BQ24125: Inrush current or hot plug limitation?

Hello,

Can you please explain what this wall wart is? Can you send a copy of your schematic? Can you explain your setup in more detail? You need to guarantee your input supply to the board to make sure you don't damage it upon start up. Can you take a scope capture of it under load upon plugin?

Thanks,

Mike Emanuel

See https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/990463/bq24125-bq24125-failing-upon-repeat-power-cycles-vin-to-vbatt-short?tisearch=e2e-sitesearch&keymatch=BQ24125# for context and schematic.

Note R35 and R36 changed to 560k and 110k to avoid voltage divider error per conversation https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/1002256/bq24125-failing-upon-repeat-power-cycles-vin-to-vbatt-short-continuation/3703978#3703978.

Wall wart datasheet here https://phasiumpower.com/projects-archive/mango25s-series/

I cannot take a new scope shot as the board was damaged as described.

Here is a scope shot of normal powerup by switching on AC to the wall wart.  CH1 = input voltage, CH2 = switch node, CH3 = output voltage, CH4 = output current.  The noise seen in the second image is due to switch node signal coupling into the other channels, it is not present when the probe is lifted off the switch node in the first image and is such is not "real".

Can you please send a waveform of the input coming up if possible? In these waveforms the input has already turned on. I need to see the behavior of the input upon plug in.

Also, what is the load on VSYS? Is this a large load and does the problem occur only when the system load is enabled?

Thanks,

Mike Emanuel

I cannot, the board was already damaged.  I do not have another to spare.  The load is not large, it was in fact a near full charge battery.

In any case, powerup on the wall wart did not damage the part.  Connecting to the supply already powered up damaged the part.  And as I described, after the failure we connected the supply to various capacitors in the same range as this board and saw no spikes.

That is correct, I have since confirmed that the other damaged boards were plugged into different live supplies, one a bench supply and the other I believe was a different adapter. 

Good catch on Vcc decoupling, I missed that on review.

Can you explain why the input cap would cause a vulnerability on the supply? 

I understand the larger input cap will lead to higher inrush current through the input connector to the input cap, but I would expect this to lead to a slightly slower voltage ramp at the IN pin on hot plug.  If larger input caps cause higher currents through the BQ24125, can you please explain the mechanism?

I agree that it causes higher inrush current, but not along a path that should be harmed by inrush current.  This is at the input side of the IC.  Again, what is the mechanisim to harm the IC that is downstream from the input cap?

Unless we can probe the input voltage upon plug in, we will not be able to tell what is happening on your board. Can you order more ICs and swap them out to measure the input voltage upon hot plug?

Thanks,

Mike Emanuel