TUSB1046A-DCI: TUSB1046A-DCI

Thank you,

I don't think I understood completely,

Path through USB type C:

I have 4 USB Type C ports which work. I use an external dongle to convert from USBC to HDMI which probably has the active converter.

Path outside the USB type C:

DP via the MUX HD3SS214I works through PI3HDX511DZLE in the current design. Do I need to change it?

Are DP++ GPUs also supposed to work at DP?

Thanks,

Divakar

Divakar

The DisplayPort Alternate Mode specification for sending DisplayPort signals over a USB-C cable does not include support for the dual-mode protocol (DP++). The GPU must be configured as DP++ when using the PI3HDX511DZLE passive level shifter and configured as DP when using the Type-C port.

In your working solution, I would think the GPU already supporting DP++, otherwise PI3HDX511DZLE will not work since it is not an active protocol converter.

Thanks

David

I see.

We never consciously set it as DP or DP++. It is happening beind the scenes then. 

Can it negotiate via the USB C interface and scale itseld down to DP? The one I am using right now is Aetnina MXM card M3T3000-QN and that is an RTX 3000 NVIDIA GPU. The datasheet of the MXM card says the DP 1.4. I just googled "Is DP 1.4 = DP++?"

Yes, DP 1.4 is considered DP++ (DisplayPort Dual-Mode), meaning it can connect to both DisplayPort displays and other display types like DVI or HDMI with the appropriate adapter, as it incorporates the Dual-Mode functionality within the DisplayPort 1.4 standard

So this means when it goes through PI3HDX511DZLE it means it is DP++ or else it shouldn't work based on what you are saying - But it works. At the same time when it goes via the USB C it somehow understands and lowers itself from DP++ to DP or else it won't work.

The new GPU I plan to use is also DP 1.4 so that means it should work based on what you explained.

Thanks,

Divakar

Divakar

Can it negotiate via the USB C interface and scale itseld down to DP?

USB-C do not communicate between DP and DP++, so I do not expect the GPU is configuring itself based on the USB-C communication. But it is possible that the GPU is communicating with the PD controller and understand from the PD controller that the PD controller has entered DP mode and then configure itself appropriately. But this is something that needs to be confirmed by the GPU.

Thanks

David

Thanks,

I do not see a path from GPU to the PD controller in my design. But the PD controller can talk to the USB partner which in this case is a USB-C dongle via the CC1/2 pins. The dongle is the USB-C to HDMI converter such as ones made by Anker. The GPU AUX pairs become the SBU lines as they pass through TUSB1046AI-DCI to the USB type C connector and the dongle. SBU1/2 are meant for alternate modes and perhaps that is how the negotiation takes place.

Divakar

Divakar

SBU1/2 are meant for alternate modes to support DP link training. I don't believe the SBU1/2 can be used to negotiate between the DP and HDMI. 

Do you know how the GPU is configuring itself to be either DP only or TMDS (HDMI)?

Thanks

David

Is it possible that the GPU defaults to DP mode?

In the non-USB-C mode when the video passes via PI3HDX511DZLE to an HDMI monitor, is it possible the GPU switches to HDMi when it reads the info via EDID?

Aside, AI in google and chat GPT say EDID is read in Alt mode. This can inform if the sync is HDMI or DP.

Thanks,

Divakar

Do you know how the GPU is configuring itself to be either DP only or TMDS (HDMI)?

No I do not. 

Divakar

Please let me check and get back to you on this question.

Thanks

David

Thanks. I have posted some of the questions that came from our discussion at the Infineon PD portal whose PD we are using.

I believe so.

Thanks.