We want to implement a system with PAL video output or VGA monitor.
1. For a PAL solution we want to use the ADV7125 for digital to analog conversion and the AD724 for analog to composite video. We saw in the datasheet of the AD724 (page 11) that it may be possible to use this solution because there is an example circuit using the ADV7120 (similar part to the ADV7125).
Our question is about the voltage level of the green output of the ADV7125.
In the datasheet of the ADV7125 you can see that BLANK level of the green output is around 0.2V and WHITE level is around 0.9V (page 11). The AD724 only supports the standard 0V for BLANK and 0.7V for WHITE. Can we use ADV7125 in conjunction with the AD724 although the level is not correct? What will the consequences be? Are there any workarounds?
We don’t understand how it is possible to use the ADV7120, for example, with this level of voltage.
2. For the same reason for VGA, can we use ADV7125 to generate the green color although the level is not the standard also?
What will the consequences be?
The 4FSC clock looks like the most likely suspect -- the LTC6904 frequency error can be 0.5% to about 1% (+/-0.75% max initial frequency error), which is too large. The availability of off-the-shelf 4x PAL subcarrier crystal oscillators is slim to none, but 4x NTSC subcarrier crystal oscillators are available off-the shelf. For an immediate test, can you use a lab signal source for the 4FSC clock?
Back in the '70s (or sometime thereabouts), an attempt was made to standardize the (NTSC) Hsync to colorburst phase (also meaning that these were synchronized), but I don't think it was fully adopted as a standard. You should be okay without Hsync-to-colorburst synchronization, but the colorburst/color subcarrier frequencies must be much better than 1% -- the tolerance is on the order of +/-10 Hz for NTSC.