BQ24640: PG problem

Dave Nadler

Part Number: BQ24640

Hi - I've got a strange problem with the PG output.

First, the datasheet defines PG thus: Open-drain active-low adapter status output. Connect to pullup rail through LED and 10-kΩ resistor. The LED turns on when a valid is detected, and off in the sleep mode.
That's an error - a valid what? I'm assuming you meant "valid input power", no?

My board has a small carry-through UPS. After the board powers up and does a few self-checks, it enables the BQ24640, and the super-cap voltage ramps up to the desired value of 10.8v as expected.
If I disconnect the input power, the UPS runs the board for 20 seconds as desired. So far, so good. However...

I read the PG signal at the uC using a 100k pull-up. When I power on the board first time, everything works as expected. !PG stays low until I disconnect the power.
However, if I reset the board during debugging (with the super-caps charged), the PG signal is erratic and sometimes reads 1 and changes between 1 and 0.
This happens with input source (from bench supply through input filters and diode) at 11.6 volts and super-cap voltage 10v-to-10.8v.

Any idea what is going on???

UPS_Schematic.PNG

Thanks!
Best Regards, Dave

over 1 year ago

0 Michael Emanuel35 over 1 year ago

TI__Mastermind 34215 points

Hi Dave,

See Section 7.3.14 for more description on PG. Yes it means, "when a valid input is detected".

What happens when you try a higher input voltage, say 13 V?

Regards,

Mike Emanuel

0 Dave Nadler over 1 year ago in reply to Michael Emanuel35

Intellectual 365 points

Unfortunately I can't easily test a higher input voltage, aarrggg...
Sure I read section 7.3.14, thought that section seems to be missing
the 100mv required difference (Vcc > VSRN+100mv) - from the electrical specs, no?
Anyway Vcc (11.6) > VSRN (10.8) by a good margin.
Any idea why this is happening?
PG flakiness seems to continue for a few seconds after a reset.
Thanks!
Best Regards, Dave

0 Dave Nadler over 1 year ago in reply to Michael Emanuel35

Intellectual 365 points

Also, does my schematic look OK?
Thanks!

0 Michael Emanuel35 over 1 year ago in reply to Dave Nadler

TI__Mastermind 34215 points

Hello Dave,

My schematic review is as follows:

1. Not really sure what the green arrows inline with output side are? I assume they are just current indicators but not a component?

2. Please make sure the closest thing on the REGN node to the pin is the decoupling capacitor for it. In layout do not put the C56 super far away yet the BTST circuit close for example.

3. Confirming charge regulation voltage to be 10.79 V.

4. Confirming charge current at 100 mA. Please keep in mind with 100 mA and 25 mOhm the charge current regulation accuracy will be between +/-25% and +/-50% due to the sense voltage being 2.5 mV.

5. We do not usually use inductors and sense resistors this high, test according to your application.

Section 9 Power Supply Recommendations reads, "For proper operation of bq24640, VCC must be from 5 V to 28 V. To begin charging, VCC must be higher than SRN by at least 500 mV (otherwise, the device will be in sleep mode). TI recommends an input voltage of at least 1.5 V to 2 V higher than the super capacitor voltage, taking into consideration the DC losses in the highside FET (Rdson), inductor (DCR), the input diode drop, and current-sense resistor (between SRP and SRN). Power limit for the input supply must be greater than the maximum power required for super capacitor charging."

We recommend a full 1.5 V to 2 V higher.

If you cannot test a higher voltage, please adjust the VFB divider such that you have 2 V of headroom and see if the issue persists.

Regards,

Mike Emanuel

0 Dave Nadler over 1 year ago in reply to Michael Emanuel35

Intellectual 365 points

Thanks Mike...
1. Big green arrow is just current flow annotation; no component.
2. Layout should be OK.
3. Yes, target 10.8v.
4. Yes 100ma charge, thanks for pointing out the range/tolerance, should be OK for our application.
5. Hmmm. I used TI calculation spreadsheet for this, attached. 6.8uH/20uF might have been a better choice.

I'll need to change my bench power setup to test at a higher voltage.
In the field this board will typically run from a higher voltage...

Still unsure why any of the above could cause PG flakiness.

Thanks again,
Best Regards, Dave

1805.bqstroller calculation tools_V1.5.xls