Something pertaining to ADV7510

Hello,

So this was working for ADV7510 and now it is not for ADV7511?

First step is to see if the ADV7511 is recognizing the video format.  Are you sending the CEA format for 720p/60 with DE/H/V from your FPGA? Have you read register 0x3E in the ADV7511 to see if the VIC code matches the one for the format you are sending?

Dave

Hi~Dave:

            Yes, it is. we sent the video format with DE/H/V. (We follow

the CEA_861_D standard)

            I read the register 0x3E and get 0x10 (I think the chip

recognize 720p/60hz video format).

            Besides, I have a question that are we really necessary to

enable HDCP

            function in the chip, then we are able to pass our video

patterns to TV?

            BR

Joseph

On Wed, Oct 12, 2011 at 3:56 AM, DaveD <analog@sgaur.hosted.jivesoftware.com

Hi Joseph,

Yes that means that the ADV7511 recognizes the format. What voltage are you seeing on the TMDS outputs?

You do NOT need to enable HDCP unless you want to protect your content.  The TV doesn't care if you encrypt or not.

Dave

Hi~ Dave:

            Sorry, my typo~ my chip is ADV7510.

            The voltages for TMDS 4 pairs are around 3.3V. And I can not see

the clock channel

            drives the clock. When can I see the clock drive out ? (After

setting a certain register?)

            I check the register 0xA1 -- > 0x00, it shows that all the TMDS

channel are power up, right?

            And  0x42 ---> 0x60, shows Rx sense detected.

            Still don't understand why it doesn't show the clock and data.

Joseph

On Thu, Oct 13, 2011 at 5:39 AM, DaveD <analog@sgaur.hosted.jivesoftware.com

Hello Joseph,

One of those pairs is the clock, the other 3 are data.  If you are seeing that voltage on the pairs-- it is sending video.  What display are you connected to?  Have you tried other displays?

From what you are telling me, ADV7510 is working properly.

Dave

Hi~Dave:

        I know the TMDS has 4 differential pairs, and one of them is clock.

        However, I still can't see any clock drive out.

        The display I used can displays 1080i (Sony, BRAVIA), it can show

video well with DVD player.

        I find the register 0x44 --> 0x79. Showing N_CTS, Audio smaple, AVI

InfoFrame, Audio

        InfoFrame enable. Should I have to fill out Audio InfoFrame? I

filled out the AVI InfoFrame,

        but still no picture pattern on my display.

        Actually, I followed the instruction of quick start in the "AD7510

programming guide".

        to fill the value into individual registers it says. However,

doesn't work. Do you have

        a clear initial procedure document or reference code? So that I can

come out soon.

        BR

Joseph

On Thu, Oct 13, 2011 at 10:29 PM, DaveD <

Hi Joseph,

Clock drive out of what?  Are you talking about clock on something other than the TMDS pairs?  If you are seeing voltage on all the TMDS pairs, there is a clock.  What does the TV say?  Sync out of range or no input?

The TV working with a DVD doesn't really say anything-- you aren't sending the same thing as a DVD player which would be encrypted and probably 4:2:2 YCbCr in HDMI mode.

There is a .zip on the ADV7612 page (http://ez.analog.com/docs/DOC-1751) with scripts in a text file-- it'll be in the ADV7612 section but it has scripts for 7511-- which should be fine for your case.  The writes to i2c address 0x72 are the ones to look for.

The first section, second script is probably closest to what you are doing.  The section that matters to you is:

This might not be exactly what you are looking for but it's probably a better start than what you have.

Dave

Hi ~ Dave:

        Thanks for your information. And I will try it next week. and see

what it goes.

        Besides, what I mean on TMDS clock is that I can not see the clock

on the TMDS clock

        pair, sorry to confuse you.

        BR

Joseph

On Sat, Oct 15, 2011 at 3:47 AM, DaveD <analog@sgaur.hosted.jivesoftware.com

Hi Joseph,

Ok.. good luck!  The TMDS clock encoded on the differential pairs doesn't really look like a clock signal if you probe it with a normal single-ended scope probe.

Dave