New Retrotink 2x v.3.2 and Commodore 64 interference issue

NewHome Forums RetroTink 2 Products New Retrotink 2x v.3.2 and Commodore 64 interference issue

Viewing 15 posts - 1 through 15 (of 24 total)
  • Author
    Posts
  • #26372
    Kalel80
    Participant

    Hi,

    I received Retrotink 2x today with new v3.2 board and new firmware, and I have now tested that with different C64 coputers using S-Video. I Also tested that on my C64 Reloaded Mk2 board. I noticed immediately that there is disturbing interference in picture. There are moving vertical / diagonal lines going on screen and these are much noticeable when screen is full colored. (Example C64 blue screen). My S-Video cables are quality shielded cable and work fine with my CRT and older S-video LCD television. Also picture is very pixelated comparing older Retrotink x2 with v3.1 board and same setup.

    Vertical/diagonal lines

    With same setup and earlier Retrotink 2x v3.1 board I don’t have that issue. I borrowed that from my friend and that’s working perfectly and it shows perfect picture. Picture is little softer with old board but no interference at all. I have now both devices at my home and it’s easy to test and compare these. I noticed that there were changes made to circuit and also ground plane is different.

    Has anybody else noticed these kind issues? Is the problem that this new board is not good with c64 or is it new firmware? Or is the problem only in my device?

    #26376
    mikechi2
    Participant

    Matt and I are looking into it by cross-checking both the old board/new board with both the old/new firmware. Could you also post a picture of the output from the older 3.1 design? I think it might be related to some of the gain settings. The ground plane is identical — the digital and analog power nets were actually pulled away from each other in the new design, which is what I think you’re seeing.

    I have noticed a similar effect in the past with DVD sources, due to cross-talk between the ADV7280 oscillator and the video input, but fixed it early on with extra decoupling capacitors.

    PS – image might be a little more pixelated because James-F was kind enough to help me figure out how to increase the luminance bandwidth on S-video. I guess we’d hoped that would be a good thing over a softer image 🙂

    #26377
    mikechi2
    Participant

    Also if you can, try another input source (non-C64 related) that gives a similar solid color over s-video and let me know how that goes. That would be super helpful!

    #26385
    Kalel80
    Participant

    Hi Mike and thank you for your quick reply 🙂 I now tested Retrotink 2x v3.2 S-Video on my Neo Geo CD and there was also little bit same moving interference but nothing major. I also noticed that there is a minor interference with earlier v3.1 board too on Neo Geo Cd. So difference with those boards on my Neo Geo Cd S-Video is quite small. Biggest difference picture quality and interference coming with Commodores.

    Here is a few pictures. Difference in picture betweeen old and new board is very noticeable. Coputer on those pictures is same PAL 50hz Commodore 64c with lumafix board. (Lumafix doesn’t seem do much that interference if I tweaking it.)

    Two upper pictures are on v3.1 board and lower pictures are v3.2

    I also tested Commodore 64 computer on composite video and I noticed same interference what I can see on S-Video.

    Thank you that you figuring out this. Cheers 🙂

    #26386
    James-F
    Participant

    @kalel80
    I have tested v3.2 with various sources over s-video and the image is very clean and noiseless.
    The S-Video mode on v3.2 is transparent and will not automatically adjust the comb filtering according to input quality, hence you see noise if it exists in the source.
    Previous v3.1 automatically adjusted the comb filtering even on S-Video to clean noisy luma signal, hence you see loss of reolution in v3.1 over noisy s-video signal.
    S-Video sources do not use comb filtering because there should not be any interference between luma and chroma,, but if there is interference you will see noise on the luma signal.
    The C64 is very noisy so maybe this LumaFix64 Mod doesn’t do a good job cleaning the signal?

    As for the difference in gain;
    Looks like the Luma signal is very weak and out of spec with your C64 s-video board.
    An actual CRT would not adjust the gain automatically like v3.1 was doing, and will look the same as you see with v3.2.
    The autogain on the ADV7280 will amplify the input signal till the sync reaches exactly 300mV in amplitude regardless of how strong the luma already is, this might clip the luma signal since its amplitude is completely ignored, hence it is disabled in v3.2.

    I also tested Commodore 64 computer on composite video and I noticed same interference what I can see on S-Video.

    So you say it is the C64 source that is noisy, should the RetroTink2x fix this very noisy signal?
    Please try video composite mode (3 leds lit), this should process composite input exactly like v3.1, aside from disabled automatic gain.

    EDIT:
    I just tested my PAL consoles over S-Video and I see zero interference with v3.2.

    #26388
    Kalel80
    Participant

    Thanks. Okay maybe it could be that v3.1 board automatically adjusted comb filtering just little bit cleans c64 VIC-II chip noisines. I also know that it’s noisy and out of specs.

    I tested this new Retrotink 2x v3.2 board on my many PAL 50HZ C64 boards including new Reloaded Mk2 board and noise is still there and picture is bit distorted. With Lumafix64 or without. Reloaded Mk2 board seems to be little better, but picture is not so good comparing v3.1 board. Reloaded board was connected with thats original low ripple psu. This noise appears specially on c64 blue screen and it’s bit noticiable when I example plug my pi1541 drive on it. I haven’t tested real disk drive. But noise also shows without drive. Pictures is taken without any drive connected. It’s weird that I don’t see moving lines on grey screen what I have on my other kernal rom in Reloaded Mk2 board.

    C64 Reloaded Mk2 board S-video Grey start screen

    My C64 S-Video cables have 330ohm resistor on chroma line and Reloaded Mk2 board have this same resistor in S-video output connector. This is what has been recommended for LCD screens. I could take this resistor off my cable and test without it?

    I tested C64 Reloaded Mk2 board composite video on v3.2 Retrotink and there was no moving lines but little bit dis distortion.

    #26389
    Kalel80
    Participant

    I took picture on C64 Reloaded Mk2 with composite and 3 led lit and picture is not bad but it is pixelated on blue screen and minor wobbling. Hard to get good picture on my iphone. My Grey kernal screen is perfect and has no any raster effect and wobbling.

    3 leds on composite video

    #26391
    James-F
    Participant

    This 330 ohm series resistor attenuates the chroma line and not related.
    Th Commodore 64 is just a very noisy source and has chroma bleeding into luma, or other design related noise.
    You say that connecting a “pi1541 drive” to the C64 makes the signal worse?

    There are a lot of discussion about this on many C64 forums.
    https://www.lemon64.com/forum/viewtopic.php?t=40663

    The RetroTink2x cannot fix noisy signal.

    #26392
    Kalel80
    Participant

    Ment to say chroma line. So I have 330 resistors in chroma line.

    #26395
    Kalel80
    Participant

    Yes wavy lines shows a bit more when I have pi1541 drive connected. It’s maybe adding more noisines on a board power line and that also have effect on a VIC-II chip.

    #26402
    mikechi2
    Participant

    Matt has been testing and we’ve determined that the issue is due to the new settings.

    As James-F mentioned, we set the S-video luminance filter for higher bandwidth, thus preserving more details for clean sources – which should be virtually all S-video sources using proper cabling.

    Unfortunately, the C64 has an inherently noisy output. While the new firmware’s image is more “true” to the signal, it also means that all that noise is now being pass straight to the image.

    Matt and I are figuring out the best solution. Right now, I have a test build that has another s-video mode that turns the luma LPF back on.

    EDIT:

    I sent Matt a copy of the firmware with an additional mode for S-Video (NOISY). And once again, a thank you to James-F for helping me decipher all these various registers.

    #26407
    Kalel80
    Participant

    Okay. Sounds good 🙂 Thanks to all of you. I really appreciate the time and effort you put into this.

    #26412
    mikechi2
    Participant

    Not at all. Thank you for bringing this up to our attention and for doing some preliminary research! We’ve narrowed it down to firmware settings and are just trying to do final optimization. Thanks for your patience!

    #26413
    BuckoA51
    Keymaster

    The C64 is very noisy so maybe this LumaFix64 Mod doesn’t do a good job cleaning the signal?

    Lumafix is really for fixing the vertical jailbar issue caused by clock signal interference but the C64s output is, as you suspect, just generally noisy and badly designed, we’ve even a prototype new encoder for C64 here and it can’t fix the issues entirely.

    On my 64s I’m seeing really bad horizontal jailbars with the newer RetroTINK f/w, even on the proto new encoder machine. We’re close to a fix though the latest firmware I tested (with an additional mode for S-Video (NOISY)) cures the issue for me.

    #26418
    mikechi2
    Participant

    Matt – it occurs to me that a potential fix for the C64 is to copy the ADV7280 filters on a PCB

Viewing 15 posts - 1 through 15 (of 24 total)
  • You must be logged in to reply to this topic.