TDP142: TDP142IRNQT Re-driver Functional Debugging and SI Test Abnormality Resolution

Rannie

TDP142 is a linear re-driver and it does not have a TX pre-emphasis control. It only has RX EQ control. To change the pre-emphasis would require the DP source to change its TX pre-emphasis.

Keysight test setup ask for HBR2 preferred level setting with and without cable model, what level did they pick?

Can you please share their fully compliance report, schematic, and layout?

Thanks

David

Hi David,

Thanks for your reply. 

Attachment is the SI report. Customer HBR2 level setting, sch and layout will be posted later.

https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/138/DP1.4a-5_2C00_4Gbps_5F00_22ohm-_2B00_7.5dB.7z

BRs,

Rannie

Rannie

The report shows lane 2 is the only lane that has consistent failure while other three lanes are passing with sufficient margin. 

Before changing the TDP142 EQ, I would double check the setup, can they for example swap the lane 2 connection to a different scope channel and see if it makes a difference on the testing result?

Thanks

David

Hi David,

Thanks for your suggestions. I will let customer check the setup and swap the lane 2 connection to a different scope channel and see if it makes a difference on the testing result. But customer also mentioned that they found the VTX_MEQ_DELTA (TX_EQL3/TX_EQL2) - PLTPAT eye pattern test for Lane 1 and 3 fails. So, it may not be the issue of setup. But I will check with customer.

I sent customer layout and schematics through email. Could you please help to check and help to give the further debugging suggestions?

Below is customer HBR2 level setup:

BRs,

Rannie

Rannie

One of things I noticed is that DDI0_DDC_AUX+/- are not AC coupled, and do not have 100k pullup/pulldown as required by the DP spec. Do they have the capacitor and 100k pullup/pulldown somewhere else? 

The VTX_MEQ_DELTA (TX_EQL3/TX_EQL2) is calculating the delta between the two different pre-emphasis levels. If you look the passing case, you can see pre-emphasis 0 is 15.9dB and pre-emphasis 1 is 13.167dB, resulted in the delta of 2.733dB which passed the test.

If you look at the failing case, then you can see pre-emphasis 0 level is 14.956dB and pre-emphasis 1 level is 14.934dB. so there is almost no delta between the two levels. 

The pre-emphasis is provided by the source, and changing between the two levels comes the DP sink controller over the AUX bus. So either the source is not receiving the request or source received the request but not able to change the level correctly. If the AUX does not have the capacitor or the resistor, then this could explain the AUX communication issue.

What is the EQ they are currently using? Based on the layout, I estimate they should use EQ around 5dB.

Thanks
David

Hi David,

Thanks for your kindly reply. Customer added 10k PU and PL resistor on AUX through flywire when test but no AC couple. 

Update the latest test results, the SI test passed if they adjusted the EQ to 7.5dB and adding series resistors of 20Ohm. 

I still have some confusions: why adjusting EQ and resistor solves the problem? If it's a problem with switching between pre-emphasis level 0 and 1, will modifying the EQ and the series resistor solve this issue?

BRs,

Rannie

Rannie

Does the test pass with 10k PU or PL? They should also use 100k PU and PL on the AUX bus along with the capacitors. 

Without the 10k PU and PL, does the SI test consistently pass with EQ of 7.5dB and 20ohm series resistors? For example, if they ran the SI 5 times, does it pass with the same result? 

Thanks

David

Hi David,

The test pass with 10K PU and PL.

With 10k PU and PL, the SI test pass with EQ=7.5dB and 20ohm series resistor.

BRs,

Rannie

Rannie

With the 10k PU and PL, have they tried the SI test without the 20ohm resistor, but with the EQ of 7.5dB?

Thanks

David