Video probs w/ SNES RGB (Vertical Bands, etc)

[GigaBoots]

Just to state it one more time before things go further in a bad direction, this thread is not me accusing anyone of anything. This thread is me trying to find out what causes this problem and solve it. That's all I want.

[Voultar]

The best way to reduce the ghosting for 99.9% of SNES enthusiasts (including me), is just by using a THS7374-based solution that's properly attenuated. Borti himself has made an excellent board for this, as well as Voultar. I have spend a ridiculous amount of time with both solutions and can vouch for their quality.

I know of their reputation. That's why I went to Voultar but everyone here seems to think this is unusual for a THS7374 mod and that's what I got from Voultar. I'm not trying to point fingers here. I just want to figure out what's happening and get it solved.

Anyone that wants better quality then that should expect to pay an extremely high price for an EXPERT modder to do Borti's other mod, for a minimal quality difference. It is by no means a "standard procedure".

So are you implying here that there's a higher quality mod than the one voultar provides with the THS7374 or am I misinterpreting this entirely and you didn't realize I already have Voultar's mod?


Also yes, that mod above sounded dangerous AF.
Also, hi Bob. I'm a fan of your work.

I've sampled dozens of SNES systems, and the results you're experiencing are certainly an exceptional corner-case.




^ This "trailing" off visual appears to be a DC error. "Overshooting" on the video outputs would have a much different visual appearance, in my opinion. Look at the hard luma edges, there's nothing there. Current limiting the DAC on the ASIC can have some negative consequences beyond user-error/damaging the chip. There's a variably wide output swing on the 1CHIP ASIC. If you do that across the board, then you're going to have video plummet on a select number of 1CHIP systems. This disortion that you see is more likely caused by a component failing/oscillating +supply on the SNES mainboard, or even aged capacitors.

[RGB0b]

I just want to figure out what's happening and get it solved.

I'm also curious about this: Why are some consoles ghosting worse then others? Is it possible there's something else failing? Even a dying PSU? 7805? Caps? I appreciate that you guys are taking the time to figure this out...especially the guys why are actually scoping the problem!

So are you implying here that there's a higher quality mod than the one voultar provides with the THS7374 or am I misinterpreting this entirely and you didn't realize I already have Voultar's mod?

I didn't know you already had his, but that's the best mod available (7374 bypass). The only other stuff you can try goes beyond the scope of a "SNES RGB mod" and that's the only point I'm trying to make: The 7374 mod is still the best out there, but if you're obsessive and totally fucking crazy (like me), then you can also try things that would be considered "above and beyond"...and stuff most people couldn't do:

- Full cap replacement
- Test and possibly refurbish the PSU
- Try Borti's other fix, then re-build the bypass board to account for the different attenuation
- Scope the individual SNES to account for each one's tolerance swing and add components on the 7374 to match that exact unit.

[GigaBoots]

I'm thrilled to have won the freakishly bad SNES lottery. Who better to win such a thing than a psychopath who records everything losslessly and upscales to 4K before uploading to YouTube...

Thanks for your help guys.

So which actually makes more sense monetarily:

1. Sending my SNES off to a modder so he can check these things from the least expensive possibility to the most to see which the problem is
or
2. Buy a new SNES Jr. and get it modded w/ the THS7374?

[Voultar]

Oh wow, I totally missed this in the pic.



That's more than likely the result of an abnormal supply.

So which actually makes more sense monetarily:

1. Sending my SNES off to a modder so he can check these things from the least expensive possibility to the most to see which the problem is
or
2. Buy a new SNES Jr. and get it modded w/ the THS7374?

If you would've emailed me about having issues. I would've been more than happy to help you.

[GigaBoots]

Oh wow, I totally missed this in the pic.

That's more than likely the result of an abnormal supply.

So you're saying it could be something as simple as the AC Adapter? Let me dig around, I should have 2 more hanging out.

[rama]

Oh gee, posts are happening too fast right now :p

I believe I need to clarify some things.
- The current limiting idea is mine. I tried it first about a year ago and reported it on a German forum where people then tried their luck with it. Borti is very good at documenting findings, as well as he's an actual electrical engineer, so I am happy with him documenting it.
But I don't think he ever actually recommended doing it.
Borti also scoped this back then and verified that it solves the DAC distortion before it happens (at 10 Ohm already, by the way).

- If you only use a 10 Ohm resistor, the output is still (usually) too bright / over 0.7Vpp fully terminated. So an attenuation mod would still be required. I don't know if using 20 or 30 Ohm is actually a good general solution for the entire 1Chip situation.

- I noticed that power draw (from the game cart, for example) will make the distortion worse. So yea, power could be another culprit, and on some consoles more than on others.

Edit:
Also, this might have gotten buried on last page.
GigaBoots: You don't happen to use a PAL console with an NTSC cable, do you?

[GigaBoots]

GigaBoots: You don't happen to use a PAL console with an NTSC cable, do you?

Nope.

[Voultar]

Oh gee, posts are happening too fast right now :p

I believe I need to clarify some things.
- The current limiting idea is mine. I tried it first about a year ago and reported it on a German forum where people then tried their luck with it. Borti is very good at documenting findings, as well as he's an actual electrical engineer, so I am happy with him documenting it.
But I don't think he ever actually recommended doing it.
Borti also scoped this back then and verified that it solves the DAC distortion before it happens (at 10 Ohm already, by the way).

- If you only use a 10 Ohm resistor, the output is still (usually) too bright / over 0.7Vpp fully terminated. So an attenuation mod would still be required. I don't know if using 20 or 30 Ohm is actually a good general solution for the entire 1Chip situation.

- I noticed that power draw (from the game cart, for example) will make the distortion worse. So yea, power could be another culprit, and on some consoles more than on others.

Edit:
Also, this might have gotten buried on last page.
GigaBoots: You don't happen to use a PAL console with an NTSC cable, do you?

Borti is a good man and has all of my respect. You both are good men and very knowledgeable.

But this is a good opportunity to talk about something else that's quite relevant to these problems.

SNES PSU's are beginning to show their age. But what's in a SNES PSU? A transformer, windings and a large filter cap.

Ah, the filter cap.. These are beginning to dry, these days. The SNES Mini doesn't have a DC filter cap. The spot is there for it, but it's unpopulated!

So when the filter cap in the SNES PSU begins to fail, there's no DC filter cap on the SNES' input to help with this. So what can happen when the filter capacitor loses performance in the transformer and you have no filter cap on the console to help?

All kinda of terrible visual disturbances.

[rama]

Indeed, power is something Nintendo didn't seem to care about much.
There isn't a single electrolytic on the 5V regulated side in any SNES.
1Chips did have a spot for one (OS-CON), but of course it was never used.

This is pretty relevant to the 'SNES vertical line'

[GigaBoots]

Guess I'll need to find a modder who's interested in helping fix this. Thanks everyone for your help.

[rama]

Voultar offered to help.
I'd also take your console but shipping to Germany and back would probably be expensive.

With our current knowledge, it's hard to tell what exactly is wrong on your side.
Someone will have to take a look at your hardware.

[GigaBoots]

Voultar offered to help.

I actually didn't see that at all!

If you would've emailed me about having issues. I would've been more than happy to help you.

Sorry about that. The only e-mail I had was for sales and I felt a little weird shooting an open-ended thing like this to sales! I did PM on Twitter though, but I doubt you've seen it. I'll e-mail in a sec. Thanks.

[RGB0b]

Please post the results here too. I've had a lot of reports of bad SNES PSU's and I'm curious if that's a factor in ghosting. Also, replacing the 7805 is never a bad idea:
https://youtu.be/57f8ArzKK0U
http://www.retrorgb.com/snesverticalline.html

[GigaBoots]

Please post the results here too. I've had a lot of reports of bad SNES PSU's and I'm curious if that's a factor in ghosting. Also, replacing the 7805 is never a bad idea:
https://youtu.be/57f8ArzKK0U
http://www.retrorgb.com/snesverticalline.html

Absolutely

[Voultar]

Please post the results here too. I've had a lot of reports of bad SNES PSU's and I'm curious if that's a factor in ghosting. Also, replacing the 7805 is never a bad idea:
https://youtu.be/57f8ArzKK0U
http://www.retrorgb.com/snesverticalline.html

Absolutely

Just out of curiosity, were you able to try a different supply?

[GigaBoots]

I checked it out and it didn’t change anything. I hooked it up to a different outlet too (someone online suggested it was a ground loop perhaps) and that also did nothing. I’m going to hook up S-video cables sometime late tonight to check if it shows up there. It will be a while though, I gotta sleep eventually.

[borti4938]

Hey,

I came a bit 'late' into this thread here. But I will left some notes on the one or other post here.

Some notes on the "one resistor brightness issue solving" mod:

There's one problem with this mod: It's pretty dangerous to do. The pins are close together and in a cramped space. Lifting it and installing a resistor in that location requires some skill.
If people do it wrong, it will probably kill the console.

definitely agree. The cartridge slot is very close here and does not leave that much space for working on it.

maybe the trace itself that leads to pin 155 could be interrupted/cut and rerouted instead, but I'd have to open my SNES Jr. up to see.

No way. You have to lift the pin.

While Borti's mod seems to fix the issue, it's an invasive, irreversible mod.

That's wrong. The mod is completely reversible. Just remove the resistor and bend the pin back into place. (Of course, one should not do that more than once or twice! But who want's to do that ?)

and I really, really love that people are now being cautious of all modders, but please just be careful with your recommendations. Saying that's "the fix" will result in many people killing their consoles for a solution that has very little testing, compared to the thousands of people who have done the other recommended mods. Remember, we're looking to get the best out of our consoles and preserve them!

I know what you mean. And I totally, totally agree with it in this perspective!
Just a little thing: "thousands of people who have done the other recommended mods" is not an argument for me.
Thousands of people use Everdrives and Repros with 3.3V chips on the card in their 5V consoles with resistors instead of level shifters. Thousands of people has used (and probably still use) sync-boosted cables.
Here in Germany 75% all people does not know when to use the German word "wie" and "als" correctly

I hope, you know what I mean. No doubt again the context of the quote at all, just the argument is as like as it comes from one of my students

Just back to the topic:
Every time, someone ask me about which method to use - one-resistor or three-resistor, I recommend the three-resistor mod and say he can uses the one-resistor mod if he's really really really 100% sure what he is doing and feels comfortable in lifting pins.

I can see the ghosting on my SNES mini, very annoying. It's noticeable on a TV screen through both an OSSC and framemeister.

I'm using an older amp board from retrorgb, but I feel like ths7374 wouldn't fix it, just make it less noticeable.

That's my experience. Also my experience is that is more noticeable on an HD-TV rather than on a CRT.

I believe I need to clarify some things.
- The current limiting idea is mine. I tried it first about a year ago and reported it on a German forum where people then tried their luck with it.

That's true! And I'm sorry that this is not always clear.
Rewriting the READMEs on Github is still on my ToDo (unfortunately with little priority) and then I will point it out there, too.


Some words on the "three resistor brightness issue solving" mod:

What is this 750 ohm resistor mod you speak of FBX? I had Voultar perform the mod himself

You can be 1000% sure that you have this mod inside. It's part of his PCB design he uses

Back to basic topic or to say it with some other words:

Well the point here is he was getting clearly visible ghost images

My guess would be the PSU as Voultar pointed out, too. But as you tried some different PSUs... Mhhhh... Or also some faulty caps inside your RGB cable, @GigaBoots?
Very unlikely: I also have seen something similar (I don't remember the problem pattern exactly) with a faulty THS7374. This chip could have randomly took some damage after first few minutes of usage.

Especially because Voultar modded your console (I hope, I got it right), the problem is likely outside of the console. But other suggestions like recapping, renew 7805 and others like stated by rama and retrorgb may help

I checked it out and it didn’t change anything. I hooked it up to a different outlet too (someone online suggested it was a ground loop perhaps) and that also did nothing. I’m going to hook up S-video cables sometime late tonight to check if it shows up there. It will be a while though, I gotta sleep eventually.

Please also check Composite Video!!! I'm curious if you have the problem here, too.


And a little bit of off-topic

Borti is a good man and has all of my respect

Thanks man. This is reciprocal - you have all my respect, too.
Especially your YT-videos are a massive gain for the community. Also your technical posts are always a very good read

[Syntax]

Lifting pin 155 is pretty easy and safe.
Just bend a spring till it's kind of straight with a tiny right angle on the end and slide it behind the pins on that side till it's touching the back of 155. Apply tension.

Clean your iron free of solder and rest the tip on 155 whilst putting tension on the spring. Should pop very fast and clean.
It doesn't matter if the iron tip touches other pins, if it's clean no solder will run and the spring will only move 155.

I find it near impossible to "pick" 155 out with a needle without damaging surrounding pins.

You should always rest a card or blade on top of pins when bending them up. You don't want them to bend where they enter the chip, you want to bend them somewhere along the leg.

For the record Famicom Jr, replaced 7805S (2A version with 220ohm cap attatched but cap not needed I feel ) 20 ohm mod, bypass board.
0 issues.

I power all my consoles off the same Mega Drive 1 PSU.
I have 2 leads coming from it so I can also power my MD2. Ones pos center the other neg center so I always insulate 1 to avoid sparks:)

I'm considering switching all my consoles to use switching OKI-78SR-5 1.5A instead of the linear 7805S 2A ive been using.
They can take up to 36v dc input

[FBX]

So it seems to be the case that there's definitely something more going with Gigaboots's SNES junior after all! And again, I don't mean to dismiss the work that rama and borti have done, rather I just felt this had to be something more unique to Giga's console because I would definitely notice that level of noise on my SNES junior had it been a universal design flaw. I'm thinking he maybe needs a full cap replacement, and regulator too (just in case)?

For posterity, here's that Mega Man X stage on my SNES junior with Voultar's bypass board (making this effectively the same SNES as GigaBoots's):

Spoiler

The only issue with mine is the Framemeister hates how sharp the picture is, so I end up with faint aliasing lines all over the picture. That's why I prefer to use my 1CHIP-03 for now, but my OSSC finally shipped a couple days ago, so I'm excited to see how my SNES junior with Voultar's board looks through that.

[Ruprit]

I don't have all the issues like GigaBoot's console, but I do have two noticeable vertical bands in the same location. It's also a Mini modded using Voultar's 7374 bypass. I've also replaced the 7805 voltage regulator with a 78S05. I also own a 1chip-03 with the exact same modifications and have no issue with that console. For power, I have one original official AC adapter and the one that's recommended by HDRetrovision. Neither make a difference. This is the first time I have seen someone else describe this problem and I have been trying to find a solution for it for a few months now. I'll try recapping my mini this week and post an update.

[rama]

@FBX: I looked at your last screenshot in inverse color and with image / contrast enhancements.
Short of the slightly misaligned sampling clock on the ADC, I can't see anything wrong with it.
(The 3 vertical columns of 'noise'. They're so weak, I don't think anyone would ever notice them in gameplay:
)

It's interesting in that there's none of the SNES ADC distortion that I expected to see from any 1Chip S-CPUN.
It appears that your console doesn't have this issue at all and this is different from any 1Chip I've ever had.
Note that I could only ever test PAL machines.
PAL machines have 160 Ohm pull down resistors and a different output stage, compared to NTSC designs.
Also I never had a machine newer than 1Chip-02 PAL (late 1996) and especially no SNES Jr design.
Maybe this is important after all...

So yeah, no conclusions yet. Just something 'odd', compared to my samples.


Ruprit:
If you mean the SNES vertical line, this is typical of any SNES that doesn't have additional Vcc bypass capacitors.
If it looks anything like this video, you only need an extra capacitor:
https://www.youtube.com/watch?v=57f8Arz ... u.be&t=17s

[Voultar]

@FBX: I looked at your last screenshot in inverse color and with image / contrast enhancements.
Short of the slightly misaligned sampling clock on the ADC, I can't see anything wrong with it.
(The 3 vertical columns of 'noise'. They're so weak, I don't think anyone would ever notice them in gameplay:
)

It's interesting in that there's none of the SNES ADC distortion that I expected to see from any 1Chip S-CPUN.
It appears that your console doesn't have this issue at all and this is different from any 1Chip I've ever had.
Note that I could only ever test PAL machines.
PAL machines have 160 Ohm pull down resistors and a different output stage, compared to NTSC designs.
Also I never had a machine newer than 1Chip-02 PAL (late 1996) and especially no SNES Jr design.
Maybe this is important after all...

So yeah, no conclusions yet. Just something 'odd', compared to my samples.


Ruprit:
If you mean the SNES vertical line, this is typical of any SNES that doesn't have additional Vcc bypass capacitors.
If it looks anything like this video, you only need an extra capacitor:
https://www.youtube.com/watch?v=57f8Arz ... u.be&t=17s

What you're seeing in the inverse is high frequency content folding back in which results in an aliasing effect. The 7374 (probably) has the LPF disabled.

9.5Mhz filter for SD content


Filter disabled on 7374 with OSSC set to 16Mhz, I think. You can see the high frequency content folding in here.


That's actually a good question.

Gigabootz, what do you have the LPF set to on your OSSC? If you set it to 9MHZ, is there a difference?

[FBX]

What you're seeing in the inverse is high frequency content folding back in which results in an aliasing effect. The 7374 (probably) has the LPF disabled.

Correct, I have LPF disabled, and the Framemeister doesn't seem to like it. When my OSSC arrives (which got shipped a few day ago), I'll do some tests with the filter levels on that, though I understand the 9.5 Mhz filter already solves this issue. For now, I prefer using my 1CHIP-03 with the stock encoder because it looks just as sharp on the Framemeister without that aliasing problem.

(The 3 vertical columns of 'noise')

If you apply the inverse to the whole screenshot, you'll see the columns are more than just those three, but rather keep going all the way across the image. That's the Framemeister's weird reaction to the unfiltered SNES signal. I can adjust the Framemeister's Sync Level up or down, and those lines shift horizontally in position as I change that value. Thus they are independent of the graphics.

[GigaBoots]

Gigabootz, what do you have the LPF set to on your OSSC? If you set it to 9MHZ, is there a difference?

There's no difference. I had LPF set to AUTO in the OSSC. There's a 9 MHz pic in the google drive folder now for comparison. https://drive.google.com/open?id=0B2WY2 ... HFOVGd2TTg

The S-video shows the same problems. I've uploaded pics of those to the Google Drive folder as well.

I also checked the same CSYNC SCART cable on the N64 and it's not showing any of the same symptoms.

N64 Pure White Waveform

Spoiler

SNES Pure White Waveform

Spoiler

[Voultar]

Gigabootz, what do you have the LPF set to on your OSSC? If you set it to 9MHZ, is there a difference?

There's no difference. I had LPF set to AUTO in the OSSC. There's a 9 MHz pic in the google drive folder now for comparison. https://drive.google.com/open?id=0B2WY2 ... HFOVGd2TTg

The S-video shows the same problems. I've uploaded pics of those to the Google Drive folder as well.

The S-video output is driven by the SRGB encoder that's on board. Seeing that S-video is equally affected, It's safe to say that you just have something wonky going on with your Mini that's unrelated to your mod.

Believe me, A LOT of people use this setup to speedrun MM:X. If this was a common occurrence, there would be torches and pitch-forks.

I would probably do a shotgun approach and replace caps and the linear regulator.

I'm happy to take a look.

[Harrumph]

It appears that your console doesn't have this issue at all and this is different from any 1Chip I've ever had.
Note that I could only ever test PAL machines.
PAL machines have 160 Ohm pull down resistors and a different output stage, compared to NTSC designs.
Also I never had a machine newer than 1Chip-02 PAL (late 1996) and especially no SNES Jr design.
Maybe this is important after all...

I think PAL/NTSC board design aspect could be explored further, given that this particular ghosting issue has been discussed foremost in european communities (and Syntax is in Australia, so I presume deal mostly with PAL consoles also). Because I was surprised our NA community OCD nerds like retrorgb and FBX seemed largely unaware of this issue. Would be great if Voultar could get a PAL 1-chip to look at also!

[Syntax]

I only have ntsc consoles.
Cleaned the house free of 50hz.

I do have a 3 chip pal Snes I use to test t.v. rgb mods and am always happy to share info.
I unfortunately don't have a suitable capture device for our tests to confirm my system is as flawless as I think it is.

[rama]

Filter disabled on 7374 with OSSC set to 16Mhz, I think. You can see the high frequency content folding in here.

Interresting!
I'll try and get a THS7374. This low pass filter might make my life a lot easier with my GBS8200 board + SNES.

I wonder, have you ever tried adjusting the sample rate and phase when using no filter?
I assume the OSSC allows tweaking those, right?

[Ruprit]

Ruprit:
If you mean the SNES vertical line, this is typical of any SNES that doesn't have additional Vcc bypass capacitors.
If it looks anything like this video, you only need an extra capacitor:
https://www.youtube.com/watch?v=57f8Arz ... u.be&t=17s

I don't have a capture card, but I tried to take an off-screen photo. It looks very similar to your photos and not like the typical vertical line problem as it only affects the left side rather than the center.