Snes Signal Noise on RGB

Hi Guys,

I noticed this was biefly mentioned on a certain thread here, but no one seemed to have a solution.

I recently purchased the HD Retrovision cables, which convert the Snes' RGB signal into component, and I tried it on 4 of my console revisions, with mixed results.

The revisions I have are:
- My SNS-CPU-APU-01 revision produces a "grainy" image on RGB
- My SNS-CPU-RGB-02 revision produces a diagonal pattern on the screen
- My SNS-CPU-1CHIP-01 and my SHVC-CPU-01 revisions on the other hand, produce a crystal clear image.

A youtuber recently made an unrelated video regarding power issues causing signal interference on his Snes through S-Video, however, when he fixed it, the same diagonal patterns I am seeing were visible on his screen, though I'm not sure wether he was aware of it. You can review the pattern on his video, look on the "good power" version of his images, on full screen HD resolution on a large TV, and you will notice the diagonal patterns:


I was suggested by Ste from Retrovision that this is a design flaw inherent to these 2 revisions, and to try modding them with the RGB bypass fix from retrorgb (which uses a THS7314 Amp), as explained on the link: http://retrorgb.com/snes1chip.html, however, there is no guarantee that this will work, and I have no soldering skills to test it myself.

Has anyone modded an APU or an RGB revision with this fix? does it clean out the signal noise? if so,let me know, I will greatly appreciate it.

 

I've used the THS7314 for a few other consoles, not specifically your SNES revisions. The chip helps cleaning up noise but you still need to have a good / clean sync signal. I don't know what Retrovision uses for sync (they had the choice between composite video, luma or combined sync). If they used luma or composite, maybe you can get better results with csync.
Best if you leave it to an experienced modder with the request to "make it as nice looking as you can". Ask him to add the "SNES vertical line capacitors" as well

 

Might be 'normal' with these revisions.

 

you need bortie4938's RGB BYPASS encoders, then you can kiss all these symptoms goodbye!

 

I'm no expert on the subject, but after discussing the issue with retrovision (they were already aware of it) I believe this is unrelated to Csync, we believe it is a design flaw on these revisions, so it might be something else. I will try the vertical line capacitors anyways though, but I was told the vertical lines were a power-related issue? Hopefully this is also just power related.

Correct, this is normal, I was just hoping that anyone out there also noticed and found a solution to it, it really sucks.

Where can I buy these encoders? do you have tutorials as how to install them?

 

I bet you can improve the image quality by adding electrolytics and a THS7314 for RGB.
Nintendo didn't put enough caps into their designs. The vertical line is one symptom of that. It's caused by a power sag on periodic DRAM refreshes. Caps will help with that.
Caps could also help with the noise, depending on what the source is.

"bortie4938's RGB BYPASS" is a THS7314 used as RGB amp

 

I bought mine from here: http://assemblergames.com/l/threads/borti4938-sells-his-last-modding-parts-for-the-snes.60049/ , there are a couple of install pics on the thread as well - and more info here: https://github.com/borti4938/SNES-AddOn-PCBs/tree/master/RGB Bypass

 

ive had run ins with diagonal lines on a couple occasions.

on a snes mini i rgb moded, fixed by bridging a pad and then undid it when i got a quality scart cable with c sync wired.
eg http://assemblergames.com/l/threads/snes-mini-rgb-measurements.53053/page-3

another time i was testing out a rgb to component converter using a 2 chip, once again it came down to the scart cable. C sync may be the issue here or the chroma in composite..... at a guess, might be a clue

 

On SNS-CPU-APU-01 and SNS-CPU-RGB-01/02 boards there is no way (so far) to get rid of the checked pattern. Even a bypass with a THS7314 or THS7374 does not help.

 

SNS-CPU-RGB-01/02 are pretty late / before 1Chip NTSC consoles, right?
I'll have to get one and scope it!

 

You only need 1 470uf capacitor on the voltage regulator to fix the vertical line issue. The capacitors on the RGB chip are not necessary.

 

You need a few more to remove the line about 80%. With just one cap, you remove 50%. Depending on your display, you can still see it then.

 

As I suspected. I will have to sell my APU and RGB revisions and get another SHVC and a GPM revision, just for testing. The GPM revision doesn't have any patterns, right?

 

1 cap on the 7805 removed it 100% for me.

 

That's fantastic, AndehX. I'm looking at the issue with an oscilloscope. One cap on the 7805 reduces the signal drop by 50% exactly.

 

I've had people email me with mixed results: Some say just one 470uF cap works, others say they needed to add the ones on the S-RGB chip.

Can you please post your scope results here? So far all the testing I've been involved in is "trial and error"...and while it seems to be fixing the issue, I'd really like more details on why it's happening and why some consoles need the extra caps.

 

I'd imagine the need for extra caps on the exact same model numbers is related to the original caps deterioration.

 

Well 1 470uf cap definitely 100% fixed it on my particular SNES (1-CHIP-02) Heres an example of before and after:

https://postimg.org/image/6aazfmiuf/

https://postimg.org/image/4xjqdy6zl/

 

im my case i put a 220uf on the 7805 and on the encoder

 

I'm looking for a Super Famicom with the problematic mainboards. When I have one, I'll document the scope readings as well