Jump to content



0

5200 S-video mod interferance


7 replies to this topic

#1 FABombjoy OFFLINE  

FABombjoy

    Moonsweeper

  • 404 posts
  • Location:Lansing, MI

Posted Tue Sep 23, 2003 5:18 PM

Hey all,

While on my s-video mod world tour, I'm on to the 5200 now. I've tried the mod on 2 different units (a 2 and 4 switch) using 2 seperate circuits, and both produce the same problem:

On the left side of the screen, occupying almost the entire left half, are a series of vertical bars in the luminance signal. They're not strong, but they are visible. The right side looks fine. They appear to be from a digital source, since they sometimes change 'density', and when playing Pole Position they actually seem to acquire a parallax effect. It is present in the RF signal, only not as strongly.

I've tried moving the circuit & wiring all around, but repositioning it seems to make no difference; the bars are always there & always the same intensity.

Anybody else ever seen this problem? Any solutions?

#2 cwilkson OFFLINE  

cwilkson

    Dragonstomper

  • 517 posts
  • Location:Santa Clara, CA

Posted Wed Sep 24, 2003 1:36 AM

FABombjoy said:

Anybody else ever seen this problem?  Any solutions?

Can you get a screenshot? It's tough to guess without actually seeing the picture.

-Chris

#3 FABombjoy OFFLINE  

FABombjoy

    Moonsweeper

  • 404 posts
  • Location:Lansing, MI

Posted Wed Sep 24, 2003 8:47 PM

Ok, here goes:

Picture 1
Picture 2

I know it looks like a pong, but it really is a 5200. My River Raid cart is just flakey enough to provide these good pictures.

I added a few 1000uF caps between +5 and ground on the circuit, and almost all of the lines disappeared. The darkest areas remained, tho. 1000uF is the largest cap that I have at the moment. I suppose that I could dig up some higher value caps to try out.

#4 cwilkson OFFLINE  

cwilkson

    Dragonstomper

  • 517 posts
  • Location:Santa Clara, CA

Posted Thu Sep 25, 2003 12:07 AM

Wow...that's really ugly! I don't know what that could be. It looks like a clock signal being coupled into the video. Filtering the power supply was a good idea. Make sure that any chips in your video section have their own bypass cap...0.1uF should do. Make all your wires as short as possible, and make sure that there are NO digitals signals in your video section. This is especially true for the the luma lines...make sure they are safe from digitals!

-Chris

#5 Bryan OFFLINE  

Bryan

    Quadrunner

  • 7,626 posts
  • Cruise Elroy = 4DB7
  • Location:Port St. Lucie, Florida

Posted Thu Sep 25, 2003 8:24 AM

Yep, that's noise from within the unit. Looks kinda like you have a bar for each of Antic's RAM refresh cycles.

Larger caps aren't better at filtering in this situation - what's more important are caps with low ESR, like a 1-10uF Tantalum.

-Bry

#6 FABombjoy OFFLINE  

FABombjoy

    Moonsweeper

  • 404 posts
  • Location:Lansing, MI

Posted Mon Sep 29, 2003 7:08 PM

cwilkson said:

Make sure that any chips in your video section have their own bypass cap...0.1uF should do.

Hmm... is there a mod that uses chips, or are you referring to the chips in the 5200 video section? I'm using the mod as listed in the 5200 FAQ. If there's a different or better mod, I'm game.

I switched to shielded coax for the wires from the circuit to the plug, added a few 1000uF caps to the board where the circuit's power leads are connected, and moved the circuit away from the RAM chips. These helped, but didn't eliminate the lines.

bryede said:

Larger caps aren't better at filtering in this situation - what's more important are caps with low ESR, like a 1-10uF Tantalum.

I can order some tantalum caps to add in. Are you recommending that I add them to the stock 5200 electronics, or somewhere in the mod itself?

BTW: Thanks for the help! This is frustrating, but I know there's a remedy. I'm just suprised that nobody has encountered this problem before.

#7 cwilkson OFFLINE  

cwilkson

    Dragonstomper

  • 517 posts
  • Location:Santa Clara, CA

Posted Tue Sep 30, 2003 2:12 AM

When using those big electrolytics, be sure to put a 0.1uF ceramic in parallel with each one...as close a humnaly possible. At high frequencies, electrolytics become inductive as well as highly resistive. Tantalums are much better, but still do it to some extent.

I don't know the 5200 architecture very well. So I'm not sure what the video section looks like. Do they use a GTIA? Antic? What? And I'm not sure how those chips operate. but those are the chips that I'm talking about. The chips that actually produce the video signals. And any chips that you add when modifying the 5200 video. Adding a bypass cap, at the chip, will give the each one a local reservior of charge to draw on...reducing the power supply glitches seen by all the other chips.

Those might be chroma bleeds. If the chroma signal is mixed in via a resistor, trying increasing it's value. If not, (if it's passed through a cap), try adding resistor in series with the cap. Try a 1K and see if it helps.

-Chris

#8 cwilkson OFFLINE  

cwilkson

    Dragonstomper

  • 517 posts
  • Location:Santa Clara, CA

Posted Tue Sep 30, 2003 2:16 AM

Ahh, yes. I just got hold of the FAQ. So, for the chroma output, try increasing the value of the 2k resistor. A 10K pot might be useful to find a value that suits your needs. If the resistor is too large, you'll loose the color info, but no harm done. Try to see if the vertical lines fade or blur as you increase the value of the pot. If so, it's chroma bleed. The peaks from the chroma signal are being added to the luma and producing vertical stripes. I've seen this a lot while creating the 2600 s-video spec board.

-Chris




1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users