Quantcast

Page 1 of 2 12 LastLast
Results 1 to 15 of 19

Thread: Horizontal red/magenta stripes through s-video (but not composite)

  1. #1
    Wildside Expert fluxcore's Avatar
    Join Date
    Aug 2013
    Posts
    240
    Rep Power
    10

    Default Horizontal red/magenta stripes through s-video (but not composite)

    Hi Guys,

    I've done many (8 or so) successful svideo mods on various megadrive consoles in the past, and this is the first that I've had troubles with. I've resoldered everything several times, and switched out each of the components to rule out problematic parts.

    Basically, svideo out has horizontal red/magenta stripes through everything (except white colours). Composite looks fine. Here's an image for comparison (top is svideo, bottom is composite):



    The circuit is the usual pin 15->220uf cap ->75ohm res -> chroma, and pin 16->2N3904 base->(emitter) 27ohm res ->chroma. It's installed in a PAL Megadrive 1 VA6.5, with the surface-mount MB3514 encoder.

    I'm using a 4.43Mhz xtal osc to force the signal out as PAL60, as I've done in the past, and never had this issue before.

    I'm very suspicious since composite out looks fine, and the video encoder should use the same freq for colour for each, so I would have expected colour to be fine in both cases. Perhaps it's the luma that's actually the issue? I don't know enough about svideo encoding to really know where to start troubleshooting.

    I'm guessing either the encoder chip is faulty, or perhaps a capacitor near it is dead (although I can't work out how that would affect svideo and not composite).

    Any help greatly appreciated!
    Last edited by fluxcore; 07-18-2015 at 09:19 PM.

  2. #2
    Creator of the Mega Amp Raging in the Streets Ace's Avatar
    Join Date
    Jun 2008
    Location
    Montreal, QC, Canada
    Age
    31
    Posts
    3,737
    Rep Power
    45

    Default

    Reading up on the MB3514 seems to indicate you can directly take Chroma/Luminance from the encoder without any buffer. I suggest simply connecting the Chroma/Luminance outputs with 75ohm resistors in series with 220uF capacitors.
    HATES ATGAMES WITH A PASSION


    Mega Amp: An all-new audio circuit for your Sega Genesis/MegaDrive and clones.

    Note: If you want to contact me on Skype, identify yourself or your contact request will be rejected.

  3. #3
    Wildside Expert fluxcore's Avatar
    Join Date
    Aug 2013
    Posts
    240
    Rep Power
    10

    Default

    Quote Originally Posted by Ace View Post
    Reading up on the MB3514 seems to indicate you can directly take Chroma/Luminance from the encoder without any buffer. I suggest simply connecting the Chroma/Luminance outputs with 75ohm resistors in series with 220uF capacitors.
    Great theory, and you're totally right, I forgot the MB3514 only requires caps and resistors. Unfortunately switching to that didn't fix the problem :/

    I just swapped the colour encoding xtal osc, tried tapping 5V directly from the 7805s, and put a filter cap across the xtal osc 5V and gnd... still no dice :/

  4. #4
    Wildside Expert fluxcore's Avatar
    Join Date
    Aug 2013
    Posts
    240
    Rep Power
    10

    Default

    Okay, well I bought some replacement MB3514 chips in case it was just faulty (why would they bother testing the svideo functionality during mega drive quality control?), but the issue remains.

    I'm assuming that the signals coming into the chip are fine, since composite and RGB output look as expected.

    I'm currently going through schematics trying to decide if I should rebuild the svideo part of the circuit around the MB3514 to match the datasheet version in case that solves the issue :/ The sega implementations differ quite a bit from the datasheet.

  5. #5
    WCPO Agent Helder's Avatar
    Join Date
    Jun 2013
    Posts
    755
    Rep Power
    16

    Default

    Posted a new Svideo schematic on another thread, maybe try that circuit out and see what you get.

  6. #6
    Road Rasher
    Join Date
    Mar 2010
    Posts
    406
    Rep Power
    25

    Default

    You said you wired a 4 pin 4.43 osc, but did you wire it bare ? There was some discussion that the bare signal coming out of a 4 pin osc was too "harsh" for subcarrier purposes and needs some resistors/caps. Helder even made a board accounting for these (can't find the post of course...).

    I svideo-modded a PAL model 2 with a MB3514, but I also replaced PAL master clock so the subcarrier generated by the VDP (in 60hz mode) is the correct NTSC value, and it works fine. Maybe composite hides those defects.

  7. #7
    Wildside Expert fluxcore's Avatar
    Join Date
    Aug 2013
    Posts
    240
    Rep Power
    10

    Default

    I attached the osc out to what I determined to be the equivalent position as on the MD2 mods I've seen - which is after a couple of resistors and caps, so I think it's fine. I've done that bit on another MD1 revision before and had no problems - but of course if they've chosen different cap/res values in places on the revisions then who knows how that affects things...

    Come to think of it, it's not actually magenta stripes through everything, it seems to be an 'every-second row colour shift' or something, blue turns to magenta stripes, yellow kind of turns browny, and red turns sort of orangey. Really seems like the colour just isn't separated out correctly, but why would that only happen every other row?

    I'll take a look at the other svideo circuits. I have tried the 'traditional' MD1 mod (for CXA chip) and now the usual cap+res mod for the MB3514, but I'll give anything a go I guess, this is just annoying me!

  8. #8
    Road Rasher
    Join Date
    Mar 2010
    Posts
    406
    Rep Power
    25

    Default

    Could you tell me the names and values of the 2 resistors before pin 6 ? To compare to MD1 schematic and my MD2 implementation.

    Code:
    sbcr trace--/\/\/-----||--pin 6 mb3514
                 R1   |  |
                      |  |
                      >  |
                   R2 <  =
                      >  |
                      |  |
                      ---
                      |
                  gnd 
    also, did you do other model 1 with mb3514 or this is your only one ?
    Last edited by Ryoandr; 08-15-2015 at 08:44 PM.

  9. #9
    WCPO Agent Helder's Avatar
    Join Date
    Jun 2013
    Posts
    755
    Rep Power
    16

    Default

    Quote Originally Posted by Ryoandr View Post
    You said you wired a 4 pin 4.43 osc, but did you wire it bare ? There was some discussion that the bare signal coming out of a 4 pin osc was too "harsh" for subcarrier purposes and needs some resistors/caps. Helder even made a board accounting for these (can't find the post of course...).

    I svideo-modded a PAL model 2 with a MB3514, but I also replaced PAL master clock so the subcarrier generated by the VDP (in 60hz mode) is the correct NTSC value, and it works fine. Maybe composite hides those defects.
    Post here:
    http://www.sega-16.com/forum/showthr...l=1#post706284

  10. #10
    Wildside Expert fluxcore's Avatar
    Join Date
    Aug 2013
    Posts
    240
    Rep Power
    10

    Default

    Quote Originally Posted by Ryoandr View Post
    Could you tell me the names and values of the 2 resistors before pin 6 ? To compare to MD1 schematic and my MD2 implementation.

    Code:
    sbcr trace--/\/\/-----||--pin 6 mb3514
                 R1   |  |
                      |  |
                      >  |
                   R2 <  =
                      >  |
                      |  |
                      ---
                      |
                  gnd 
    also, did you do other model 1 with mb3514 or this is your only one ?
    R1 in your diagram (labelled R23 on the MD1) is a 332, and R2 (R24 on MD1) is 102, so 3.3k and 1.0k respectively. And I attached my osc output to the point you've labelled as 'sbcr trace'.

    This is the first MD1 with a MB3514 that I've tried to mod, all the others have been MD2s (mainly v1.8, but a v1.0 as well)

  11. #11
    Road Rasher
    Join Date
    Mar 2010
    Posts
    406
    Rep Power
    25

    Default

    just found a VA6.8 pic (with a 3514 on it to boot) and all the resistors / caps that are in the sbcr trace are located on the underside... Did you connect your oscillator there or directly at the encoder pin ? Ultimately, pics would help.
    Also, judging by the pic the names should be the same as other model 1 so what I need is R23 and R24 values, pic wasn't big enough to see them

    EDIT just saw your msg.

    ok, the original value for the CXA1145 is R23=10k and R24=1K
    on my 2 model 2 with a MB3514, the equivalent of R23=4.7k and R24=1K

    You could try to mimic the value of what is on model 2.
    Last edited by Ryoandr; 08-16-2015 at 02:36 AM.

  12. #12
    Wildside Expert fluxcore's Avatar
    Join Date
    Aug 2013
    Posts
    240
    Rep Power
    10

    Default

    Quote Originally Posted by Ryoandr View Post
    just found a VA6.8 pic (with a 3514 on it to boot) and all the resistors / caps that are in the sbcr trace are located on the underside... Did you connect your oscillator there or directly at the encoder pin ? Ultimately, pics would help.
    Also, judging by the pic the names should be the same as other model 1 so what I need is R23 and R24 values, pic wasn't big enough to see them

    EDIT just saw your msg.

    ok, the original value for the CXA1145 is R23=10k and R24=1K
    on my 2 model 2 with a MB3514, the equivalent of R23=4.7k and R24=1K

    You could try to mimic the value of what is on model 2.
    I'm sure I can find or scrounge a 472 somewhere, will give it a shot. My gut feeling is that the problem is on the pin 15/16/17/18 side though. Thanks for the input!

  13. #13
    Road Rasher
    Join Date
    Mar 2010
    Posts
    406
    Rep Power
    25

    Default

    have you tried stock console with just s-video wires ? Meaning console in 50hz and subcarrier coming from VDP.

  14. #14
    Wildside Expert fluxcore's Avatar
    Join Date
    Aug 2013
    Posts
    240
    Rep Power
    10

    Default

    Great thought, unfortunately my only source of Svideo monitor (LCD monitor) doesn't like svideo @ 50Hz for some reason. I'll take the whole thing around to a friend's place later in the week which will give me better debugging abilities, and I'll try that then.

  15. #15
    Road Rasher
    Join Date
    Mar 2010
    Posts
    406
    Rep Power
    25

    Default

    alright, hope you get it sorted out.

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •