TDP142: Trace Length Calculation, Cascade TDP142?

Hi,

1. No, the example assumes the GPU and the TDP142 are placed on the same board. The trace length A is the PCB trace length between the GPU and the TDP142. But depending on the loss of the flex cable, you can set the TDP142 RX equalizer to compensate for the flex cable loss.

2. Yes, you can cascade the TDP142 if needed.

Thanks

David

Hi David, 

Let say I have 6 mil trace width, which i will have 12 inch length for trace A . i have utilize all the 12 inch on board without including the flex cable, as describe in photo below 

My Questions are:

1. Is the RX equalizer able to handle this ? 

2. is Cascaded TDP142 a must in this case.

Hi,

What is the data rate and do you know how much is the loss of the 12in trace plus the flex cable?

For example, DP1.4 is 8.1G and Nyquist frequency is 4.05GHz. Let's assume ~0.7dB/in loss for FR4 PCB material at 4.05G, the 12in loss would be ~8dB, this is still within the DP142 maximum EQ of 14.4dB. 

If the loss of the trace, the flex cable, the flex cable connector, etc is more than 14.4dB, this is when you need to consider cascading the DP142.

Thanks

David

Hi David,

Thanks for the reply.

I have another question in mind, let say i want to have TDP142 cascaded (2 x TDP142).

I wonder about the connection of pins below :

• AUXp
• AUXn
• HPDIN/RSVD9
• SNOOPENZ/RSVD8

and is capacitor needed between TDP1's OUTDP and TSP2's INDP (the data signal lines)

Can you provide me example of the schematics ?

Thank you.

Desmond

Desmond

Please see below block diagram for the proposed solution.

SNOOPENZ can be pulled high to 3.3V through a 1K to disable AUX snoop

AUXP can be pulled down to ground through a 100k resistor. AUXN can be pulled up to 3.3V through a 100k resistor.

Thanks

David

Hi David

Three questions

1. If i want the AUX to be monitored by the TDP, can i connect the AUX like what HPD do ?

2. Is TDP142 pin compatible with any of the DPv1.4 re-driver and DPv1.2 re-driver 

3. My simulation shows I have loss of 30 dB from DP sink to DP source, do you have any recommendation ?

Thanks you.

-Desmond

Desmond

1. The TDP142 snoops the traffic on the AUX bus for power management. It snoops native AUX writes to DisplayPort sink’s DPCD registers 0x00101 (LANE_COUNT_SET) and 0x00600 (SET_POWER_STATE). TDP142 disables/enables lanes based on value written to LANE_COUNT_SET. You could connect the AUX line like the HPD, but you have to be careful with the loading on the AUX bus to make sure the loading does not cause data corruption such that the TDP142 incorrectly disable the wrong number of lanes.

2. Which DP1.4/1.2 redriver are you referring to?

3. What data rate are you seeing the 30dB loss? You may need more than 2 DP142 to compensate for the 30dB loss.

Thanks

David 

David

2. I try to search on TI products to see if any 1.2 redriver have pin compatible with it, but i do not found.

3. i'm looking at the nyquist freq, 4.05GHz.

Let see this scenario:

Will it work if i set as below :

TDP142 (B) = 14.4dB

TDP142 (A) =  8.6dB

Desmond

1. With DP142 supporting up to DP1.4 8.1G, DP142 provides backward compatibility and would work as a DP1.2 redriver as well. 

2. How much is the loss at board B? Is it possible to put a DP142 on board B?

Do you have a DP142 EVM that you can use to evaluate the DP142 performance with this flex cable?

Thanks

David

Hi David,

the loss in Board B is around 15dB, cable loss is 2dB, this is checked using simulation tools. Unfortunately i do not have spaces to put a DP142 on Board B.

I do not have DP142 EVM on hand. 

-Desmond