Products Mentioned

ADV7611 Recommended for New Designs
The ADV7611 is offered in automotive, professional (no HDCP), and industrial versions. The operating temperature range is -40°C to +85°C. The UG-180 contains...
Datasheet
ADV7611 on Analog.com

ADV7611 LLC Amplitude

MatthewDSE

on Jun 29, 2021

Greetings,

We are performing a board bring up on a new design using the ADV7611. With this part we are receiving HDMI video and outputting 24-bit RGB from the CMOS. The HDMI input is detecting correctly, but as we were checking the outputs we noticed that the LLC output is swinging between -3V to 5V. The other sync signals HS, VS and DE all seemed normal at 0V-3.3V se we can't explain this. Also, the measured frequency of the LLC is what we would expect for the 720p60 input.

So far we have tested multiple boards and they are behaving the same. Below is a screen shot that shows our scope capture. We checked the scope against its reference clock to verify the probe was accurate and then re-tested the LLC. I have also tried lowering the drive strength, but this had no affect.

Should this signal be 3.3V like the other syncs?

Thank you,

Matthew

0 FormerMember

on Jun 29, 2021 8:38 PM

LLC for 720P60 should be 74.25MHz

I assume you are using a 10x scope probe with >=200Mhz and all the scope filters are turned off.

0 MatthewDSE

on Jun 30, 2021 1:09 PM in reply to FormerMember

Hello GuenterL, thank you for for the quick response. Yes, I have confirmed the probe is set to 10x, with DC coupling and without any filters set. We reset the scope after the initial measurement to re-calibrate it against the reference output clock because I thought a setting might be off.

On our board we have a 0 ohm series resistor (at R312 below) with the LLC clock. We have also tried depopulating this to disconnect it from the video sink and saw no change in the signal. The "LLC" net that is shown was a mistake in the schematic and is not connected anywhere. We verified this in the PCB also.

Please let me know if there is any other information I can provide you.

0 FormerMember on Jun 30, 2021 1:57 PM in reply to MatthewDSE

Normally it is good practice set those series resistors to 22-33 Ohms to match the PCB impedance.

You might also want to look at the DR_STR_CLK settings. Overdriving the pin can do weird things on boards. Try a lower setting.

All the DR_STR... settings control the drive strength of the various output pins.

What I would try.

1) install a 22 Ohm series resistor

2) Lower the drive strength to lowest.

3) Put a 50 Ohm load to ground at the end of the trace. Fast outputs work better driving into a load.

Make sure the scope bandwidth is 5x (3rd and 5th harmonic) the fundamental frequency of the clock you are trying to measure. For 720p60 this would be 74.25 MHz x 5 = 371. I think your scope is but I can't really tell from the image. (A 100Mhz scope measuring a 74MHz square wave would just show a sine wave)

When I check my boards I see a rounded off square wave.

If you figure out the sine wave issue you can also adjust the clock phase for better setup and hold times at the sink end. There's several post on this when you get to that point.

In practice you determine the DR_STR... and Phase adjust on the first new board and then use the same setting s for all other boards. It can be board dependent. (I have two different designs using the ADV7612 and they each have there own unique settings) Let me know if you need help with this.

0 MatthewDSE

on Jun 30, 2021 5:18 PM in reply to FormerMember

Hello GuenterL,

Thank you again for the detailed suggestions. We added a 22 Ohm resistor in series, a 57 Ohm load and turned down all the drive strength outputs. The amplitude dropped considerably with these changes and is now about 3Vp-p. The signal is still dipping below reference by a volt, but it is not nearly as bad as before. What do you think of the signal now?

The clock waveform is actually round on our setup too after adjusting the scope period and should be easier to look at. I will take a look at those discussion on adjusting the phase next.

Thank you again.

0 FormerMember on Jun 30, 2021 8:55 PM in reply to MatthewDSE

looking better. Some of the valleys actually have flat bottoms which indicates this is the true ground level. With the 57 Ohm load you may need to up the drive strength. Why it appears to be going 1V below ground? I don't know. I'd have to have your set up in front of me to see what's really happening