BQ34Z100 - Charted Average current not stable with pulsed load

Martin
It could be a current sampling issue. Could you send the log file for this data?

Tom

Av Current Pulse load issue.xlsxHere you go Tom,

Thanks for looking at this.

Regards

Martin

Martin,
The passed charge is about the same for the reported Current and Average Current, so that is good. The gauge is running a new simulation every time that the current pulses start, so we may need to adjust your Load Select to provide a better simulation with the pulses. Can you export a senc file from the bq34z100, so that I can check some parameters?
Tom

ISS U.xlsxSenc export is a bit problematic in HDQ.

I've attached a GG in an Excel.

Regards

Martin

Martin,
You can try a couple of options to tweak the simulation.
1) You can try switching to Load Mode = 0. This option makes the simulation more predictable.
2) Try using Load Select = 6 and the User Rate. You can increase the User Rate parameter to turn the simulation to better fit your discharge profile.

Tom

Hi Tom,

Thanks. I've had that in the back of my mind as a fall back but I'm concerned that we have to. It would make our product less generic and very much more user specific.

If we test another battery of ours that uses a BQ20z65 we do not experience this effect running the same test on our Maccor tester.

Is it because the conversion time on a BQ20z65 is 4 times faster than a BQ34z100 ?

Please see attached other effect on average current when 1A load is switched 1s on, 1s off.

Best regards

Martin

Hi Tom,
I have some further information. Maybe it will ring a bell with you, but obviously I'm still puzzled:-

If the load is switched between 2 current levels but never goes to zero in between, we get a flat average discharge, no sawtooth average current waveform and the gas gauging prediction is correct. It is only when the load has a repeated part duration at zero that we get the problem.
Is it some kind of "relax " issue when the DSG current momentarily goes to zero?

Best Regards

Martin

Hi Tom,

We have kind of solved the problem. However, I'm not sure it is the best solution or if it is not really an inherent problem with the BQ34Z100 though.

When the load switches to zero (or less than the sleep current) the fuel gauge goes to sleep. If this happens regularly over a discharge this affects the average current calculation and hence the RC prediction to the extent shown in the waveforms above.
Why does routinely going to sleep affect the average current so adversely?
and
Is there a way of solving this?

Best Regards

Martin

The only way to change this behavior is to disable sleep mode.