SNES/Famicon PCB Revisions and RGB Video

[mvsfan]

I have a Sync-On-Luma lightgun cable for my PS1. No problems at all. On the PS2 i just use component since im converting everything else to that anyway.

[RGB0b]

I just finished taking a ton of screen captures. I'll eventually try and get an many SNES revisions as possible, but here's what I have so far, including many different SNES Mini mods and a txt file with a full description of each. These are untouched bmp's without a watermark, so please be respectful and not use them without giving credit. I've had most of my pics stolen and used in other sites, eBay auctions, etc...it's quite annoying: http://www.retrorgb.com/files/SNESVersi ... ne2016.zip

I've only just started comparing these, but at first glance, I think I know why I used to think Mini's were better then 1CHIP's: Brightness. Technically, a SNES Mini modded with the correct resistors added to attenuate the brightness output IS better then a stock 1CHIP...but if you plan on modding, they are identical. I haven't tested yet, but I wonder if you're able to simply add the 750 Ohm resistors to a 1CHIP and have the output be as good as a SNES Mini: http://www.retrorgb.com/snes1chip.html#brightness

What does everyone think?

[Fudoh]

can't download your zip file. File transfer breaks off after a few MB. Any chance you can put this on a filehoster of your choice ? Sendspace maybe ?

[CkRtech]

What does everyone think?

The general community (including the YouTubers, gaming journalists, etc.) seem to have adopted the misconception that the mini has the best RGB picture output. I don't think that modifying a non-RGB native Mini with an amp and saying it is the best versus a stock 1CHIP for example is a fair argument due to a major control (encoder/amp) turning into a variable.

Pull the RGB signals, use proper input resistance to the inputs for identical amps in each system (1CHIP and Mini), get proper 75 ohm impedance on output, and then do a side by side.

The stock 1CHIP vs a THS-amped Mini is apples to oranges.

Fudoh - Huh. Zip worked for me.

Edit: That is some serious red in the 1CHIP-03 with THS vs the rest.

[RGB0b]

can't download your zip file. File transfer breaks off after a few MB. Any chance you can put this on a filehoster of your choice ? Sendspace maybe ?

Weird, sorry about that. Try this: https://www.sendspace.com/file/crkd15

[blizzz]

I can confirm Fudoh's problem. The download broke after 24.9 MB for me. Downloading the senspace link now.

[Josh128]

What does everyone think?

The general community (including the YouTubers, gaming journalists, etc.) seem to have adopted the misconception that the mini has the best RGB picture output. I don't think that modifying a non-RGB native Mini with an amp and saying it is the best versus a stock 1CHIP for example is a fair argument due to a major control (encoder/amp) turning into a variable.

Pull the RGB signals, use proper input resistance to the inputs for identical amps in each system (1CHIP and Mini), get proper 75 ohm impedance on output, and then do a side by side.

The stock 1CHIP vs a THS-amped Mini is apples to oranges.

Fudoh - Huh. Zip worked for me.

Edit: That is some serious red in the 1CHIP-03 with THS vs the rest.

What about non-amped mini? Mine is not amped, simply wired up without one and the picture is immaculate. Why even install an amp? I doubt a 1Chip is any better. Not saying the mini is better, just saying Im very skeptical you are going to get a better picture in any case.

[FBX]

What about non-amped mini? Mine is not amped, simply wired up without one and the picture is immaculate. Why even install an amp? I doubt a 1Chip is any better. Not saying the mini is better, just saying Im very skeptical you are going to get a better picture in any case.

RetroRGB sent me a comparison pic of 7374, 3-wire, and 7314 all on the mini. There was no question in my mind the 7374 had the best PQ out of the three methods.

The general community (including the YouTubers, gaming journalists, etc.) seem to have adopted the misconception that the mini has the best RGB picture output. I don't think that modifying a non-RGB native Mini with an amp and saying it is the best versus a stock 1CHIP for example is a fair argument due to a major control (encoder/amp) turning into a variable.

I'm not sure what the default non-modded RGB option for the mini would be, but I'm guessing the "3 wire" method? If so, I just looked at PQ from that and the stock 1CHIP-03, and I honestly have to say the 1CHIP-03 looks a little better! However, both were out-shined by the 7374, but as you say, it's not a fair comparison. It is however, the best PQ throwing out conventional rules.

[the_crayon_king]

So I took the pics from retrorgb and picked two of my favorites:
Half the picture is Snes mini 7374 (left) and 1CHIP-03 w THS7374 (right)


I can't tell the difference. @retrorgb are these with LPF on or off ?
[edit] oh didn't see the read file LPF off on the 1CHIP-03 but what of the SNES Mini THS7374 ? I am assuming its the same.

I think besides the SHVC-CPU-01, and SNS-CPU-GPM-02 the differences look mostly negligible and the 'best' may fall into the end users personal preference.

[FBX]

So I took the pics from retrorgb and picked two of my favorites:
Half the picture is Snes mini 7374 (left) and 1CHIP-03 w THS7374 (right)

I was also going to point out they are identical, but I figure that's because they are using the same mod.

[the_crayon_king]

So I took the pics from retrorgb and picked two of my favorites:
Half the picture is Snes mini 7374 (left) and 1CHIP-03 w THS7374 (right)

I was also going to point out they are identical, but I figure that's because they are using the same mod.

I do not know the various intricacies with these revisions but I'm surprised they aren't perceptibly different.

[Ikaruga11]

Isn't the SNES/Super Famicom picture stretched on anything that isn't a multi-scan PVM or BVM? The SNES outputs 256x224 to 512x448, which are 8:7 aspect ratios. When you display it on any consumer TV or monitor, the image is stretched to 4:3 or 16:9 depending on your settings. Obviously 4:3 is much better, but still suffers from slightly blurred picture quality and incorrect aspect ratios due to being stretched out to fill the screen, resulting in rectangles instead of squares and ovals instead of circles.

[CkRtech]

Isn't the SNES/Super Famicom picture stretched on anything that isn't a multi-scan PVM or BVM? The SNES outputs 256x224 to 512x448, which are 8:7 aspect ratios. When you display it on any consumer TV or monitor, the image is stretched to 4:3 or 16:9 depending on your settings. Obviously 4:3 is much better, but still suffers from slightly blurred picture quality and incorrect aspect ratios due to being stretched out to fill the screen, resulting in rectangles instead of squares and ovals instead of circles.

The SNES does not output a 8:7 aspect ratio. The internal dimensions such as 256x224 would be 8:7 provided that the pixels are output as square, however the output is designed to fill a 4:3 CRT. Therefore, the pixels are rectangular by design.

The aspect ration on the display is independent of any sort of blurring - It is simply a stretch.

[FBX]

Isn't the SNES/Super Famicom picture stretched on anything that isn't a multi-scan PVM or BVM? The SNES outputs 256x224 to 512x448, which are 8:7 aspect ratios. When you display it on any consumer TV or monitor, the image is stretched to 4:3 or 16:9 depending on your settings. Obviously 4:3 is much better, but still suffers from slightly blurred picture quality and incorrect aspect ratios due to being stretched out to fill the screen, resulting in rectangles instead of squares and ovals instead of circles.

The SNES does not output a 8:7 aspect ratio. The internal dimensions such as 256x224 would be 8:7 provided that the pixels are output as square, however the output is designed to fill a 4:3 CRT. Therefore, the pixels are rectangular by design.

I'm not so sure. Doesn't the SNES use the same timing as the NES? If so, this is what I understood to be the case according to NESdev:

Multiplying the pixel rate by the scanline length gives 39,375,000*6/4/11*640/(135,000,000/11) = 280 pixels per scanline. The PPU puts signal in 256 of these and a border at the left and right sides. The color of this border is the same as the backdrop color (usually the value in $3F00). This makes the pixel aspect ratio on a 4:3 TV to be 240/280*4/3 = exactly 8:7, or about 1.143:1.

[CkRtech]

Well, the 4:3 output that I am referring to is the display aspect ratio - not the pixel aspect ratio. I didn't specify which aspect ratio I was referencing, but I was inferring display aspect ratio. Your quote references the pixel aspect ratio. Independent of the calculations themselves - Wouldn't a pixel aspect ratio that is not 1:1 be considered a rectangular pixel? (Of course, you aren't technically going to get perfectly square pixels with NTSC, anyway)

[RGB0b]

What about non-amped mini?

That's the "3-wire" mod that's using the onboard amp.

I think besides the SHVC-CPU-01, and SNS-CPU-GPM-02 the differences look mostly negligible and the 'best' may fall into the end users personal preference.

I agree. I also think most people wouldn't be able to tell the difference between any of the 1CHIP/Mini mods, other then brightness.


Out of curiosity, does anyone know the motherboard revisions of the first SNES consoles to be released? I'm referring to the November 1990 Japanese consoles and the August 1991 North American consoles (I don't have any experience with the Korean models). I'd really like to get one of each, just to have as a reference.

[RGB32E]

Out of curiosity, does anyone know the motherboard revisions of the first SNES consoles to be released? I'm referring to the November 1990 Japanese consoles and the August 1991 North American consoles (I don't have any experience with the Korean models). I'd really like to get one of each, just to have as a reference.

DarthCloud had a spreadsheet with S/N and revision data. I have a VERY early US SNES and it's a SHVC-CPU-01 (S/N UN10007560). From the data available, UN1* models are all SHVC-CPU-01. Were you thinking there might be some differences in SHVC-CPU-01 models depending upon when they were manufactured?

[RGB0b]

DarthCloud had a spreadsheet with S/N and revision data. I have a VERY early US SNES and it's a SHVC-CPU-01 (S/N UN10007560). From the data available, UN1* models are all SHVC-CPU-01. Were you thinking there might be some differences in SHVC-CPU-01 models depending upon when they were manufactured?

Thank you, I completely forgot about that spreadsheet...and I even link to it on my site!

No, I just wanted to make sure I got the first and last consoles made and apparently, I already own them! I'd eventually like to test every revision out there, plus do more in-depth tests like CkRtech suggested, but I'm only one guy and don't have an unlimited budget to buy these things. I realize that there are a million other factors involved too; wear and tear, bad caps, etc, so I'd want to test a few of each version, just to be sure. This is a great start though. Other people can help too if they'd like...all you need is a Framemeister, USB3HDCAP, (other capture cards will work, but might effect the look) and FBX's profiles. Actually, maybe FirebrandX can upload the exact 4x +40 zoom profile, to make configuration easier for people who want to try? Combine that, plus his snapshot method: Right when you start Link To The Past, get the princess, exit the sanctuary, then save and quit. When you re-start, choose Link's house and don't touch anything; Then you'll be standing in the right spot and can take a screenshot from the exact same spot each time.

Hey, on a side note, my SHVC-CPU-01 has a broken audio board (the separate board that sits on top of the motherboard). Does anyone have one I can buy off of them...maybe someone has a broken console, but the audio board might still work? I broke that last year when poking and prodding for via's for the SPDIF mod. I'm sure it can be fixed by re-connecting the trace I broke, but I figured I'd ask, in case someone has one lying around.

[the_crayon_king]

DarthCloud had a spreadsheet with S/N and revision data. I have a VERY early US SNES and it's a SHVC-CPU-01 (S/N UN10007560). From the data available, UN1* models are all SHVC-CPU-01. Were you thinking there might be some differences in SHVC-CPU-01 models depending upon when they were manufactured?

Thank you, I completely forgot about that spreadsheet...and I even link to it on my site!

No, I just wanted to make sure I got the first and last consoles made and apparently, I already own them! I'd eventually like to test every revision out there, plus do more in-depth tests like CkRtech suggested, but I'm only one guy and don't have an unlimited budget to buy these things. I realize that there are a million other factors involved too; wear and tear, bad caps, etc, so I'd want to test a few of each version, just to be sure. This is a great start though. Other people can help too if they'd like...all you need is a Framemeister, USB3HDCAP, (other capture cards will work, but might effect the look) and FBX's profiles. Actually, maybe FirebrandX can upload the exact 4x +40 zoom profile, to make configuration easier for people who want to try? Combine that, plus his snapshot method: Right when you start Link To The Past, get the princess, exit the sanctuary, then save and quit. When you re-start, choose Link's house and don't touch anything; Then you'll be standing in the right spot and can take a screenshot from the exact same spot each time.

Hey, on a side note, my SHVC-CPU-01 has a broken audio board (the separate board that sits on top of the motherboard). Does anyone have one I can buy off of them...maybe someone has a broken console, but the audio board might still work? I broke that last year when poking and prodding for via's for the SPDIF mod. I'm sure it can be fixed by re-connecting the trace I broke, but I figured I'd ask, in case someone has one lying around.

If they are interchangeable between revisions then I have one.

[Josh128]

What about non-amped mini? Mine is not amped, simply wired up without one and the picture is immaculate. Why even install an amp? I doubt a 1Chip is any better. Not saying the mini is better, just saying Im very skeptical you are going to get a better picture in any case.

RetroRGB sent me a comparison pic of 7374, 3-wire, and 7314 all on the mini. There was no question in my mind the 7374 had the best PQ out of the three methods.

How can adding an amp improve the picture when the brightness is fine already? An amplifier can only add noise to an analog signal, not take away from it. I dont follow-- In what way did the amp picture look better than the straight wired one?

Did you post the pics anywhere in this thread?

[CkRtech]

Other people can help too if they'd like...all you need is a Framemeister, USB3HDCAP, (other capture cards will work, but might effect the look) and FBX's profiles.

When comparing different revisions of the SNES/different mods/etc, it seems like the Framemeister might add a few extra variables (artifacts), etc. While the device is great for gaming on flat panels, wouldn't it be better for the sake of down to the pixel comparison among revisions to use something that can directly capture 15khz RGB at original resolution?

[Pasky]

Other people can help too if they'd like...all you need is a Framemeister, USB3HDCAP, (other capture cards will work, but might effect the look) and FBX's profiles.

When comparing different revisions of the SNES/different mods/etc, it seems like the Framemeister might add a few extra variables (artifacts), etc. While the device is great for gaming on flat panels, wouldn't it be better for the sake of down to the pixel comparison among revisions to use something that can directly capture 15khz RGB at original resolution?

Yes and if you want a real comparison you need to not use a card that's down sampling from 4:4:4. Both the XRGB mini and SC-512/X-Cap1 and clones down sample to 4:2:2.

[FBX]

Other people can help too if they'd like...all you need is a Framemeister, USB3HDCAP, (other capture cards will work, but might effect the look) and FBX's profiles.

When comparing different revisions of the SNES/different mods/etc, it seems like the Framemeister might add a few extra variables (artifacts), etc. While the device is great for gaming on flat panels, wouldn't it be better for the sake of down to the pixel comparison among revisions to use something that can directly capture 15khz RGB at original resolution?

Yes and if you want a real comparison you need to not use a card that's down sampling from 4:4:4. Both the XRGB mini and SC-512/X-Cap1 and clones down sample to 4:2:2.

Neverthless, the approach actually really does show the difference in PQ from one console to the next. The ZOOM is set to 10X on these screengrabs retroRGB has been doing. Since the only variable that changes from one grab to the next is the SNES console itself, you get a pretty good idea what's going on. Of course you don't have to take my word for it. Just download the package he posted.

Did you post the pics anywhere in this thread?

[CkRtech]

Since the only variable that changes from one grab to the next is the SNES console itself, you get a pretty good idea what's going on.

It seems like the xrgb adds a bit of a gradient between the colors, though. I didn't really notice it until I installed a NESRGB mod for the first time a couple of years ago. I thought it was an artifact caused by my cable of choice/install/noise, etc, but it was the xrgb (or at least the standard/picture/zoom/whatever mode I was using at the time).

Using the xrgb to zoom in like that appears to inflate whatever image processing artifacts occur by quite a bit. I appreciate the time that goes into making the caps of each console, but a true side by side comparison of the different SNES models (if we apply the type of precision that is typical here at shmups) demands a direct capture without the extra processing and scaling. Having the xrgb in the chain just adds too many variables. I don't think we are going to start saying that a CPU-01 is as good as a Mini just because of xrgb processing, but a true comparison is going to come from a vanilla 240p analog capture.

Here is the SNES Mini 3-wire 1.1k Brightness image sized up with nearest neighbor scaling to show the gradient I am talking about.

[blizzz]

It seems like the xrgb adds a bit of a gradient between the colors, though.

Three reasons for that: 1. Fixed sampling size and phase, 2. Downsampling to 4:2:2, 3. Scaling to 1080p

I don't think there's any capture card that can perfectly capture a 256x240 signal without any issues. Maybe an oscilloscope would be the best way to see how sharp a picture is. An easy way to get a good result would be to use the OSSC with a capture card that captures in 4:4:4.

[Pasky]

It seems like the xrgb adds a bit of a gradient between the colors, though.

Three reasons for that: 1. Fixed sampling size and phase, 2. Downsampling to 4:2:2, 3. Scaling to 1080p

I don't think there's any capture card that can perfectly capture a 256x240 signal without any issues. Maybe an oscilloscope would be the best way to see how sharp a picture is. An easy way to get a good result would be to use the OSSC with a capture card that captures in 4:4:4.

Cut out the OSSC, it's not necessary. Capture at native resolution then use nearest neighbor in a paint program to scale it up for comparison.

[RGB0b]

Capture at native resolution then use nearest neighbor in a paint program to scale it up for comparison.

I did that when I first got my USB3HDCAP and I had a hard time telling the difference between the 7314, 7374 and 3-wire mods. To be honest, I even have a hard time telling the difference between the three on my Sony BVM...and I'm pretty sensitive to this stuff.

Using the xrgb to zoom in like that appears to inflate whatever image processing artifacts occur by quite a bit. I appreciate the time that goes into making the caps of each console, but a true side by side comparison of the different SNES models (if we apply the type of precision that is typical here at shmups) demands a direct capture without the extra processing and scaling.

I think that's exactly why it's so much easier to see the differences. One can argue that you're not seeing the differences between the SNES versions, but the differences between how the Framemeister processes each. It still provides a way to visually see the differences...especially when you zoom into those screen caps as you did.

I don't think there's any capture card that can perfectly capture a 256x240 signal without any issues. Maybe an oscilloscope would be the best way to see how sharp a picture is. An easy way to get a good result would be to use the OSSC with a capture card that captures in 4:4:4.

I think if we wanted to go crazy, a combination of both would be the best method: Use an oscilloscope to test each motherboard revision and post that for the technical results, like what was done with the GC/Wii comparisons. Then also post zoomed screen caps, so the average person can see just by looking at it. Maybe OSSC/4:4:4 capture card is a better idea for this then, but you'd be lacking the zoom mode.

Either way, I love discussions like this. Even if we decide I should throw out the screen caps I took and start again, it was NOT a waste of time, just another step in the right direction!

[FBX]

But really these pics are plenty fine enough. Just look at the GPU versus the 1CHIPs in your package. It becomes pretty obvious the difference in quality without having to go ape on the capture method being used.

[Blair]

well, I ended up getting a one chip. I didn't expect the output to be so bright and detailed. I'm not sure if it's really better looking than my UN25. but it's definitely different.

I'll have to take some comparison screenshots when I get some time later, I'm pretty happy with it though.

Can you open them both to tell us what boards are in each?

Yes I can. (just got my security screwdrivers in the mail).


this is what's inside all of my super Nintendo's


SNES UN23 - this is my oldest system, the chipset is [snes 02] video quality is a bit soft, and a little dim. clean with no interference

SNES UN25 - this is my favorite super Nintendo, the chipset is [snes 02] video quality is sharp and even with no interference. it's perfect!

SFC S24 - I just got this from Japan, the chipset is [rgb 01] sharpness is pretty similar to the UN25, however it has interference patterns everywhere and these faint vertical lines evenly spaced all across the screen, extremely distracting when scrolling. very disappointing.

SNES UN80 - this is my second favorite super Nintendo, the chipset is [CPU - 1CHIP - 01] video quality is very sharp, somewhat overly bright and some of the graphics have a bit of a ghosting or "shadow effect" (only really noticeable without scanlines, but it annoys me that the video quality isn't as clean as my UN25)


All systems tested on...


[Displays]

PVM-14m2u CRT

PVM-20L5 CRT

DVDO VP50pro + SONY Multiscan CPD-E400 CRT

DVDO edge green + Samsung LN40 LCD


[Cables]

Hori S-video cable (imported)

Retro_console_accessories Coaxial RGB Scart cable. (C-Sync)


[Power supplies]

All original official power supplies

Connected to a trip light power strip with good filtering.



I really wish there was some way I could fix that Super Famicom, any ideas?

[blizzz]

I really wish there was some way I could fix that Super Famicom, any ideas?

There's no fix I know of for the SNS-CPU-RGB-01/02 and SNS-CPU-APU-01 consoles. Even a THS7314 won't fix it.