TPS25751: Cannot write to TPS25751 after swapping from TPS25750

Jesse Mills

Part Number: TPS25751
Other Parts Discussed in Thread: TPS25750,

Tool/software:

Hello,

We had a functional design with the TPS25750 internal switch variant. The design has been updated to swap in the TPS25751.

After the swap (new board spin, professional assembly) we cannot communicate with the TPS25751. When writing the address from the MCU over I2Ct to the TPS25751, the address is written (has not changed) 0x21.

Scoping the I2C bus, we see the MCU is writing to the TPS, and no response is given.

I have done some basic checks, like voltages on the 3V3_LDO is accurate. 1V5_LDO is also reading 1.5V.

Are there any firmware modifications I have to make on the TPS25751 firmware? Does the target address change? Both ADC1 and ADC2 are pulled low with 100K, so I would expect 0x21 to be the address.

Thanks

Jesse

1 month ago

0 Chris Sterzik 1 month ago

TI__Guru 62881 points

Hello,

Can you retest with 0 Ohms (short to GND) on ADC1 and ADC2? Can you see the analog waveforms?

Regards,
Chris

0 Jesse Mills 1 month ago in reply to Chris Sterzik

Prodigy 155 points

Hi Chris,

I shorted the ADCs to ground, and it appears I am getting some replies now.

Does this mean the ADC quiescent current, or some other parameter on the TPS25751, has changed from the TPS25750?

Is there any I2C address that the TPS will accept, assuming what I am reading here is in fact ADC1 and ADC2 = '1', '1' instead of '0', '0' even though they are tied to ground?

Thanks

Jesse

0 Jesse Mills 1 month ago in reply to Jesse Mills

Prodigy 155 points

Hi Chris,

I spoke too soon. I just received a read with no response from the TPS:

Is the boot time different on this IC, or have other customers brought up similar issues lately?

Thanks

Jesse

0 Chris Sterzik 1 month ago in reply to Jesse Mills

TI__Guru 62881 points

Hello,

Jesse Mills said:
I spoke too soon. I just received a read with no response from the TPS

The leakage can range from -1uA to +1uA so a 100K resistor could result in a maximum voltage of 0.1V. https://www.ti.com/lit/ds/symlink/tps25751.pdf#page=18

Jesse Mills said:
Is the boot time different on this IC, or have other customers brought up similar issues lately?

The specification for Tboot has not changed, but it is only provided as a typical number (see same reference as above). I am seeing close to 16ms.

Regards,
Chris

0 Jesse Mills 1 month ago in reply to Chris Sterzik

Prodigy 155 points

Thanks Chris.

We adjusted the wait time for boot to 20ms and that seems to have solved the issue.

I will still take a note to reduce the pull down impedance in case that comes back to bite us.

0 Jesse Mills 1 month ago in reply to Jesse Mills

Prodigy 155 points

Thank you for taking a look at this

+1 Chris Sterzik 1 month ago in reply to Chris Sterzik

TI__Guru 62881 points

Hello,

If i attempt to read immediately after the glitch on the IRQ then the address is NACKd.

I added a delay after the IRQ line going low:

Hope that makes sense. You can see the full code here.

https://www.ti.com/lit/an/slvafv8a/slvafv8a.pdf#page=16

Regards,

Chris

0 Jesse Mills 1 month ago in reply to Chris Sterzik

Prodigy 155 points

Thank you Chris. We don't monitor the IRQ line, but that is good to know. That seems to be equivalent to a longer wait time on boot.

Power management

Power management forum

TPS25751: Cannot write to TPS25751 after swapping from TPS25750