We have recently taken a PCB with the ADV7610 HDMI Receiver (BGA version of the 7611) for radiated emissions testing and have found problems similar to what Padmanathan has described. We have significant emissions radiating from the shielded HDMI cable in our system. We are trying to meet CISPR22 and FCC Part 15 Class B emissions requirements and are using a commercially available HDMI source that claims Class B compliance to drive our device. However, we are failing the test by a significant margin, with levels depending on the length and type of cable used.
Our PCB layout is quite simple, with high speed ESD protection devices placed behind a Type A HDMI connector and a total etch length of ~620 mils between the connector and the ADV7610.
The frequency of the offending emissions is dependent on the settings of the HDMI source. Our goal was to accept up to 1080p60 resolution in which case the fundamental clock is 148.5 MHz and its associated harmonics show up as radiated signals (594MHz being the worst). I did a quick verification with a current probe and found that setting the HDMI source to 480p60 did produce an emission at 81MHz on my cable, although I haven't measured it in a semi-anechoic chamber.
One of the areas of the ADV7610/7611 that is not well documented is the method of terminating the TMDS pairs. Is there a technical document that details this subsystem?