This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TPS2547: System wake-up abnormally after system entering S3 from S0

Part Number: TPS2547
Other Parts Discussed in Thread: TUSB1002A

Hi team,

My customer reported an issue that the system wake-up on itself after entering S3 from S0 with large fail rate when: connecting an USB disk to 2547 charging port and mouse to another USB-A port but both do nothing ==> manually turn PC into S3 and see the system wake-up to S0 again. Connecting the USB disk to other USB3.0 non-charging port will not see the issue.

Schematic as below: ILIM_SEL, CTL2, CTL3 are all pulled high to 5V, only CTL1 is controlled via Embeded Controller. 

It should be in CDP/DCP_Auto with load detection.

In normal case, S3 signal is triggered and kept low:

In fail case, seems there is an USB3 signal transmission then the S3 pin is pulled-high again.

Scale-up on the changing part:

And I checked the D+/D- signal (yellow as U2P and green as U2N) to check the mode change:

Seems to me 2547 is in 2V/2.7V divider mode but not DCP_Auto after entering S3, is this behavior correct?

Scale-up on the Vbus drop part, seems D+/D- changed after S3 signal.

Attach the un-scaled version if you'd like to check the whole process:

I also confirmed CTL1 waveform as below and seems it is working as expected after entering S3:

Scale-up and seems that S3 is leading the process.

So in short, USB3 signal appears ==> S3 signal changed ==>2547 mode changed. But seems that in the OK case, there is also USB3.0 signal after entering S3.

Due to the limitation on the board I'm not able to captured the Ibus current. Could you kindly help to analysis and suggest what is the possible cause of the issue? Is our charger part of the issue? And if any test we could do to clarify if it is USB communication related or others?

It will be helpful if you could also help to explain the conditions that could trigger the wake-up process by 2547. I saw a power-detection and a HID but actually don't know which one should I focus regarding this case.

Thanks very much for your help.

Max

  • Hello Max,

    The expert covering this device is currently out of office, and will get back to you next week. Thank you in advance for your patience.

  • Hi Max,

    To clarify, the wake up fails when the test uses the other USB type-A port to wake up the system using a mouse? Or, when using a mouse the type-A port with the TPS2547, the system cannot wake up?

    Thanks,

    Emma

  • Hi Emma,

    I noticed the misunderstanding in the description and thanks for double checking on that.

    Let me change another way to describe the expected behavior and current behavior:

    The expected behavior: after turn system into S3, system stays at S3 if it is not waked up.

    Current issue: If connecting a usb disk on the 2547 port, after system is turned to S3, it will not stay at S3 but jump out to S0 again on itselt. This issue is confirmed with some fail rate on different boards.

    Other tests applied: if The same usb disk is connected to other non-charging (non-2547) port, the issue disappeared. ( which means the system will keep at S3 after it is turned to).

    One more update, customer also tried to change the TUSB1002A USB3 10G redriver of the 2547 port and the issue also disappear. But using this same redriver on the OK board there will still be no problem. 

    In this case, personally I'm confused whether the wake-up process is related to USB3 signal or loading on 2547 port or other behaviors.

    Do you have some suggestion on this issue? And how does 2547 do the wake-up behavior? I noticed a HID and a load detection power wake-up. Is current issue possibly related to this feature? And if some more tests are required to clarify?

    Thanks

    Max  

  • Hi Max,

    Wake on USB is done using USB 2 on the DP and DM pins so this is unrelated to USB 3, which uses the super speed pins. 

    Power wake is done with S4 and S5, and requires an embedded controller to monitor the STATUS pin, which is also not applicable to your test. 

    Could you provide more detail into the TUSB test? Are you adding in a redriver, switching one out for another one, or removing the redriver, to get the test to work.

    Thanks,

    Emma

  • Hi Emma,

    It's basically a ABA test.

    Change redriver of fail board to a redriver from the good board, and fail board works. While, the pass board still works with the redriver from the failed board.

    Change back the redriver of the failed board and it still fail. Besides, it is a USB3 Gen2 redriver, in the case you mentioned wake up is based on USB2, I'm not sure if this phenomenon is related.

    By the way, after turning to S3 with the expected setting to be DCP_AUTO, I see the D+D- waveform seems to be 2V/2.7V divider mode. Is this an expected behavior?

    Thanks

    Max

  • Hi Max,

    Thanks for the explanation. 2V/2.7V is Divider 1 mode. DCP_auto can charge in either Shorted/1.2V Pull-Up/ Divider 1/Divider 2, so for the device to charge in Divider 1 during DCP_auto mode is expected behavior.

    As changing the redriver of the super speed lines will correct the issue you are seeing, and the TPS2547 does not touch the super speed lines, this issue does not sound related to the TPS2547. Please follow up with any further questions.

    Thanks,

    Emma

  • Get it. 

    Thanks Emma.