[marqs]

The exact source dot clock is irrelevant for stability, but line length has to be consistent due to line-locking behaviour. Can you still test it in 240p passthru mode and report if the green LED keeps turning off? If so, then there might be a mode detection problem instead of PLL locking issue.

[marqs]

So 0.xx still detects sync which at least indicates the board has not been damaged. I’d use something other than NES/SNES (unless dejittered) and/or a display directly to test stability. Did you test that nothing is shorted on the resistor side?

[marqs]

Uzebox video timings are SW-generated so perhaps they are not fully consistent across lines/frames, potentially causing PLL locking issues (indicated by the LED turning off). Either that or noise on the sync line. You could try adjusting “Analog Sync Vth” and/or setting “ADC PLL BW” to ultra low under sync options.

[marqs]

I recommend downgrading to 0.xx firmware and testing if that still works (no need to revert the mod).

[marqs]

Your safest bet would be a cable which has active sync combiner + 75ohm driver powered from VGA pin 9 or external 5V. The cable/adapter should bypass RGB lines without any additional circuitry. The AV in/out adapters are also back in stock although not in large quantity due to challenges sourcing scart connectors.

[marqs]

Is this diode the correct way round. Thanks

Doesn’t seem to be.

[marqs]

It looks like the diode is still installed the wrong way.

[marqs]

Have you tested its game mode with any other 1080p50 source? You can also try fine-tuning refresh rate with “Off (preset Hz)” and via “Output opt -> Adv. disp timing” if that makes a difference, but overall it sounds like a TV issue (especially if you can replicate it in A-LM mode).

[marqs]

Which fw version you have, and have you already tried adjusting sync opt -> analog sync vth? Also, is the RGB mod wired for sync with jumper J8 closed?

[marqs]

There are shadow mask presets under Post-proc->Shadow mask that might be what you want.

[marqs]

Both the audio and audio metadata are sent out continuosly so the card should be able to restore audio after any interruption. It’s possible to add compatibility option for a longer HDMI handshake when a mode is changed which might help, though. Meanwhile, I’d recommend also to try to report the issue to Elgato support.

[marqs]

I’m not sure if it’s related, but you have unofficial board which may or may not match official v1.7 schematic. You should at least check that pin25 of TVP7002 is not grounded or connected anywhere else on the PCB.

[marqs]

So you’re looking for 1920×1200@85Hz  output, presumably with CVT timings? That’s 281MHz which is quite high but probably still works as digital output (and based on some reports possibly also for extra_av_out).

[marqs]

This is the same issue which is reported here and which I’ve occasionally bumped into too, but it’s not easy to reproduce so I have only a hunch where the issue could be and this would need a larger undertaking to debug. You can unplug the power for a few seconds and everything should look normal once you restore power.

[marqs]

The premise in LM modes has been that they only apply integer multiplication in X and Y directions. That works fairly well for analog content where aspect adjustment can be made by controlling sampling, but has its limitations with digital sources and ADV7280 which has fixed samplerate. A simple nearest neighbor / linear horizontal resampler could be added for LM but I’m afraid the end result in many cases would not look good enough to justify the effort. Does 1080p work at all on your TV? With A-LM the signal should be spec-compliant aside from refresh rate which is set by the source.

[Author]

Posts