in series, Agilent E3631A power supply

In the document, step 1 says that I must face a thermistor to the battery cells, but...

1) I do not have any external thermistor. Do I need three thermistors, one for each cell?

2) Which one could I use? What about the TMP36GZ?

In step 5, furthermore, it's said that I must charge the cell to the manufacturer specified voltage, but...

1) How can use only the BQ24133EVM-15V to charge three cells with cell balancing?

2) How can i monitor that cells are full charged only with the BQ24133EVM?

I'm sorry for the huge amount of questions.

Thank you very much

Regards,

Giuseppe

You will only need data from one cell to characterize so ideally you would extend the leads to one of evm thermistors so it can be attached against side of that cell. But if you can't then we can still proceed for a first pass attempt with temp data from the evm mounted thermistor to get some sort of initial chemistry match, it just will not be as accurate as it could be.

I am not familiar with the bq23133evm so if it is complicating things just charge the cells using a current limited bench pack in constant voltage mode. I presume your cells are fairly closely matched to start with? Just watch the cell voltages and use the one which reaches full charge first to determine when to stop charging. This will then be the cell you will chararacterize.

Your bigger problem I see is how you will control the discharge current as you do not seem to have a constant current load in your list of equipment. So you will need to come up with some kind of constant currrent load circuit for the discharge portion of the measurements.

Ok, no problem for the load: i've a device that discharges at C/10 rate with constant current, so I can start the test

Thank you

Hello Mr. Ewer,

I've executed the test following the instructions on the pdf file. Here, attached, I send you the log. Could you, please, take a look to help me find the right chemistry?

Many Thanks 

Giuseppe3223.chemistry_cln.txt

Hello Mr. Ewer,

just let me give you a more updated status: I managed to execute the MathCad script by myself and, finally, found that the best matching chemical ID is 1189. 

I think, now, that I should use the bqEasy wizard to set all the parameters and the chemistry to flash my board, is it right?

What did you mean when you told about learning cycles? Could you, please, explain which are the next configuration steps I should go into before I can say BQ20Z95 is correctly setup?

Again, thank you sincerely

Regards,

Giuseppe

Follow the learning cycle described in  http://www.ti.com/lit/an/slua334b/slua334b.pdf. 

There are two procedures described in sect 2 but I would use the procedure for "When no close Qmax value is known:" starting top of page 2.

I was just looking at your discharge data to see if I got the same match as you and you have a nasty negative spike in your discharge voltage at 16:24:14 when voltage drops from 3.974V to 1.548V and does not recover back completely for about 10 minutes. I believe that would likely affect the accuracy of the match. I would repeat this test again.

I am wondering if you stopped discharge a little early as it would have been good to capture a little more discharge to catch more of the knee at the end. Also it is good to log the 2hr relax period  as you stop charging to full and log the 5 hour relax period after you stop discharging.

Right. I sent you that log before trying by myself. 

To let the machine work, indeed, I've had to modify the log: yesterday, at 16:24, moving the board I turned it off and track some wrong values. 

Please, look at this file, now

Thank you

Giuseppe1106.chemistry.xls