BQ28Z610: Battery pack with Bq28Z610 is being used with BQ24075 power path and has conflicts

Dave,

Yes when the bq28z610 enters shutdown the charge pumps to the FETs are disabled. When the bq24075 is connected to the pack pins it should apply a voltage which will wake the bq28z610 device. The bq28z610 needs this voltage on the pack pin to wake and recover the FETS. This is expected behavior of our device. 

Thanks,

Eric Vos

Yes it is supposed to but we are having a portion of our devices where it does not so saying it is supposed to does not solve the problem

Dave,

So you have confirmed you have voltage on the pack+ pin of the gauge and the gauge is not waking, but remaining in shutdown?

Thanks,

Eric Vos

Yes that is correct I can see the BQ24075 having the battery connection going high and then low and then high and low etc at 4.2 volts, the pack does not reset, but when I just connect the pack to a resistance pulled to 4 volts, the pack begins charging.  Then I hook it up to the BQ24075, and it charges like normal.

Dave,

Are you able to put a probe on the pack+ pin where the gauge is measuring so i can see what the signal looks like to the gauge?

Thanks,

Eric Vos

Eric,

No it is sealed.  It does not occur on all devices just a few.

Dave,

instead of using the charger if you apply a DC power supply do the failed units wake? 

It could be that the charger is turning off too quickly and/or some added resistance is not allowing the gauge pack pin to get high enough for long enough

Thanks,

Eric Vos 

Hey Eric  As I mentioned previous it does come out of the shutdown with a DC power supply through a resistor.  There is no resistance issue as after it comes out of it, and I put it on the BQ24075 circuit, the charge current hits design maximum.  The startup voltage spec I believe is set to close to half the charge voltage out of the BQ24075, and the signal while checking for battery is 50% duty.  The typical charge voltage for BQ24075 is 4.2, the BQ28Z610-R1 has a Vstartup of 2.25.  The charge voltage can be as low as 4.16, and the Vstartup can be as high as 2.45.  Won't this be an issue?

I have gleaned more info - in the tech ref for the BQ28Z610 it says it comes out of it on greater than Vstartup.  It also says later that it is Vstartup-.  Then the specifications on the main datasheet has no reference for Vstartup, but rather refers to this voltage as Vshutdown (+ and -).  So if it is -, it should work if everything is nominal, but it does show a Vshutdown-  can be as high as 2.2.  so the question is what kind of filtering or delay is on this trigger?  I thought it specifies 500 mS which is much longer than the pulse coming from the BQ24075.

Dave,

I spoke with a designer and the pulse width from the charger needs to be around 100mSec to wake the gauge completely. Some devices are faster, but not guaranteed. 

Thanks,

Eric Vos

Eric, thank you for this.  I just noticed that when we have this problem, the "INIT" flag is off.  How does this affect the recovery?

Yes I agree you would think.  However, the reality is that I have about 5 % of a pre-production run that is exhibiting this behavior, and in every case, I can hook up the battery pack to the bench, read that it is not initiated, and see that the FETS are off with no flags saying anything about being in shutdown.  I hook it up to our BQ24075 circuit, and it does not come out of it.  I hook it up to the same voltage but not pulsed, comes right out of it.  I don't think the BQ28Z610 is recognizing the charger.

And the pulses out of the BQ24075 are 250 mS long 50% duty.

Dave,

If you can communicate to the gauge please grab a log file of the register settings and we can see why the gauge is holding the FETs off. Again if it is communicating it is not in shutdown mode. 

thanks,

Eric Vos