TPS51200: PGOOD always "low"

Hi Steve,

Can you check the PGOOD to GND impedance to see if there is anything abnormal. From your description, PGOOD should be high. If your PGOOD pull up connection is correct, my suspect is PGOOD impedance might be low, which caused PGOOD voltage low.

Regards,
Weidong

Thank you very much for the quick reply. When I get access to the board later today, I will measure the impedance from PGOOD to ground.

We are wondering what the difference is between TPS51200DRCR and TPS51200DRCRG4 shown in the datasheet. We'd like to be sure we have the correct part on the board.

We measure 10.5k from PGOOD to ground (this seems low). We measure 10.0k from PGOOD to +3.3V.

TI Tech Support Chat...

Fabrizio: Hi, my name is Fabrizio. How may I help you?
Steven Longworth: Could you please tell me what the difference is between TPS51200DRCR and TPS51200DRCRG4 (the "G4" suffix is not described in the datasheet)
Fabrizio: none, it is just 2 names for the same part
Steven Longworth: got it... thank you!
Fabrizio: this is because G4 suffix stands for echo friendly
Fabrizio: but now it is a standard
Steven Longworth: Great... I'll pass the information to our parts person
Fabrizio: thank you for chatting
Fabrizio: and the G4 suffix are gradually been taken off

We are still investigating why PGOOD measures 0V in our design even though we measure VO as 0.891V and REFOUT as 0.899V, well within 20% specified in the datasheet for the PGOOD signal to be asserted high.... and why we measure 10.5k resistance from PGOOD to ground.

Hi Steven,

Can you provide the schematic for review?

Also, can you replace TPS51200 with a new part and check PGOOD signal?

Thanks

Qian

We removed and replaced the TPS51200 with a new one.  This did not resolve the issue.

Hi Steven,

Can you send the full TPS51200 schematic in pdf file to below email?

Q-CHEN@TI.COM

Also, please probe REFIN, REFOUT, VO and PGOOD at the same time and provide the waveform.

Thanks

Qian

I just sent you the schematic. The board is not available at the moment, but we are planning to put a scope on the signals you suggested as soon as we can.

We understand the TPS51200 has an asynchronous delayed response (10us to pull low, 2ms to release) so we believe if there's any noise on output, it could trigger the delay to continually reset.

In other words........ we believe PGOOD is designed to stay low during soft-start (when the current limit is set to 1/2).... and this could happen if the part thinks it's continually being powering up...... perhaps caused by noise on VLDOIN, VREFIN, REFOUT or VO (same as VOSNS).

We got access to the board and measured Vo.... we found 0.9V but with 200mV ripple, maybe 2 to 6MHz on the regulator output. We assume the high ripple is the reason PGOOD is always low. The device is also very hot to the touch. We believe we have a stability issue here .

We have four 10uF CERAMIC capacitors on the regulator output.... but also two 220uF tantalum caps and two 4.7uF tantalum caps on the regulator output at the load (at the DDR2 memory).

Any ideas on how to bring the regulator into stability?

Hi Steven,

I agree the high Vout ripple should be caused by the stability issue.

Is 4*10uF cap close enough to TPS51200 VO pin?

Is VOSNS trace connected to load side and close to any noisy signal?

Can you change R507EM in your schematic to 10ohm and test again?

Btw, can we discuss by emaill for quick communication?

Thanks

Qian

Thanks for the udpate.