DP130 failure in Pre-Emphasis Level in compliance test.

Hi Peter,

They can try adding a boost to PE, and expect PE L2 gets higher than PE L1 does.

If fail is present in other lanes, the failing test is the same? (PE delta PE2-PE1, try 2?

Could they verify DP130 is configuring correctly to 5.4Gbps?

Regards

we see the issue duplicate in EVM as well.
Sent the detail report by mail

Hi Peter,

 

From the test report, it seems sometimes DP130 is not configuring correctly.

Could you verify DPCD registers have the right values for each test?

Could they configure DP130 by I2C and rerun the tests?

 

Regards

Hi Mr. Garcia,

The failed items about Pre-emphasis Level Delta are passed by changing register 15h's value to 0x10 eventually. Can you let us know why doesn't TI set Boost=10 as default?

Regards,

S.P. Lin@Adlink 

Hi,

In most of the application nominal pre-emphasis is enough to pass testing, that's why BOST=10 is not the default value.

Regards

Hi Mr. Garcia,

Here are further questions need your help,

1. The board set boost=10 to pass the pre-emphasis level delta has approximately 2-inch routing trace length from DP130 to DP receptcle. What about the trace length of the most applications?

2. Is there any side-effect while boost=10?

3. DP130's spec. describes that 10% will be increased once boost=10 is set and the unit is dB. Can you explain the definition of the 10%? We've tried to observed the before/after change but it seems the effect of boost is not obviously on the waveforms.

Regards,

S.P. Lin@Adlink

Hi,

In most application the redriver is placed as close as possible to DP connector.

Should be no side effect of setting BOOST=10b, the increase is very small

The change is 10% in dB, for example in pre-emphasis level0, PE=3.5dB, setting BOOTS=10b(10% dB) would result in 3.85dB

Regards

Hi Supporter,

As I mentioned before, to set the boost-related register to 10b can make the failed pre-emphasis level delta be passed. However, after running the complete compliance test, we found another test item called "non-transition voltage range measurement" will be failed and confirmed that the failed is directly caused by setting boost to 10b. I mean, non-transition voltage range measurement is always passed until setting boost to 10b. On the other hand, we've checked the relationship between pre-emphasis and, output voltage swing and EQ. And it appears both of VOD and EQ are the functions of pre-emphasis, that is, they are respect with pre-emphasis level. So, even changing EQ's level with I2C  wound never affect the pre-emphasis level and also, during the compliance test, link training is enabled and pre-emphasis level is fixed at level-2 by the auto test program therefore output swing is also decided.

As my understanding, either to increase the pulse level of transition bit or to decease the voltage level of non-transition bits will enhance the level of pre-emphasis. Thus, I am considering the failed of non-transition voltage range possibly is a side-effect caused by boost. We also tried to set the boost register to 00b but it still cannot make the two mentioned test items are passed. As a result, I'd like to get your help to see if there are any other actions can be taken to make the two mentioned test items passed.

By the way, we've also tried to increase the level of supplied voltage to around 3.465V but the result was still failed (although the result of pre-emphasis level delta is significantly improved). The IC we are using is SS type.

Regards,

S.P. Lin@ADLINK

Hi,

Thanks for reaching out again, I'm analyzing this information with the team, I'll come back when I have suitable data.

Regards

Hi SP Lin,

Do you have a FAE that could escalate this issue?

Regards

SP,
We use unigraf DP compliance test box to perform DP compliance test. the kit is as the same as professional testing laboratory using. The kit will communication with the GPU and do the correct pre-emphasis setting in compliance test.
www.unigraf.fi/.../dpr-120

Hi Peter,

Do you remember we confirmed this before? The aux controller we are using is the same as yours. I don't think that is a problem because the test result of items we are talking about can be changed as long as we change the setting of DP130's registers. And the pre-emphasis level is fixed to a specific level by the demand of test program no problem. However, the problem is: we found one failed item and others passed, after we changed the setting of DP130 for the failed item via I2C, this failed item passed but cause another item failed. So, we are asking whether there is any action can be taken to solve the both. And also, maybe TI can explain the detail about the boost function for us.

Regards,
S.P. Lin

Hi Peter,

Thank you for placing HelpMe ticket, support will continue there.

Regards

Hi Moises,

Are you still the owner of this issue? We are eager to know TI's verification result since we have not only a project cares about this.
So, we need your help make the progress more efficient.

Your kindly help will be appreciated.

Regards,
S.P. Lin@ADLINK

Hi SP,

This issue has been escalated, I'm no longer in charge of the escalated ticket, ask you FAE for HelpMe ticket status.

Regards