BQ25792: Charger in trickle charge

Hi Raphael,

If the charger is outputting the 100mA trickle charge current as designed, then the pack protector must require more than 100mA in order to close the FET. Does the pack have a spec for this?  Is there a load on BAT that could be stealing the 100mA?  Do you have a sourcemeter that could be configured as a current source and inject slightly more current into the protector to see if that is the case?

Regards,

Jeff

Hi Jeff,

Thank you for your fast reply.

Actually, I have 9 identical boards working perfectly (with the same battery).
But one has this strange behaviour (with other battery packs as well).
So I am trying to figure out if this is a soldering problem, a PCB problem, the BQ25792 or another chip (MCU?), ...

One information: if I plug a 2S2P battery pack, empty but still above 6V, fast charge is working OK.
It is as if the BAT FET was not working in LDO mode.
I am not even sure that 100mA are sourced in trickle charge.
The rise to 2,2V on Vbat can happen with only a few µA (the cells voltage is close to 6V normally).

FYI, no current is consumed on Vbat.
DC/DC converters are powered through Vsys.

Is it possible that the chip be damaged, not working in LDO mode but working in fast charge?
Or some external parameter preventing trickle and/or pre-charge?

Thanks!

Some additional information...

On a working board, with no battery but a CTN on TS, I have Vsys = Vbat = 8,4V.
On the faulty board, with the same setup, I have : Vsys = 7.2V and Vbat = 2.2V.

Voltages on the TS pins are identical of course.

HI Raphael,

Trickle charge current comes from the output of a current source and does not flow through the BATFET.  I suppose it is possible the chip is damaged.  Does the issue follow the board or the battery pack?

Regards,

Jeff 

Hi Jeff,

I have this issue on 4 boards out of 10.
Same fabrication batch, no visible assembly problem.
Independently of the battery pack.
And they are completely new.

If I raise Vbat artificially to 6V (100ohm between Vsys and Vbat for example), fast charge starts.
So I suppose that the BATFET is working.
At least in commutation, maybe not in linear mode.

Do you think it is possible that the IPRECHG register be initialized with a 0mA current value?
It looks like the trickle is working, certainly because it is a separate current source, but not the precharge.
Or maybe a clamping voltage set for a strange reason to 2.2V?

Raphaël

Hi Raphael,

IPRECHG default is higher than 0mA. What are the status and fault reporting when this happens?  Can you take an oscope shot of VSYS, VBAT, IBAT when this is happening?  There is a brief period during VBAT rise when trickle charge current source turns off and precharge turns on.  Could it be getting stuck in that window?

Regards,

Jeff

Hi Jeff,

To keep you informed.

Fortunately, the 2V Vbat limit was high enough for the MCU to boot.
So I had access to the registers of the BQ25792.

There was no specific error.
The charger was stuck in 'trickle charge'.
And the battery was not detected (Vbat not present, which is normal because Vbat was below Vbat_uvloz).

I finally found a way to workaround the problem: I disabled the PFM in forward mode.
As soon as I write into register 12h, Vbat rises to around 5V for a sec, and then climbs to 6,4V (fast charge).

I have no explanation.
Maybe in PFM there is not enough energy from the buck-boost to source the 120mA to the batteries?
As you suggested, maybe there is a problem when switching from trickle (current source) to pre-charge (BATFET)?
If you have any information on this behaviour, please share.
Do you think it is 'safe' and repeatable to use this workaround on volume production?

Thanks and regards.

Raphaël

HI Raphael,

That is odd.  The converter isn't involved in providing trickle charge.  I will ask the designers but their response may be a few days out.

Regards,

Jeff

Can you provide scope plot showing SYS, BAT, IVBUS and IBAT when you are stuck in trickle charge?  If you are currently bucking (VBUS>VBATREG) can you repeat when boosting (VBUS<VBATREG) or vice versa and report what happens?

Regards,

Jeff