TPS25840-Q1: PCB‘A Screams During TPS25840 Operation

Hi Eric，

1)I have made some modifications to the board：

1. Try another inductor model——I changed 10uH to 2.2uH.
2. Shift RT to change the fundamental SW frequency——I changed 8.9kΩ to 49.9kΩ.
3. Add an RC snubber from SW to GND——I added R=10Ω，C=120pF.
4. Add a series resistor from BOOT to SW (R23).  Something small like 1 ohm.——I changed 0Ω to 1Ω.

2)Oscilloscope Waveform：

1.Output voltage waveform before change (PIN2 of inductor L1) ：

2.Output voltage waveform after change (PIN2 of inductor L1) ：

3)Could you please help to explain：

1.Screams become less sharp but still exist.Are there any suggestions for further optimization?

2.Why does the TPS25840 have an output voltage of 2.6 V when Client mode is used?But the output voltage of TPS25830A is zero？

Thanks

Hi Rubin,

Rubin Chen said:

1.Screams become less sharp but still exist.Are there any suggestions for further optimization?

You should be able to find a frequency for a given inductor that doesn't produce the noise.  Did you try each of my suggestions independently?

Rubin Chen said:

2.Why does the TPS25840 have an output voltage of 2.6 V when Client mode is used?But the output voltage of TPS25830A is zero？

Are you driving any signal on DP_OUT or DM_OUT while in client mode?  If so, I would advise against this.  The DC/DC will be off in client mode.  Also, is the noise occurring during client mode?  I only expect you should be getting audible noise when the DC/DC is on.

Regards,
Eric

Hi Eric，

1）I reconstucted the board according to the sequence of 1→2→3→4.

1.Try another inductor model——I changed 10uH to 2.2uH.——The noise is not changed.

4.Add a series resistor from BOOT to SW (R23).  Something small like 1 ohm.——I changed 0Ω to 1Ω.——The noise is not changed.

3.Add an RC snubber from SW to GND——I added R=10Ω，C=120pF.——The noise is not changed.

2.Shift RT to change the fundamental SW frequency——I changed 8.9kΩ to 49.9kΩ.——Screams become less sharp but still exist.

2）

I did not drive any signal on DP_OUT or DM_OUT.And I did not add any load through the TYPE-A interface.

I pull the CTRL1&CTRL2 down（as show in the following figure），I think the chip has entered client mode.

In client mode，I think the output voltage SW should be 0 volts.But it is 2.6V now.

Is it normal？

Thanks

Correct the above sequence:1→4→3→2.

Rubin,

Thank you for the details.  Let me check with our design team for additional ideas.

Noise like what you describe has always been due to the inductor in my experience.  I am surprised it is not going away.  I will need to speak with other team members to brainstorm other ideas.  It will be a day or two before I have additional feedback.

Regards,
Eric

Eric，

OK. I look forward to your reply and support.

Thanks

Hi Rubin,

My other team members also agree it is most likely the inductor producing the noise.  One suggestion was made to bypass the nFET by installing 0 ohm for R17.  This isn't a long term solution - just a debug activity for now.

Regarding the output voltage - I misspoke when saying the DCDC should be off.  If you measure the DCDC output at R6, I would expect 5V.  If you measure pin 1 of L1 (SW node), you should see a square wave with amplitude = VIN.

I still need to brainstorm some more tests we can perform to debug this circuit.  I apologize for the limited feedback at this time.

Regards,
Eric

Hi Eric，

I'm very sorry that I didn't provide you with the processing BOM before. Due to the 15mΩ sampling resistor not arriving in time,the sampling resistor R6 was not soldered. After soldering R6 of 0Ω, VOUT output 5V normally and the noise disappeared.

Could you help to explain why the noise appeared when R6 was not soldered?

Thanks