A fix for SNES jailbars

[maxtherabbit]

annd we're good
diagonal noise GONE
jailbars in SMW night screens GONE

Which board revision is this? Under which circumstances did you see the diagonal noise?

SHVC-CPU-01

It was visible in just about any solid color

[fredJ]

I just tried this on an RGB-01 and it is most likely C70-75.

[copy]

My 1CHIP-02 has been looking very good with the added 10uF caps. The thick wide jail bars are gone.

I was poring over the PCB scans linked by rama, and it looks like I missed one cap earlier: C69, which appears to decouple WRAM pin 1 (VCC), and also connects to CPU pin 101 (no idea if that's VCC or something else).

So, the (hopefully) complete list for a 1CHIP should be:

Code: Select all

 C5 -- CPU
C57 -- WRAM
C63 -- VRAM
C64 -- VRAM
C69 -- CPU and WRAM

I'll add a 10uF to C69 this weekend. I'm wondering if it may help my last minor issues (very thin jail bars, and moving diagonal noise in the lower right quadrant).

[DJ Kevgeez]

Do you have a picture of the issue?

https://www.youtube.com/watch?v=4cECK0Nqtgs

[scopeu]

Hi everyone,

I have jailbars problem and I would appreciate it if you could help me for Snes mini.

At which points should I add smd cap. for snes mini? I can't see it because there is no circuit diagram. Please help me.

C90-95 or C70-75 I dont know which points?

[scopeu]

To fix the issue, I replaced all the SMD ceramic decoupling capacitors (stock: 0.1uF) with 10uF X7R ceramic capacitors on CPU, PPU1, PPU2, WRAM, and both VRAM chips, although you probably just need to replace the one on PPU2 (or S-CPUN for 1CHIPs) since that is what produces the final analog RGB output.

Hi.

I assume C90-95 are on the SNES Jr. (Mini) board revision? Which cap numbers I should stack or replace on an SNS-CPU-1-chip-01?
I am ready to make this fix, thank you so much. But i dont know which caps place, point i have to solder.

Please help me

[TooBeaucoup]

So which exact caps should be used for an SNES Mini? Should you completely replace the caps or stack them? I see the OP used X7R obviously, but there's a couple dozen variants of the 10uf on digikey, mouser, etc... I've also seen people recommend completely replacing and some that stacked the caps. How exactly do you stack them, just solder the new cap to the top of the original? I don't believe I've ever seen picture examples of that. Anyway, I've tried everything to get rid of my Mini jailbars with my Retrotink 5X but to no avail. I think changing the caps is going to be my only option at this point. I also saw the one guy above who did his like this. Which would be easier for me than dealing with little SMD caps. So I'm just curious what is the best way to go about this.



Also, is this the list for the 1-chip mini as well as the others? The info in the thread wasn't perfectly clear to me since so many people were doing theirs to different revisions. I would be humbled by any expert guidance.

C5 -- CPU
C57 -- WRAM
C63 -- VRAM
C64 -- VRAM
C69 -- CPU and WRAM

[maxtherabbit]

X7R is ok, NP0 C0G is better

just make sure they are the same size and you can stack them easily, the 3-chips use 0805 but I'm not 100% about 1-chip/jrs

leaving the original caps in place is better because it increases the range of frequencies which are bypassed

1) put new cap on top of old cap, hold it in place with tweezers or something small to press it down
2) drop a little no clean flux on it
3) touch each side of the stack with the wetted tip of your soldering iron

stacking the SMD caps is easy as falling down

[TooBeaucoup]

X7R is ok, NP0 C0G is better

just make sure they are the same size and you can stack them easily, the 3-chips use 0805 but I'm not 100% about 1-chip/jrs

leaving the original caps in place is better because it increases the range of frequencies which are bypassed

1) put new cap on top of old cap, hold it in place with tweezers or something small to press it down
2) drop a little no clean flux on it
3) touch each side of the stack with the wetted tip of your soldering iron

stacking the SMD caps is easy as falling down

Looks like the Mini is 0603. Would you be able to link one you'd recommend? It looks like most of the NP0 C0G that I've searched through are rated in pf and not uf. Not sure about those differences as I know nothing about capacitance. I just don't want to blow up my Super Nintendo. LOL!

[maxtherabbit]

https://www.digikey.com/en/products/det ... T/12324801

[incrediblehark]

I've been following this just recently as I found interference on my Super Famicom and 1-Chip SNES. and hoping this will clear it up. Would this link be a suitable part for the 3-chip snes systems?

https://www.digikey.com/en/products/det ... T/13969905

Also are there any known similar fixes for NES/Famicom? Getting rolling horizontal interference like my snes systems. Although I'm not ruling out power supply.

[maxtherabbit]

Sure it's fine

[NewSchoolBoxer]

I like that syboxez came here with a solution and asked for criticism and verification. I respect maxtherabbit's contributions but no one's being critical. Just accepting the first thing that seems to work and not considering the trade offs. Increasing the stock value of C11 by a factor of 10 to fix jailbars but in turn causing black bars in some games is good example.

Reading over the TurboGrafx 16 / PC Engine fix, he recommends replacing one 0.1 µF capacitor each in the analog and digital blocks of four with one 47x higher. Reduces resonant frequency by about 3.5x if same ESL. I judge by the write up that not all consoles have jailbars. More a symptom of an aging electronic device.

Here the proposed solution is replacing almost every AC decoupling capacitor at each chip with one 100x greater. If you want to uniformly reduce EMI, could try EMI shielding tape or EMI cages first. Aluminum foil wrapped completely with duct tape to prevent a short would be an okay budget attempt. I have to doubt that jailbars are caused by ripple voltage at each and every chip.

I think the strategy should be to measure voltage levels and slew rates from each chip on a console with no jailbars, then replace capacitors with same values and measure, then compare against another console of same hardware revision that has jailbars. Show the operating frequencies of the chips on oscilloscope, calculate the resonant frequency of the capacitor block around them and make Bode plots before and after signal passes through the block.

An ideal AC decoupling capacitor block allows the chip's square waves or analog signal through and filters out the noise. I don't want to say increasing Ceq by 100x is crazy when it seems to work but you reduce the resonant frequency by a factor of 10x, assuming same ESL.

Did Nintendo settle for 0.1 µF because, as Tim W suggested, higher valued ones didn't exist at 0603 / 0805 size in the 90s? Without theorizing too hard, jailbars could be from unfiltered harmonics that get into the ~3.58 MHz NTSC color burst. PAL could be completely unaffected.

https://circuit-board.de/forum/index.ph ... PCB-Scans/
leading to
http://stripclub.arcade-tv.de/

Should have everything

D-Do German speakers realize the double entendre? No complaints.

[maxtherabbit]

Here the proposed solution is replacing almost every AC decoupling capacitor at each chip with one 100x greater. If you want to uniformly reduce EMI, could try EMI shielding tape or EMI cages first. Aluminum foil wrapped completely with duct tape to prevent a short would be an okay budget attempt. I have to doubt that jailbars are caused by ripple voltage at each and every chip.

not replacing, adding bulk decoupling in parallel

[NewSchoolBoxer]

Here the proposed solution is replacing almost every AC decoupling capacitor at each chip with one 100x greater. If you want to uniformly reduce EMI, could try EMI shielding tape or EMI cages first. Aluminum foil wrapped completely with duct tape to prevent a short would be an okay budget attempt. I have to doubt that jailbars are caused by ripple voltage at each and every chip.

not replacing, adding bulk decoupling in parallel

Thanks, was hard for me to follow the action. I haven't a used a bulk capacitor before but this article (minus the pictures) clears things up. Preserve the small capacitor network's high frequency noise suppression and add one large capacitor for low frequency noise suppression. Also charge the small ones if needed. Still downshifts the resonant frequency but only half as much.

Significant thing is I don't see anyone in wild west of online diagrams use a ceramic for the bulk capacitor. Tantalum or electrolytic are standard due to their lower Q factor damping the circuit to avoid voltage spikes.

Article mentions avoiding the impedance spiking at the pixel frequency, which is good point. Seems to be 5.3692 MHz (4 master clock cycles) and color carrier is 6 master clock cycles NTSC. Can measure frequency response and find optimal bulk type and value at each chip.

[maxtherabbit]

Here the proposed solution is replacing almost every AC decoupling capacitor at each chip with one 100x greater. If you want to uniformly reduce EMI, could try EMI shielding tape or EMI cages first. Aluminum foil wrapped completely with duct tape to prevent a short would be an okay budget attempt. I have to doubt that jailbars are caused by ripple voltage at each and every chip.

not replacing, adding bulk decoupling in parallel

Thanks, was hard for me to follow the action. I haven't a used a bulk capacitor before but this article (minus the pictures) clears things up. Preserve the small capacitor network's high frequency noise suppression and add one large capacitor for low frequency noise suppression. Also charge the small ones if needed. Still downshifts the resonant frequency but only half as much.

Significant thing is I don't see anyone in wild west of online diagrams use a ceramic for the bulk capacitor. Tantalum or electrolytic are standard due to their lower Q factor damping the circuit to avoid voltage spikes.

Article mentions avoiding the impedance spiking at the pixel frequency, which is good point. Seems to be 5.3692 MHz (4 master clock cycles) and color carrier is 6 master clock cycles NTSC. Can measure frequency response and find optimal bulk type and value at each chip.

Electrically I'd agree that tants or electros would be more ideal. But from a physical fitment perspective you can't beat stacking the MLCCs.

[TooBeaucoup]

https://www.digikey.com/en/products/det ... T/12324801

Thanks for the link, unfortunately those are sold by a marketplace seller and require a minimum order of 4,000. LOL! All of the X7R caps on Digikey and Mouser were either way back ordered or required some absurd minimum order. I had to just pick an alternate series of ceramic caps with similar specs. We'll see what happens!

[maxtherabbit]

My fault, I checked in stock but didnt look at MOQ.

Stupid component shortage

[TooBeaucoup]

So did we have a definite list for SNES Jr? I see all kinds of lists in this thread but none that I can tell definitively say for SNES Jr. I only ask because when looking at my SNES Jr. board, I can't seem to find every specific cap number from any of these lists. My board seems to be missing some from every list I see in this thread. I know nothing about schematics, so looking at boards or wiring diagrams means nothing to my brain.

Sent from my SM-G781U using Tapatalk

[TooBeaucoup]

Alright, so without knowing for 100% certain which caps to swap/stack on my SNES Jr. I stacked a 10uf X5R capacitor onto C5 which appears to be for S-CPUN as it's the only cap that sits underneath that chip. This seems to have eliminated 99% of my jailbars. I can just barely see them still in certain shades of green, but even still, they are so faint, you'd never notice them unless you were meticulously squinting and hunting to find them. The Final Fantasy III title screen is completely black and free of bars which was awful before adding the cap. I may poke around a bit and see if I can confirm which other caps there are on the Jr. board and add more later, but for most people with a Jr. I think all you'd need to absolutely replace/stack is C5 and that'll fix 99.9% of your visual issues. That pretty much rivals the Super NT.

[maxtherabbit]

Alright, so without knowing for 100% certain which caps to swap/stack on my SNES Jr. I stacked a 10uf X5R capacitor onto C5 which appears to be for S-CPUN as it's the only cap that sits underneath that chip. This seems to have eliminated 99% of my jailbars. I can just barely see them still in certain shades of green, but even still, they are so faint, you'd never notice them unless you were meticulously squinting and hunting to find them. The Final Fantasy III title screen is completely black and free of bars which was awful before adding the cap. I may poke around a bit and see if I can confirm which other caps there are on the Jr. board and add more later, but for most people with a Jr. I think all you'd need to absolutely replace/stack is C5 and that'll fix 99.9% of your visual issues. That pretty much rivals the Super NT.

Spoiler

awesome! congrats on the fix

[TooBeaucoup]

Alright, so without knowing for 100% certain which caps to swap/stack on my SNES Jr. I stacked a 10uf X5R capacitor onto C5 which appears to be for S-CPUN as it's the only cap that sits underneath that chip. This seems to have eliminated 99% of my jailbars. I can just barely see them still in certain shades of green, but even still, they are so faint, you'd never notice them unless you were meticulously squinting and hunting to find them. The Final Fantasy III title screen is completely black and free of bars which was awful before adding the cap. I may poke around a bit and see if I can confirm which other caps there are on the Jr. board and add more later, but for most people with a Jr. I think all you'd need to absolutely replace/stack is C5 and that'll fix 99.9% of your visual issues. That pretty much rivals the Super NT.

Spoiler

awesome! congrats on the fix

I'm actually surprised this isn't more talked about. I checked my two fat SNES consoles just for the hell of it, which are both non 1-chip and they both exhibit minor jailbars as well. It's weird this isn't more widely discussed.

Sent from my SM-G781U using Tapatalk

[NewSchoolBoxer]

'm actually surprised this isn't more talked about. I checked my two fat SNES consoles just for the hell of it, which are both non 1-chip and they both exhibit minor jailbars as well. It's weird this isn't more widely discussed.

Sent from my SM-G781U using Tapatalk

Finally getting an oscilloscope next week. Going to take a closer look at things. Adding a large capacitor filters low frequency noise. In theory, 90s new console mostly didn't have that problem. Aged PPUs may not have as clean a rise and fall time as they used to that adds noise and we know very well that capacitors lose capacitance and increase resistance with age.

I have a 2-chip SNES and now a 1-chip SFC. I haven't noticed jailbars on either but then I don't examine with a fine tooth comb. Issue probably is widespread on 25+ year old console but may be hard to notice. As for not more widely discussed, I think 90-98% of people play on blurry Composite that hides defects to this day and know nothing else. Upscale to HDMI or VGA from Composite with $20 converter. Most threads are about jailbars on HDMI that probably don't show on CRT.

I think the greater issue is the analog to digital conversion (ADC). Specifically the phase of the sampling rate. A ~1° sampling mismatch will work at first. First pixel, second pixel, etc. still are scanned within their time intervals but each sample is a little more off. Finally at around 180°, one luma-colorburst pair get dropped and the rest shift down a spot. Result is one jailbar in the middle of the screen. More than that and you get many jailbars until you cycle close to 360° where things get better.

You know what causes a greater delay and therefore more degrees of mistmatch? A LPF or any chip processing. Really any capacitive or inductive element will shift the phase but in opposite directions and can cancel out. It's possible the very large capacitor you add isn't helping by filtering low frequency noise, but rather by shifting the center frequency way down so that the phase change is smaller (or much larger) and closer to what a new console had.

[TooBeaucoup]

^^^ You may be right about many of those things. I just find it odd because Genesis and PC Engine jailbars have been discussed at great length on many sites, so I'm kind of surprised SNES isn't talked about more. I know that the fat SNES is designed quite differently than the Jr, but it is weird to me that my RGB modded Jr console showed jailbars much worse than my much older fat models. However, I suppose that could be because the 1-chip models are so razor sharp that it shows more obviously? Either way, I'm glad that I have eliminated 99% of the issue with these capacitors.

[Gunstar]

Has anybody had any luck with this fix and an NTSC-J APU-01 SNES? I've used 10uf 10v X7Rs (Murata brand from Digi-key) and it's not really changed anything, maybe a bit fainter?

I put them over C70-75. SNES also has 470uf 16v on the stock Voltage Regulator and 22uf on the DC output circuit.

[IgnacioF]

I just tried this on an RGB-01 and it is most likely C70-75.

Dear Fredj, can this be done with electrolytic capacitors? For now I have only access to those. Problem is, that electrolytic capacitors are "polarized" and I have no idea which side is positive and which negative on those tiny little caps on the Snes!

Cannot find any diagram showing either.

If you know anything please let me know! I will be SO THANKFUL

Best regards,
IgnacioF

[IgnacioF]

I just want to explain to those who do not know much about capacitors:

The capacitor has to be ceramic as in NON polarized
The capacitor has to be 10UF, 6.3v
The capacitor SIZE has to be 0805

In essence, in this thread nothing is really explained much, and I ended up buying a 10uf, 6.3v but in a much SMALLER size, as no one here explained that the 0805 number refers to size.

I hope this info helps, I know it will.

[Cuttsj]

Hey folks,

I have a 1 chip 01 SNES using scart with a retrotink 4k.

I replaced the voltage regulator with a 78S05, and it did help with the main vertical bar in the middle, but I still had smaller jailbars.

I added a 470uf / 6.3v leaded capacitor as described here from RetroRGB https://www.retrorgb.com/snesverticalline.html.
It came in this kit.
https://console5.com/store/nintendo-sne ... odels.html

It did help!

Here are the results.
Before.

After.

Capacitor

[SMLMcKenzie]

I would like to resurrect this thread as I performed the "Jailbar Fix" on my Super Famicom and would like to share my results.

I am quite surprised that this fix is not talked about more often as it really works!

Just to add more context, my setup is a Super Famicom SNS-CPU-RGB-01 2 chip model. The photos that were taken below are of the Super Famicom connected to a Sony 75" TV via a Retroscaler2x in line doubling mode and a cheap Aliexpress S-Video cable. I am using a SNES USB to barrel jack power supply of dubious quality that came with the console. I had previously already fitted a 1000uF 16v capacitor to the voltage regulator to get rid of the big line down the center of the screen which was successful but unfortunately highlighted the dreaded jailbars.

Here is a photo of what the output from my Super Famicom looked like before I applied the fix:



As you can see the jailbars are quite noticable.

I ordered a 10 pack of 10uF 6.3V 0805 capacitors from eBay for about £3. As a bit of a soldering novice I decided to solder 1 capacitor at a time and test the Super Famicom before soldering another. The capacitors I soldered to were C70, C71, C72, C73, C74 and C75. Here is a photo of one of the capacitors soldered on top of the existing capacitor at C70 on the motherboard.



Here is a photo of the image after fitting the capacitor to C70:



As you can see an instant improvement after only adding 1 capacitor (ignore the dark dots above Mario's head as these were only picked up by my phone camera and not visible with the naked eye). The lines are still just visible but greatly improved.

I continued adding capacitors to C71, C72, C73, C74 and C75 and each time there was a very slight improvement.

Here is a photo of the final result with all capacitors added:



To my eye the jailbars are completely gone. The remaining noise on the image I would imagine is coming from either the power supply or the cheap S-Video cable. I think it is time to invest in a better RGB video cable!

It is a pretty easy mod to perform I reckon. Just make sure you have a pair of tweezers as I don't think you would be able to do it otherwise the capacitors are so small!

Thank you to syboxez for starting this thread and figuring all of this out. This fix really works and makes a noticeable difference.