TPS65987D: TUSB1146 DP Alt Mode not driving HPD high

Hi Merlyn,

The project configuration of the DP HPD GPIO seems correct. 

Do you have a PD log of the USB-C port CC line messaging?

We are also looking for a DP alt mode message called DP "attention".

Thanks and Regards,

Chris

Should the Application Customization Tool show this in the debug log or do I need a different CC line debugger?

Hi Merlyn,

No, this is separate from the App Customization Tool. You would need another CCline debugger to see the PD log. If you don't have one, we would need to find alternative ways to debug your issue.

Thanks and Regards,

Chris

Ok I'll work on getting one, but what other debug options are there?

In the meantime, I should have mentioned that I am only using 2 DP lanes. So in theory I should be using A,B, or F pin configs for DFP_D, but when I do I no longer see Port 0 DP Mode Select Event going high on the GPIO. I must include pin config C to see the event go high when connecting to a monitor.

Hi Merlyn,

DP pin assignments A,B and F have been deprecated from the DP spec and are not supported anymore, the PD controller should only be negotiating C, D or E.

In the project you shared, I only see C,D and E connected.

With the EVM connected and powered, first check the Data Status register and make sure the HPD bit is low. Then connect the monitor over type-C and check the same register has changed the HPD status.

For HPD to go high, the monitor needs to send a DP "Attention" message over the type-c PD communication. If we do not get this message, we will not update this register and will not assert the HPD bit. Without the PD log, we can't observe the attention message, but the status register should get us closer.

Thanks and Regards,

Chris

Ok back to using just C,D,E pin configs. I was just "trying stuff".

Can I get reg 0x5F to appear in the application? Doesn't show it in list of debug registers or in snapshot.

Ok, I found that there was a raw register read in the scripting tab.

So running a raw register read of 0x5F I get:

No monitor connection result: 0x0

Monitor connected result: 0x513

So looks like bit 15 (HPD) is 0 (low)

Merlyn Bluhm said:

Ok, I found that there was a raw register read in the scripting tab.

Correct, the scripting tab in debug mode should have all of the configuration and status registers.

If we do not see the HPD, there is a chance that the monitor is not sending the attention message, or we are not reaching that point in DP communication. Can you do another  test where you power up the system, check and report the DP SID config and status registers?

At this point, it is difficult to determine what is happening without the PD analyzer. The additional registers may provide more information.

From the read of 5F, once the monitor is connected it seems like we enter DP in pin assignment C or D which is a good sign. Is there anything attached to the HPD pins that could possibly be .

Can you also try retesting with this bit set?

Thanks and Regards,

Chris

Reading reg 0x58 (DP SID status) on startup gives 0x0. After connecting to monitor 0x405000004060000008a0000000103

Setting the Auto-entry bit doesn't seem to have any effect.

And for what it's worth, I disconnected GPIO3(HPD) on the phy and left it floating and it still doesn't go high.

Hi Merlyn,

From what you shared, it seems like an Attention message is either never sent by the monitor, or our device sees the attention and does not assert HPD. Everything else looks correct. Without the PD log, it will be difficult to proceed. Please give me a couple days to try testing your pjt on an EVM. Which GUI and base fw version are you using?

Thanks and Regards,

Chris

GUI is 6.1.4 and fw is TPS65987_88_F907_14_14.bin

Thank you

Thanks Merlyn,

I'll get back to you in the next couple days.

Thanks and Regards,

Chris

Ok I was able to get my hands on a debugger (CY4500). It looks like the attention message is sent?

FYI messages 1-67 are generated when connecting to external monitor. Messages 68 up are generated after the monitor goes into power saving mode.

Logs are attached. Will have to use the EZPD Protocol Analyzer Utility from Infineon to read decoded messages.

DPaltmodedebug.zip

Hi Merlyn,

Nice! This should help a bit.

I'm having issues opening the ccgx file. I've opened them in the past but this one seems to not work with the software. Could you try resending a new PD Log? It is hard to debug the csv file.

I'm going to sideline the EVM test now that we have logs.

Thanks and Regards,

Chris

Nvm, needed to update the utility. Let me review the logs and get back to you.

Can you try taking the logs with a different type-c monitor? I see a potential issue but want to know if it is seen across multiple monitors.

Thanks and Regards,

Chris

Ok, only other USBC "monitor" I have are some Viture XR glasses, but having the same problem.

Logs attached.

8540.DPaltmodedebug_viture.zip

Hi Merlyn,

Just to confirm, you are using the TPS65987DDK, not the DH, correct?

This will affect what FW version should be used.

Thanks and Regards,

Chris

No, I'm using TPS65987DDHRSHR