Hello,
I am using the ADV7280A with a STANAG 3350 Class C (NTSC RS-170A timing equivalent) composite input and 8-bit embedded BT.656-4 output.
Symptom:
The first active row of one field shows blinking colored dots (unstable chroma) on saturated colors such as SMPTE bars. Luma appears stable. Row 1 and the following rows are clean.
Isolation:
We checked the raw BT.656 output directly with an FPGA Integrated Logic Analyzer (ILA). The ADV7280A outputs 487 active lines, and the chroma variation is already present on the first active line of the raw byte stream before our FPGA crops it to 720x486 lines for the downstream modules.
Attempted fixes:
We tested several documented registers and EngineerZone-recommended initialization sequences, including 0x27, 0x2B to 0x2E, 0x3D, 0x38 to 0x39, and 0xEB to 0xEC, without any change to the first-line behavior. We have also checked with different video sources.
For reference, our baseline ADV7280A configuration is:
0x00 = 0x00
0x02 = 0x14
0x03 = 0x0C
0x04 = 0xBD
0x1D = 0x64
0x0F = 0x00
0x3A = 0x07

Questions:
1. Is this a known ADV7280A behavior for STANAG Class C / non-broadcast NTSC sources, specifically on the first active line of a field?
2. Could this be related to NTSC field-boundary / SCH burst phase behavior, or to how the ADV7280A handles chroma decoding / comb filtering on the first active line?
3. Is there any documented or undocumented register sequence that can improve first-line chroma stability, or is the recommended approach simply to mask this line downstream?
Thanks in advance for the support.
Kind regards,
Jonathan