The resistor value of hdmi_extswing of DM8168

What value are you currently using?

The resistor should be 5K9.

Some early EVM boards and schematics had incorrectly installed 698R.

BR,

Steve

Hi Steve:

currently, we use 5.9k ohm. The fail item is rising edge  hits the mask 7723.HDMI.rar

not sure if there is any registor to adjust the driving strength or can we adjsut  the resistor value of HDMI_EXTSWING 

The resistor will change the characteristics of the eye, yes, but you need to be careful since the value is chosen so that the output drive characteristics should meet the HDMI specification.

We have done extensive characterization on the DM816x family and found that it is good all the way up to about 225MHz.

One possible cause of this eye closing is your board layout. Have you made sure that all board impedances are correcly matched for the HDMI signals? (few/no vias, solid reference plane under the traces, correct trace widths, inter trace spacing, etc...)

Do you have any other components on the signals (i.e. ESD etc...)? Are these components correctly placed on your PCB (i.e. no stubs etc...)?

The other thing to try is moving the mask. The HDMI specification does allow you to move the mask in order to make it fit.

BR,

Steve

Hi Steve:

We keep the impendances to be 100ohm and only two vias on the trace. But, the routing length is up to 4300mil and we also add the ESD components and CM choke on the routing trace. Maybe these factors will cause the loss.

Except for adjusting the resistor value, is there the register improving the drving strength or voltage swing ?

We also fail on TMDS-VL item. The spec of VL is  2.7V<= VL <=2.9V, but our test reuslt is 2.69V. 

 If there is no register to improve the TMDS characteristics, mabye we will try to adjust the resistor value.

Except for TMDS characteristics , is there any side effect if we change this resistor value ?

It is difficult to really determine other side effects, but I think you should be OK making relatively minor adjustments to the resistor value. I would not change it by more than 10-15%.

There are no registers for changing anything else.

Don't forget that the 100R requirement is differential, which translates to 50R single ended.

BR,

Steve

Hi Steve:

Do you have test report  regarding the HDMI test item 7-2 TMDS-VL ? I am not sure the 5.9K ohm of HDMI _EXTSWING if can pass this item on EVM

This item is related to voltage swing. We fail on this item for HDMI CTS.  seems that the voltage swing is too large.

We test two boards. The symptom of these two boards is not consistent.

For board A, D1+ is 2.705V and D0- is 2.675V, but for board B, D1+ is 2.6725V and D0- is 2.635.

The  D1+ is large than D0- on the board A, but D1+ is small than D0- on the board B.

If the layout and schematic is similar , the result should be similar.

But, the test result shows the different TMDS channel has the different test result.

Do  you have any idea about this phenomenon.

The attached is our test result.2287.TMDS_VL.docx