N64 noise on solid white screen ... ??

[Syntax]

I swear mario 64 doesnt have the noise on an emulator.

You dont know how releaved i am knowing im not the only one that can see this. Thanks dude

[Thomas83lin]

No problem. I'm starting to think this noise might be normal for this mod(no offense). I would like to try a simple rgb mod to compare to though. It could be the N64 is just a really noisy console

[Syntax]

Im a week off trying a dac based RGB mod (Bortis design)

Ill report back after i have fit it.

[Thomas83lin]

Ok, sounds good. I would like to know myself.

[kel]

I have been following this thread for the past few days as I also have these same issues with this mod but have not had any time to post yet. I noticed the noise straight off the bat compared to my early NTSC N64 with Borti's simple RGB amp which has no noise whatsoever although I forgot to check for interference on Paper Mario/Superman with that mod.

I tried a few things to get rid of the interference on Paper Mario/Super man on Tim's board. I tried the RC filter that Tim suggested and tried a couple of ferrite beads also that I had lying around but the only thing that helped was the LDO method suggested in this thread and even that doesn't completely get rid of it.

The 3.3v is very noisy but only at the RGB board everywhere else on the console is clean and even if I connect a separate wire to a clean 3.3v supply on the console it becomes noisy at the board so I don't think it is just being transfered through the ribbon cable. It seems that it enters somewhere on the board itself. The noise looks like it might be from the 12.5Mhz DS# signal.

3.3v at RGB board (channel 1) and 12.5Mhz DS# (channel 2.)



3.3v at power supply (channel 1) and 12.5Mhz DS# (channel 2.)



I will also be trying out one of Borti's boards within the next week (N64Advanced). I will see how that one goes compared to this one with regards to these issues.

EDIT: Just tested Paper Mario on my early N64 with borti's simple RGB amp and if you look very very closely then the interference is actually there but barely noticeable so it seems that even the VDC-NUS DAC suffers to a certain extent or maybe it has something to do with the THS7374 being as though this mod and Tim's RGB board both use this same amp, IDK. The UltraHDMI doesn't have this issue at all though as far as I can see.

[Thomas83lin]

@kel that's interesting, maybe viletim might have some ideas on that. Myself I decided to shorten my ribbon cable to almost nothing in a attempt to get some result. I got nothing, its the same as it was before. My mod does looks a lot neater though.
http://www.mediafire.com/view/ynta6x84u ... C00536.JPG

btw does anyone know of a color test rom for the N64?, right now I'm running 240p test suite nes rom through the everdrive64. I can say it still works good to see the noise clearly, but I'd like something made for the N64.

[borti4938]

The 3.3v is very noisy but only at the RGB board everywhere else on the console is clean and even if I connect a separate wire to a clean 3.3v supply on the console it becomes noisy at the board so I don't think it is just being transfered through the ribbon cable. It seems that it enters somewhere on the board itself. The noise looks like it might be from the 12.5Mhz DS# signal.

3.3v at RGB board (channel 1) and 12.5Mhz DS# (channel 2.)

Wow - that's a lot. How much is it?
Noise peaks around 600mV-700mV peak to peak for a duration of 66% each cycle?

Here is what I've measured on my oscilloscope:
N64 Advanced:

• 5V rail: appr. 200mV peak to peak
• 3.3V rail: appr. 340mV peak to peak
• 2.5V rail: appr. 200mV peak to peak
• 1.2V rail: appr. 200mV peak to peak

N64RGBv2:

• 5V rail: appr. 200mV peak to peak
• 3.3V rail: appr. 200mV peak to peak
• 1.8V rail: appr. 150mV peak to peak

I can also see that it's related to the 12.5MHz DS# clock as one can clearly see peaks in the spectrum at multiples of 12.5MHz.

Unfortunately I haven't taken any pictures - were just quick measurements without having an USB drive lying around. I'll take some on Monday.

Low value resistors would help prevent reflections on the video data lines. The problem is they need to be installed on the source end of the ribbon cable to be effective.

You are right - my fault! I was thinking of my modding board where I put them on the board to have it more convenient for installation. More effective they are at source. This can be clearly measured. Although they are also (measurable) helpful to have them on board if one have not them at source.

[kel]

Wow - that's a lot. How much is it?
Noise peaks around 600mV-700mV peak to peak for a duration of 66% each cycle?

Here is what I've measured on my oscilloscope:
N64 Advanced:

• 5V rail: appr. 200mV peak to peak
• 3.3V rail: appr. 340mV peak to peak
• 2.5V rail: appr. 200mV peak to peak
• 1.2V rail: appr. 200mV peak to peak

N64RGBv2:

• 5V rail: appr. 200mV peak to peak
• 3.3V rail: appr. 200mV peak to peak
• 1.8V rail: appr. 150mV peak to peak

I can also see that it's related to the 12.5MHz DS# clock as one can clearly see peaks in the spectrum at multiples of 12.5MHz.

Unfortunately I haven't taken any pictures - were just quick measurements without having an USB drive lying around. I'll take some on Monday.

Yeah it is quite a lot.
That's about right give or take. On that measurement it was around 700mV-800mV peak to peak but having the 2 probes connected adds around 100mv, so yeah it would have been around 600mV-700mV peak to peak.

With the LDO supplying 3.3v to the RGB board it goes down to about 600mV-700Mv with both probes connected and around 500mV-600mV when only probing 3.3v so roughly 100mV less that without it but still quite a lot. The duration is around 55%-60% of each cycle before it seems to level out again.

LDO supplying 3.3v at RGB board (channel 1) and 12.5Mhz DS# (channel 2)



Not sure what else can be done to get this noise any lower. I was wondering if there could maybe still be a GND issue? I have added the GND strap to the heatsink from the board but each GND on the board only uses a small via to connect to the GND plane on the bottom side although I can't see under the CPLD to be fair.

Your measurements on your boards are quite a lot lower. It will be interesting to see what I get once I have your N64Advanced board installed.

[Syntax]

Replacing all the surface mount caps made the visual noise much lower.

Maybe we need more decoupling near the DAC?

[viletim]

I made some noise measurements on the console that I used for the installation guide.
https://etim.net.au/temp/forum/n64rgb_3v3_noise/

One measurement is at 100 ns per division, full bandwidth. The other is at 10 ms with the oscilloscope's 20 Mhz low pass filter on. Repeated again in a different position.

The low frequency noise is from the motherboard or power supply, while the high frequency is from the data cable or the N64RGB board.

kel,
How did you make this measurement exactly? The noise level you measure is incredibly high.

Syntax,
Did you replace them with larger values?

[Syntax]

I used the same values for all caps that I replaced.

[kel]

I made some noise measurements on the console that I used for the installation guide.
https://etim.net.au/temp/forum/n64rgb_3v3_noise/

One measurement is at 100 ns per division, full bandwidth. The other is at 10 ms with the oscilloscope's 20 Mhz low pass filter on. Repeated again in a different position.

The low frequency noise is from the motherboard or power supply, while the high frequency is from the data cable or the N64RGB board.

kel,
How did you make this measurement exactly? The noise level you measure is incredibly high.

Syntax,
Did you replace them with larger values?

In that case maybe it's something I'm doing wrong. To be fair I've only been learning how to use this oscilloscope over the past couple of weeks or so and it wouldn't surprise me if I've not quite managed to get my head around one of the settings or something.

I used the settings as per the pictures that I posted, 50ns per division, full bandwidth at position B on your photo's. One thing that I did notice after viewing your pictures is that my volts/div setting may have been a bit high at 1v per division. After lowering this to 100mV per division I'm getting a measurement of about 280mV-380mV peak to peak with the LDO supplying the board, which is close to Borti's N64 Advanced board measurement that he posted earlier in the thread.

[viletim]

I used the same values for all caps that I replaced.

To clarify, you replaced the four 100n ceramic capacitors with another set of four 100n caps?

In that case maybe it's something I'm doing wrong. To be fair I've only been learning how to use this oscilloscope over the past couple of weeks or so and it wouldn't surprise me if I've not quite managed to get my head around one of the settings or something.

Did you have your probe and oscilloscope set to 10:1 attenuation? Did you use a probe with a ground spring (as opposed to a ground wire)?

[Syntax]

Sorry Tim you misunderstood, I only replaced the surface mount electrolytic capacitors that are present on the top side of the N64 Motherboard.

Which 4 ceramic 104's are you speaking of?

[kel]

In that case maybe it's something I'm doing wrong. To be fair I've only been learning how to use this oscilloscope over the past couple of weeks or so and it wouldn't surprise me if I've not quite managed to get my head around one of the settings or something.

Did you have your probe and oscilloscope set to 10:1 attenuation? Did you use a probe with a ground spring (as opposed to a ground wire)?

Yes, both probe and oscilloscope were set to 10:1 attenuation. I used the probes that came with the oscilloscope which have a ground wire with a crocodile clip at the end, shown in the picture below. It is attached to the probe with a type of push on clip.



EDIT: Heh, just found the spring clips that came with the probes , will try again later and report back.

EDIT#2: Wow that made a massive difference, I'm now getting 50mV-60mV peak to peak at the RGB board instead of the 500mV-600mV that I was getting with the ground wire with crocodile clip and at the power supply I'm getting 20mV peak to peak. Thanks Tim for pointing me in the right direction.

Do you think the 50mV-60mV peak to peak high frequency noise could be the cause or am I barking up the wrong tree? I can't seem to get a good measurement of the low frequency noise, what setting did you use to take that measurement?

[thebigcheese]

Mostly I am replying here so that I can get notified of results, but also so I can contribute that I have also experienced some interference and it drives me crazy. It's worse on some games than others, for sure, but Kirby 64 has been my main test game: https://www.youtube.com/watch?v=jK-JdsBtJV8. It is most prominent on RGB, so I mostly assumed it was something to do with Tim's N64 RGB. I tried shortening up the ground plane as I had installed this before that became an official step in the install, switched to using a shielded Mogami wire running from the N64 RGB to the output to prevent any possible interference there... no luck. I have been thinking about trying Voultar's board as I have an earlier N64 that could use it, but if that's not the issue then I would rather address the actual problem then throw new mods at it. Having said that, now that I know the heavy-duty wire didn't help, I could probably stand to tidy things up anyway. I have not tried recapping the system or PSU yet, but it would be pretty cheap to do that, so if that helps others, keep me posted so I can order a kit

[Thomas83lin]

I've been reading up on the SSDS3's RGB video noise, and the fix they implemented for themselves. They mentioned it was caused by the video amp referencing digital ground instead of analog ground. I wonder if the noise we have been getting is caused by a similar problem? Thinking about it I'm not sure how one would separate the two, but just throwing out some ideas.

[leonk]

as far as i see, there is only 1 common ground in the n64.

[Thomas83lin]

as far as i see, there is only 1 common ground in the n64.

Yeah I'm not sure how it works or how one keeps them separated, but I also no longer believe that's the problem either.

I believe the background noise is coming from the framemeister itself. I can't speak for everyone or people using the ossc, but I'm almost sure my noise is coming from the framemeister and not the mod kit. I can somewhat confirm this cause I also get it on completely different systems like the Ps1. Only difference on the Ps1 is that it mostly only shows up on the blacks, honestly I had never seen the noise on the Ps1 until I looked for it using 240ptestsuite but it is there.
btw all my rgb cables come from retro console accessories so they should be completely in spec.

Here's a thread I found which also discusses the problem and I can confirm changing the A/D setting does move the noise but really doesn't get rid of it. viewtopic.php?f=6&t=52828

So there you have it, one thing I would like to add is that the 3.3v ldo does fix the interference problem that was showing up on Paper Mario and Superman64. The background noise in my opinion is unrelated and is being caused by the conversion process from the framemeister.

[leonk]

the noise I see is also visible on my PVM. Even when I use composite cables in N64.

Not XRGB problem.

[Thomas83lin]

the noise I see is also visible on my PVM. Even when I use composite cables in N64.

Not XRGB problem.

I also have a pvm, do you have a video capture maybe our noise is different

[thebigcheese]

the noise I see is also visible on my PVM. Even when I use composite cables in N64.

Not XRGB problem.

Same. It's not as noticeable because CRTs are better at masking it, but it's definitely still there.

[Thomas83lin]

the noise I see is also visible on my PVM. Even when I use composite cables in N64.

Not XRGB problem.

Same. It's not as noticeable because CRTs are better at masking it, but it's definitely still there.

any particular game were it shows up bad. I'd like to test to see if I can see it myself on a pvm. I'll try to take some video captures if I can. I do have a startech pexhdcap and framemeister\hd60 pro

[thebigcheese]

Kirby 64. I linked to a direct capture earlier, but it's also noticeable on PVM.

[leonk]

paper mario. right at the start when it looks like a movie theatre. Look at the solid white. You'll see some random horizontal noise in the white.

[Thomas83lin]

paper mario. right at the start when it looks like a movie theatre. Look at the solid white. You'll see some random horizontal noise in the white.

I uploaded a capture of the Paper Mario intro earlier in this thread arter installing the 3.3ldo, I'm curious if it still looks the same as yours? http://www.mediafire.com/file/vx83kp3c5 ... mpress.mp4

My noise shows up mostly on large solid dark colors like the everdrive 64 menu. The noise I see doesn't show up or at least I can't see it on a pvm. It also isn't as noticeable on the capture card.

@thebigcheese
On kirby64 I couldn't really see any noise but I'll upload a capture and if you see it or not let me know. I'm mostly just curious if we are getting the same noise or not.
http://www.mediafire.com/file/zrl7gexhh ... apture.mp4

[Ikaruga11]

Go back to paper mario. In the intro there's going to be a spot where most of the screen is white (movie theatre screen). Pay close attention to the white. Let us know if you see random horizontal noise in there.

Sorry for the long delay. I just checked Paper Mario again and the startup screen does indeed have jailbars, although it's almost completely unnoticeable. This may be due to the fact that I'm using S-Video, and thus the blur filter is also there, the white bloom from the lack of a voltage regulator, and the low TVLine count of the 20" Insignia CRT. It's almost impossible to see on my setup, but oddly Pokémon Stadium has far more visible jailbars on black/white transition screens.

[Rupert H]

I'm having exactly the same issue. I recently did a complete overhaul of my AV set-up and it's about as fine-tuned as I can imagine getting it.

• Sony PVM-2730QM (Upgraded the speaker wires to OFC and added pin connectors to ensure everything is nicely shielded)

• RGB Pakapunch Pro cables

• gscartsw_lite 1.5

• RGB-modded US launch N64 (I don't even know how this was modded back in '96, but obviously there's no THS7374)

After all this I'm getting the same screen issues you describe. Jumping pixels on Super Mario 64, intense flicker on white screens (most apparent on Paper Mario and Banjo-Kazooie). I'll be watching this thread with great interest for a solution.

[leonk]

I think this noise has always been there. The only reason I notice it now is because:

- RGB mod (vs Svideo/composite)
- Borti's deblur filter (N64 blur function very good at reducing it)
- XRGB mini / LCD TV vs PVM/CRT (again .. very good at capturing and amplifying high frequency noise vs PVM/CRT just blurring it away)

[Thomas83lin]

I think this noise has always been there. The only reason I notice it now is because:

- RGB mod (vs Svideo/composite)
- Borti's deblur filter (N64 blur function very good at reducing it)
- XRGB mini / LCD TV vs PVM/CRT (again .. very good at capturing and amplifying high frequency noise vs PVM/CRT just blurring it away)

I've compared my paper mario capture at 10% speed to yours and I just don't seem to get all that flicker that you get. I do have some slight horizontal bars on the transitional white screen. I also don't get that noise on the letter M at all either. That's weird?

Imho I believe most of the flicker not the bars could be solved by using the ldo 3.3v and the noise on the letter M looks like data line noise. but I can't explain why you get it on crt using composite or svideo cause you shouldn't.

The noise I'm getting, that I see on the Everdrive 64 menu is not the same as what you guys are experiencing. I still believe my noise is being caused by the framemeister. I can see it on certain solid colors on the Ps1\N64 and Sega Genesis.. Adjusting the A\D setting on the framemeister, I can actually move this noise around on the screen so my noise is definitely related to the framemeister or similar.

The noise is hardly noticeable but it is there, It very well may be that the framemeister is very sensitive to noise and there may be some slight noise in the rgb that isn't noticeable on a pvm.

So I'm at a loss, but I do think its time for me to let it go and move on.