We have a board with 2 ADV7188 and 2 ADV7393 that was working fine for our previous project, but is giving us problems for a new project: after roughly 2 hours of use, the ADV7188 inputs stop working, so far "forever".
We use CVBS inputs (PAL or NTSC), with the "same" register initialization as on the eval board (what are the ADI recommended write doing? They don't seem to be documented), and almost the same schematic as that board (with additional ESD diodes to the +3.3V and the 0V, and R of 18+56 instead of 19+56). These parts are really close to the ADV, but separated by a via (going to the AIN1 we use, so we changed that register, to the same value on both projects).
Register 0x10 is indicating that the ADV doesn't lock (any more), but we also have a few more deltas when we read the other documented registers (hex values):
Reg OK KO
10 4d 6a
15 0d 00
41 17 11
49 07 01
50 02 00
51 17 01 (also sometimes read back at 07 when OK, and at 24=reset value when KO)
Does reg 0x51 ever indicates the locked line counts? It'snever written at 7 nor 17 (nor 1)
Any other undocumented register we should read/compare?
The signal arrives correctly onto the board, but ramps up (for each video line) at the ADV input pin, after the 100nF cap. So we suspect a clamping issue. We tried the 4 digital clamping configurations, with the same results (no lock).
Is there any way to check if the current sources died? What could kill them ? Like having to deliver way too much current?
Any other possible reasons?
It wasn't the case when the last board stopped working, but can the ADV7188 be connected to an active video signal before its own power source is switched on?
Please send us any ideas you have, all hints may help us !