Moire effect / wavy lines on some resolutions

[Sublevel]

*Update: It seems i need some kind of a Low Pass Filter between the VGA ouput and JPAC input. Original post down below.


In most of the games i've tried i get this annoying moire effect when the screen is scrolling vertically (or horizontally when i have the monitor in horzontal position). I also get this ringing effect on the edge of sprites/backgrounds like the picture is too sharp. However, in games like Ketsui and Espagaluda which both run at 448x224 @ 59.17Hz i do not get the moire effect. How come? Has it something to do with integer scaling?

Please see attached logs. In ESP Ra.De. i get the moire effect but not in Ketsui.

My setup looks like this:
Dedicated Groovymame computer (Win7, Radeon 5450) inside an New Astro City > VGA > JPAC > JAMMA connector > Nanao MS929 monitor rotated vertically

Things i've already tried:

Different presets/monitor settings in VMMAKER
Replace VGA cables
Replace video card
Replace computer PSU
Replace step-down converter for the NAC
Moving cables around inside the NAC
Adjusting pots on the Nanao monitor (adjusting focus somewhat hides the moire)

[Calamity]

Just in case, try with normal resolutions instead of super resolutions. E.g. add 320x240@60 to the list of super resolutions, and generate and install the list of modes, then launch esprade.

[Sublevel]

Thanks Calamity, will try that when i get home tonight. But if i remember correctly, in the past i've tried installing normal resolutions instead of super resolutions with the same moire effect in game. However, i did notice the moire effect was gone when using normal resolutions and checking stuff in ArcadeOSD, but when starting Groovymame the moire was back again.

[Calamity]

In your logs, I can see this error:

Direct3D: Unable to find D3DXCreateEffectFromFileW

I don't know if it could have any relation to the issue. I see this very often in users' logs. I'm not sure if it's caused by not having the proper DirectX distributables or by a mismatch between GM's exectuble and the rest of files in its folder. I recommend to get the full package from my Google drive site.

[cools]

In your logs, I can see this error:

Direct3D: Unable to find D3DXCreateEffectFromFileW

I don't know if it could have any relation to the issue. I see this very often in users' logs. I'm not sure if it's caused by not having the proper DirectX distributables or by a mismatch between GM's exectuble and the rest of files in its folder. I recommend to get the full package from my Google drive site.

I'm getting this on the only machine I've updated to 203, and I'm using the full package (had to recreated my mame.ini  ). I can't confirm whether I have correct DirectX or not, it's Windows 7 Home Premium 64bit. I'm having performance/sync issues with the machine too so would like to resolve it and see if that fixes them

[krick]

Full DirectX9 package installer...
https://www.microsoft.com/en-us/download/details.aspx?id=8109

[cools]

Thanks, that fixed the message but not the problem, which I'll follow up in another thread.

[Sublevel]

Tried normal resolutions and installed the Full DirectX9 package and i still got the moire effect in game. The monitor chassis recently got recapped and the image quality is sharp as hell complete with ringing around sprite edges. When i defocus the monitor with the focus pot (or using bilinear filtering in mame) the moire/ringing is somewhat reduced, but then i also get convergence issues so that solution is a no go. Could it be that this monitor combined with JPAC/VGA input is just too sharp for it's own good? And that the resolution for Ketsui (448x224) somehow comes out as slightly blurrier than Esprade (240x320) which eliminates the moire effect? Just a wild guess.

[Calamity]

You'd better post some pictures of your real screen (and probably delete that one of a NEC monitor on your first post that's quite misleading).

Is the ringing around the sprite edges something really all around the sprites or just on the right side, like a shadow?

[Sublevel]

I will take some pictures and post later today. If i remember correctly it's all around the sprites, not like a shadow effect.

[Paradroid]

Is the ringing around the sprite edges something really all around the sprites or just on the right side, like a shadow?

Sorry to threadjack but I have the exact issue you described with a TV that I recently RGB-modded (right-side shadow that trails any sudden change of contrast or color).

Interested to hear your theory if you had one...

Sent from my SM-G955F using Tapatalk

[Calamity]

Interested to hear your theory if you had one...

My Polostar did this too, and it sucked for that. It was fixable to some extent adjusting the contrast down (or maybe the brighness?). It probably has some simple explanation from an electronics point of view, for some reason it can't resolve those sudden changes in the signal.

[Sublevel]

Oh by the way, does "filter 1" add any input lag to Groovymame?

[Calamity]

It doesn't add lag, but it degrades the image.

[Sublevel]

Please see attached pictures. The ArcadeOSD one shows the moire effect and the rest the ringing on pixel edges.

[Calamity]

I see. Thanks for the pics.

Have you tried calibrating the contrast?

[Sublevel]

Yes, and to no avail. It should also be said that my cab came directly from japan complete with the original power supply and ISO transformer still connected. I haven't rerouted any cables myself so i don't really know if everything is hooked up correctly regarding grounding etc.

[pakoman]

I was going to suggest that: maybe a grounding or degauss issue? Is there any degauss switch in the cabinet?

Sometimes magnetized stains are related to the angle between the cabinet and north and just rotating the cabinet improves a lot.

[Sublevel]

I played around with the cab this weekend and checked the grounding. I have no idea how grounding should be wired in this cab but there are a lot of green cables going from the back (inside) of the cab to different spots like the control panel, coin door etc. I moved them around a bit but didn't notice a difference. But if it was a grounding issue, shouldn't all games be affected? I want to point out that i only get this moire effect, no other noise or interference at all.

I also e-mailed Andy regarding the JPAC to hear if he had any idea and got this response: "It would normally be caused by a difference between the pitch of the resolution and the dot pitch of the CRT. The first thing to check is that the resolution is correct and matching the game. If it is, it must be the CRT. So the only way to overcome this would be to de-focus the monitor slightly or mess around with the convergence ring magnets although the latter is a risky approach."

So i tried messing around with the dot clock setting in ArcadeOSD and i can in fact "move" the moire around a bit by doing so, but it doesn't get rid of it. Any ideas how i should approach this further? Or should i just settle with defocusing the picture?

Thanks again all 

[cools]

What version of CRT_EmuDriver are you using?

(I think the ringing is a monitor setup problem, but having just spent lots of time finding a problem with my 5450 and the Crimson driver I figure it's worth asking)

[Calamity]

Ok, I've been thinking of your issue and I agree with Andy's diagnosis. Your CRT is just too sharp.

Now, you provided a clue that may lead us to a software fix. When you said Ketsui does not show the issue, I didn't notice that 448 scales on 2560 by a fractional factor, while Espagaluda is 320 pixels wide, that scales by a perfect integer factor of 8.

So it's probably fractional scaling what provides the slight blur you need to get rid of the interference pattern.

If that is true, there's something you could test. Instead of 2560x, use 2600x for all your super resolutions. Edit the user_super.ini file and replace all 2560 values by 2600. Then create and install the modes again.

I suggested 2600 because it's not a multiple of either 256, 264, 288, 320, 400, etc, so fractional scaling will be applied on the horizontal axis.

[Sublevel]

Thanks for the suggestion Calamity, but sadly that didn't work. I still get the moire effect when installning 2600x*** modes. I also tried switching over to latest release of Catalyast 12.6 and there is no difference.

I've noticed that all vertical games with a resolution higher than 320x*** is free from this issue. I tried a game with a res of 384x240 (free from moire), please see attached log if that's in any way helpful. I noticed in the log a bit further down "orientation: normal". Shouldn't that be "rotate_r" as i have my monitor rotated vertically? Also, shouldn't vertical resolutions be listed as e.g. 240*384?

[Calamity]

Shouldn't that be "rotate_r" as i have my monitor rotated vertically? Also, shouldn't vertical resolutions be listed as e.g. 240*384?

No, GM is right there. Since you are using the rotate_r preset, "vertical" resolutions are turned into "horizontal" resolutions, so 384x240 is correct.

I'll try to think something else. Meanwhile, consider installing the DirectX package linked by Krick above, so you get rid of the error message at least.

[pakoman]

What about playing a bit with the screen size (not resolution)? You will get black bars/ overscan, however you may find something balanced between screen size and moire effect.

Or even adjusting the geometry pots (Pinch Cushion, etc) of the monitor or adding a small amount of divergence to one of the colors may help.

[Evewalker]

Just to potentially add more data to help solve this issue, I have the same problem.   Ketsui and any game running at its resolution is perfect, but the wavy moire effect is present elsewhere.   I've tried everything from hardware and software swaps, cable replacements, and monitor adjustments with no improvement.

I can upload my logs later tonight, but to test, I did comparisons with a few actual PCBs which ran perfectly without the issue, while the GM versions had the disturbance.   The PCBs I tested were Donpachi and Strikers 1945.   The monitor is a Nanao MS8 29 FTB rotated vertically in an Egret 29.

[Sublevel]

Shouldn't that be "rotate_r" as i have my monitor rotated vertically? Also, shouldn't vertical resolutions be listed as e.g. 240*384?

No, GM is right there. Since you are using the rotate_r preset, "vertical" resolutions are turned into "horizontal" resolutions, so 384x240 is correct.

I'll try to think something else. Meanwhile, consider installing the DirectX package linked by Krick above, so you get rid of the error message at least.

I did install the DirectX package but i guess i still get the error message. I extracted the files and ran the setup file in the main folder.

What about playing a bit with the screen size (not resolution)? You will get black bars/ overscan, however you may find something balanced between screen size and moire effect.

Or even adjusting the geometry pots (Pinch Cushion, etc) of the monitor or adding a small amount of divergence to one of the colors may help.

If i stretch the image vertically a lot by using the pots, the moire is somewhat reduced. But i have to stretch it so much it defeats the purpose. Minimizing the image does not help.

Just to potentially add more data to help solve this issue, I have the same problem.   Ketsui and any game running at its resolution is perfect, but the wavy moire effect is present elsewhere.   I've tried everything from hardware and software swaps, cable replacements, and monitor adjustments with no improvement.

I can upload my logs later tonight, but to test, I did comparisons with a few actual PCBs which ran perfectly without the issue, while the GM versions had the disturbance.   The PCBs I tested were Donpachi and Strikers 1945.   The monitor is a Nanao MS8 29 FTB rotated vertically in an Egret 29.

Sorry to hear you also got this problem. Please do upload your logs, thanks 

[Calamity]

I'd say what Pakoman meant was to use the software stretching options in MAME's internal ui, namely "Screen Horiz Stretch" under the "Slider Controls" menu.

[schmerzkaufen]

Can you try a different monitor ?

Also can't aging caps or flyback cause all kinds of weird issues that start showing with a select number of source signals and then with time expand to all ?

I've had a couple of old sets that would start acting only with certain sources before going all senile.

[cools]

I know you said you've tried them, but logs where you are using normal resolutions would be good.

I can understand why you *might* get moire with super resolutions (integer or fractional scaled), however it's baffling that it's occurring with normal resolutions.

Pics to accompany logs too please.

[pakoman]

I'd say what Pakoman meant was to use the software stretching options in MAME's internal ui, namely "Screen Horiz Stretch" under the "Slider Controls" menu.

Well, I meant the pots in the monitor to avoid distorted scaling, but both options could do. And specially those that control the curvature of the screen (pinch cushion, pillow cushion...).

I remember from the times I used to collect jamma boards that Mame looked way darker than the same jamma board in the same arcade cabinet and gamma needed to be adjusted in Mame. That was a few years ago (AdvanceMame 0.106 dos IIRC, before old good Groovymame times ), so I don't know if Mame improved in that regard nowadays. Maybe not the case, however, the darker the colours the more likely the moire effect to appear, the brighter and the more they "bleed" the more they hide moire effects, chequered patterns, sharp pixels, etc.

Could you just check the brightness levels in both Mame and jamma are the same?  Maybe the jamma board with darker colours also produces moire patterns...

[Calamity]

I can understand why you *might* get moire with super resolutions (integer or fractional scaled), however it's baffling that it's occurring with normal resolutions.

Pics to accompany logs too please.

Yeah, that's the point. Anyway, both 320 (not necessarily other) and 2560 would be undistinguishable signal-wise (unless integer scaling is broken from some reason in the actual gpu).

I'd also like to see full screen pics of an affected game.

[Evewalker]

Here are some logs from my setup, both using Standard Resolutions, and Super Resolutions installed.

Ketsui has no moire effect at all.  Donpachi has a severe moire effect when scrolling using both Standard and Super Resolutions installed.  Using the actual PCB gives a functionally identical picture except for the lack of Moire effect and wavyness.

[Evewalker]

Could you just check the brightness levels in both Mame and jamma are the same?  Maybe the jamma board with darker colours also produces moire patterns...

This was a good idea, and I did check this.  The arcade PCB as expected was a bit brighter, but I adjusted the brightness levels to match and the wavy effect was still there, unfortunately.

[Zebidee]

Something obvious for you to try.

Make sure that your arcade monitor is grounded to nothing except itself. Make sure that it doesn't share ground with other parts of your cab.

Why? Because your arcade monitor is powered by an isolation transformer, which isolates your power supply and effectively provides a local ground for that circuit (. This way you won't get interference from the AC ground line, and it is actually safer too!

If your monitor doesn't have an ISO, then feel free to ignore this advice.

[Sublevel]

I will try and post some pics and new logs tonight. It's difficult to find the time with a 1 year old son 
Thanks to all contributing to this thread so far!

Something obvious for you to try.

Make sure that your arcade monitor is grounded to nothing except itself. Make sure that it doesn't share ground with other parts of your cab.

Why? Because your arcade monitor is powered by an isolation transformer, which isolates your power supply and effectively provides a local ground for that circuit (. This way you won't get interference from the AC ground line, and it is actually safer too!

If your monitor doesn't have an ISO, then feel free to ignore this advice.

If it was an grounding issue wouldn't all resolutions be affected? I don't think the monitor is grounded to anything special. There is a spring-like wire going around the monitor that attaches to the neckboard.

[Zebidee]

If it was an grounding issue wouldn't all resolutions be affected? I don't think the monitor is grounded to anything special. There is a spring-like wire going around the monitor that attaches to the neckboard.

It could still be grounding, just the effects not apparent in all modes. However, that is just a theory. Leave that spring-like ground wire from tube to neckboard, you need that there. Usually there should be nothing connected between monitor chassis and ISO except 2 power lines (active and neutral).

One way to test the AC ground interference theory is to plug in something to the same powerboard/outlet that has a powerful electric motor, like a mixer/blender or a drill. See if running that makes a lot of interference.

Keeping video signal wires away from high voltage areas of the chassis like the flyback is always a good idea.

I have a suspicion that this moire effect is common in Nanaos. Many Many years ago I had an Astro Cab (like a Candy) with a 29" Nanao monitor and, iirc, I am sure that I had this same moire issue. I went over everything, including grounding, and didn't get any better than small improvements. I was using an ArcadeVGA and JPAC.

My Nanao was one of those dual-res ones that could do EGA (25khz) as well as CGA, but it was always set to CGA anyway. Can't remember the model, but it would have been ms8 or ms9.

All this leans toward Andy's diagnosis earlier, that the moire has something to do with a mis-match between phosphor dot-pitch and game resolution. Maybe the standard Nanao tube has a weird phosphor mix/layout/pitch to better suit EGA? However, I don't think that there is anything wrong with your convergence. Pic look good in that regard. Your monitor is not "too sharp". Leave the convergence rings alone.

Fiddling with focus/convergence might minimise moire, but only because it is harder to see everything through the fuzz.

If dot-pitch/resolution mismatch is the cause, you might find that the problem disappears with another tube! Anyway, you can't really test that without a matching tube lying around...

I also suspect that it is not CRT_emudriver causing the moire, as ArcadeVGA causes it too (if I am recalling correctly above, please test if you have an ArcadeVGA lying around).

Which brings us back to the JPAC. Maybe you could try running signals directly to your Nanao, bypassing the JPAC, to see what happens? The picture will be faint, for sure, but am interested to see what happens to the moire.

Evewalker notes that the problem goes away when running a PCB (ie no JPAC) vs MAME:

Here are some logs from my setup, both using Standard Resolutions, and Super Resolutions installed.

Ketsui has no moire effect at all.  Donpachi has a severe moire effect when scrolling using both Standard and Super Resolutions installed.  Using the actual PCB gives a functionally identical picture except for the lack of Moire effect and wavyness.

[pakoman]

Could you just check the brightness levels in both Mame and jamma are the same?  Maybe the jamma board with darker colours also produces moire patterns...

This was a good idea, and I did check this.  The arcade PCB as expected was a bit brighter, but I adjusted the brightness levels to match and the wavy effect was still there, unfortunately.

Then if the jamma / Groovymame are outputting the same signal (except for brightness! ) the moire effect must be produced by the computer at some point...

Maybe the VGA cable / computer / etc is not shielded/grounded enough and some resolutions use more voltage than others. The computer speakers?

Have you tried adjusting slightly the V-Sync / H-Sync pots in the monitor?

From the computer side you may try a different computer, using AdvanceMame, avoiding the Jpac (direct connection VGA-Monitor), a different JPac, different electric sockets...

PS using CRT Shaders with a QHD monitor also produces moire effects

[Zebidee]

Even the best CRT monitors can show a moire effect if you put up the right test patterns.

I'm 99% convinced that the moire effect is caused by phosphor dot pitch being too large/wrong, and has nothing to do with PC, video card, PCB, JPAC, monitor chassis, grounding, AC interference or anything but the CRT tube itself.

It is another reason to avoid Nanao monitors, despite the fact that they often seem to be otherwise good (they are also difficult to repair).

I'd be on the lookout for a TV donor tube.

This site has a very simple explanation for the effect: http://www.displaymate.com/moire.html

Summary of ways to minimise moire:

1.    Some displays have "a Moiré Reduction Control" (but this might not be effective and may introduce artifacts)

 2.   Slightly de-focus the image using the monitor's focus control

3.    Adjust up contrast/brightness

4.    Use different resolutions.

5.    Avoid dim backgrounds.

6.    Avoid dithering, and avoid 16 color Windows modes, which generally have a lot of dithering. Dithering occurs much less frequently in 256 color modes and is generally absent in 32K color modes and above.

[Evewalker]

Even the best CRT monitors can show a moire effect if you put up the right test patterns.

I'm 99% convinced that the moire effect is caused by phosphor dot pitch being too large/wrong, and has nothing to do with PC, video card, PCB, JPAC, monitor chassis, grounding, AC interference or anything but the CRT tube itself.

It is another reason to avoid Nanao monitors, despite the fact that they often seem to be otherwise good (they are also difficult to repair).

I'd be on the lookout for a TV donor tube.

The problem is that no actual arcade PCBs, multiboards, or JAMMA test pattern generators display any sort of issue on the same monitor.  All display perfectly.  This problem is only coming up with a PC connected via a J-pac and GM.

I'm going to try a few different video cards tonight to see if that changes anything.  I have both 4000 and 5000 series AMD cards to try out, and I'll report back.

[Zebidee]

The problem is that no actual arcade PCBs, multiboards, or JAMMA test pattern generators display any sort of issue on the same monitor.  All display perfectly.  This problem is only coming up with a PC connected via a J-pac and GM.

OK, sorry, maybe its a bit confusing because I thought someone (you!) was reporting that they got got the moire effect with PCB earlier:

Could you just check the brightness levels in both Mame and jamma are the same?  Maybe the jamma board with darker colours also produces moire patterns...

This was a good idea, and I did check this.  The arcade PCB as expected was a bit brighter, but I adjusted the brightness levels to match and the wavy effect was still there, unfortunately.

But I see that you are NOT getting moire with PCBs now, at least with Donpachi and Strikers. Just a note about those games - they are vertical games with lots of horizontal lines (rotated to be vertical ofc), and Sublevel was reporting that he/she didn't get the moire effect with games that had many horz lines (more than 320 horz lines?).

It may also just come down to individual video modes (as people report moire in some modes, not others), and your PCBs are probably creating a different video mode (at least slightly different) to the one GM is choosing when you run the game, if you understand what I mean.

What I'm trying to say is: not getting moire when running those PCBs is not conclusive, we need more evidence.

I'd be very interested to hear about your moire outcomes from running PC/GM direct to the monitor, bypassing the JPAC.

[Zebidee]

Is the ringing around the sprite edges something really all around the sprites or just on the right side, like a shadow?

Sorry to threadjack but I have the exact issue you described with a TV that I recently RGB-modded (right-side shadow that trails any sudden change of contrast or color).

Interested to hear your theory if you had one...

Paradroid, a few possibilities, but sounds like the TV is having a little trouble responding to voltage level changes. Do you have any "blooming" (slow to change when screen goes from very bright/dark/vice versa) and geometry issues (between dark/light areas) as well?

I'm guessing that you have already 75-ohm terminated the RGB lines and put a capacitor on the end, near the jungle IC? Your capacitor is probably a 0.01 - 1uF ceramic?

If you are using 0.01 or 0.1uF you might find that changing the capacitor value up a little helps, up to 1uF ceramic. If that doesn't help, try putting an additional cap (this time an electrolytic, start from 1uF or 10uF and work up from there until 680uF or so) in-series on each RGB line, before the ceramic cap. Positive towards jungle (negative towards nearest ground).

If that gives no joy, try attaching the negative lead of the electrolytics directly to ground and seeing how that goes (this will also help filter AC interference).

This cap combo will help to keep the voltage levels more constant. The electrolytic will hold more charge but release it "slowly", the ceramic will hold less charge but release it quickly. With the right mix, this helps your TV to respond better, more smoothly anyway, to voltage changes.

[Sublevel]

OP reporting back here. I did some more testing and found some interesting results. I hooked up some of my consoles (SNES, PC Engine, PS1) to the Nanao trying different games/resolutions and NONE showed neither the moire or the ringing effect. The picutre quality was absolutely glorious! A nice, stable picture without any noise or interference. It was fantastic to see what this monitor actually is capable of. I hooked up the consoles like this:

Console > RGB Scart > Sync Strike > VGA cable > Extron RGB unit > VGA out (buffered) > JPAC > Jamma harness > Nanao monitor

So i guess the issue is not the monitor? I have also noticed on my GM computer setup that i get the moire/ringing even on the Windows desktop (640x480) before launching Groovymame. So i feel the issue is either the computer/components itself or a CRT Emudriver issue. Things i've already tried regarding the computer:

Moving it away from the Astro City/Nanao monitor
Tried plugging the PSU into different wall sockets
Tried a different PSU
Tried three different GPUS (one HD 4350, two 5450s)

Here's a video of the moire:
https://ufile.io/ihz8c
Only 480p, why i don't know.

[Calamity]

Have you tried passing the VGA from your PC through the Extron and then into the JPAC?

Also, have you always tried this through the VGA output connector of the video card? Have you tried using the DVI-I connector through an adapter, instead?

[Zebidee]

Have you tried passing the VGA from your PC through the Extron and then into the JPAC?

An excellent idea, guessing the Extron must be filtering a lot of noise out of the signal.

[Sublevel]

Have you tried passing the VGA from your PC through the Extron and then into the JPAC?

Also, have you always tried this through the VGA output connector of the video card? Have you tried using the DVI-I connector through an adapter, instead?

Yes and yes. No difference, still moire/ringing present. Trying the DVI-I ouput with a dvi to vga adapter was one of the first thing i tried. Guess i could try it again with a different adapter.

[Zebidee]

Those DVI/VGA adapters are just straight pass-through passive adapters anyway (the analog VGA signals for RGBs are all there already on the DVI-I head, the adapter just passes these through to relevant VGA pins), so don't expect to any big difference from using one of those when the Extron does nothing.

[Calamity]

Those DVI/VGA adapters are just straight pass-through passive adapters anyway (the analog VGA signals for RGBs are all there already on the DVI-I head, the adapter just passes these through to relevant VGA pins), so don't expect to any big difference from using one of those when the Extron does nothing.

Yeah but usually the DVI connector is mounted right on the pcb while the VGA connector uses a ribbon cable to the pcb that is more prone to interferences, that's why I suggested that.

Sublevel, I'm out of ideas. Do you have another computer to try with? If possible without an AMD card/CRT Emudriver.

[Sublevel]

I (kind of) did my homework before getting into all of this and made sure all the GPUS i was going to use did not have the ribbon cable for the VGA output. All of them are mounted on the pcb directly. I do not wanna mess around with my main computer so no, i do not have another one to do tests with. I guess a complete format / re-install is all there is left to try right now.

Thanks all for the input. Will report back.

[pakoman]

You don't need to do a complete format, just can try Groovyarcade or boot from an USB drive. For testing this should be more than enough.

[Calamity]

You don't need to do a complete format, just can try Groovyarcade or boot from an USB drive. For testing this should be more than enough.

Yeah, good point.

[Zebidee]

I'd suggest bypass or decoupling capacitors for the signal lines to remove noises, similar to what I mentioned for Paradroid above, but wonder why/if the Extron hasn't done this already.

Which brings us back to dot-pitch for phosphor clusters, resolution mis-matches and your Nanao tube.

[Paradroid]

I'd suggest bypass or decoupling capacitors for the signal lines to remove noises, similar to what I mentioned for Paradroid above, but wonder why/if the Extron hasn't done this already.

Ground loop isolators can also be helpful with such problems. Would be an expensive experiment if you couldn't arrange to borrow one first though...

[Sublevel]

Once again, thanks for the input!

I was about to give up when it dawned on me. I just couldn't understand why i got an interference free signal connecting my consoles to the Nanao so i did some more testing. It turns out my Scart switch (Hydra) has some kind of low pass filter which makes the signal clean when outputting the consoles to the Nanao. So when i include the scart switch in my computer / Groovymame setup i get a nice, clean picture! Yay!

So the question now is: is there any other kind of low pass filter device/circuit i can include in my GM setup, so that i dont have do chain it together with my scart switch?

Like: GPU > VGA cable > LPF > JPAC > JAMMA harness > Nanao

[Zebidee]

So the question now is: is there any other kind of low pass filter device/circuit i can include in my GM setup, so that i dont have do chain it together with my scart switch?

Funny you should ask this - I almost wanted to suggest you put a low-pass filter on each of the RGB lines, but didn't want to get too complicated.

A simple low-pass filter would involve a resistor in-series and then a capacitor to ground (negative to earth, away from display). I'm really not sure what values you would need to use - maybe start with something like 75 ohms resistor and 220uF capacitor. Electrolytic is probably best as they pass AC current better. The cap value might need to be even higher like even 1000uF or more,  but start with 75R/220uF.


PCRGB----->-----RESISTOR---------------|-------DISPLAY
                                                           |
                                                      CAPACITOR
                                                           |
                                                       GROUND

PS - This sounds fussy, but is really quite simple and a lot cheaper than an Extron!

[Zebidee]

A more robust approach might be using a video-amp IC such as the THS7314 or THS7316. These have a low-pass filter built into them. You can also buy tiny little PCBs to mount them on!

It is very simple. Check out the retrorgb article on how to set them up and use:
http://retrorgb.com/thsamps.html

You'd need to get the IC, the PCB mounting board, a 0.1uF ceramic capacitor and three 75 ohm resistors. I'm not sure which IC (THS7314 or 7316) would be best for you. You can read the discussion on the site (and on shmups forum) and decide... or get both, and try them out.

I think the most appropriate place for this would be between PC VGA out and JPAC. You can get 5v to power it from VGA pin 9 (or from USB or make up a molex connector).

[Sublevel]

Thanks for the suggestion! I think i'll go with the THS7314. So just to clarify, i need to splice a VGA cable and then add the amp in the middle?

GPU output > vga cable *amp* vga cable > JPAC?

[Paradroid]

It turns out my Scart switch (Hydra) has some kind of low pass filter which makes the signal clean when outputting the consoles to the Nanao. So when i include the scart switch in my computer / Groovymame setup i get a nice, clean picture! Yay!

Serious?! So, you finally solved the issue?! Awesome.

[Zebidee]

Thanks for the suggestion! I think i'll go with the THS7314. So just to clarify, i need to splice a VGA cable and then add the amp in the middle?

GPU output > vga cable *amp* vga cable > JPAC?

Yes. You'll still want the JPAC (or a video amp) to boost the signals to the 3-5v expected for an arcade monitor.

Even with the retrorgb instructions, I'd go a bit further than that and follow the diagram (figure 2) on page 6 of the datasheet at
http://www.ti.com/lit/ds/symlink/ths7314.pdf

This means putting both a 0.1 - 0.01uF ceramic cap AND a 22uF - 100uF electrolytic cap between 5v and ground near the Vcc in pin, as shown in that diagram. This will help filter AC interference from your power input.

I would also suggest adding termination resistors (75 ohm seems a good starting choice) onto the IC's RGB inputs, again as per that diagram (figure 2). Terminate them to video ground.

I am slightly concerned that the THS7314 outputs amplified video signals in the 0 - 2vp-p range (roughly doubled, 2V/V) so that it can supply inputs to 2 receivers at once. However, I think that the JPAC can handle those inputs and still give consistent outputs. I also note that between the 75 ohm load resistors on the video outputs, and 75 ohm termination resistors on your display or JPAC, you get a voltage divider that will cut voltage by half anyway. If you want to try to reduce the signal output levels anyway, you should try to do it at the inputs rather than the outputs. But I would just give it a go first.

Finally, I suggest you get a few of these ICs and the little PCBs, so that you can scrap and start again if needed.
Depending upon your skillz, you might want to try the AC-coupled circuit (figure 3) as well. Hopefully you have a range of capacitors and resistors lying around too.

[Sublevel]

It turns out my Scart switch (Hydra) has some kind of low pass filter which makes the signal clean when outputting the consoles to the Nanao. So when i include the scart switch in my computer / Groovymame setup i get a nice, clean picture! Yay!

Serious?! So, you finally solved the issue?! Awesome.

Yep, found a solution at least  Now i can fully enjoy my Groovymame cab! GM is such an awesome emulator.

Thanks for the suggestion! I think i'll go with the THS7314. So just to clarify, i need to splice a VGA cable and then add the amp in the middle?

GPU output > vga cable *amp* vga cable > JPAC?

Yes. You'll still want the JPAC (or a video amp) to boost the signals to the 3-5v expected for an arcade monitor.

Even with the retrorgb instructions, I'd go a bit further than that and follow the diagram (figure 2) on page 6 of the datasheet at
http://www.ti.com/lit/ds/symlink/ths7314.pdf

This means putting both a 0.1 - 0.01uF ceramic cap AND a 22uF - 100uF electrolytic cap between 5v and ground near the Vcc in pin, as shown in that diagram. This will help filter AC interference from your power input.

I would also suggest adding termination resistors (75 ohm seems a good starting choice) onto the IC's RGB inputs, again as per that diagram (figure 2). Terminate them to video ground.

I am slightly concerned that the THS7314 outputs amplified video signals in the 0 - 2vp-p range (roughly doubled, 2V/V) so that it can supply inputs to 2 receivers at once. However, I think that the JPAC can handle those inputs and still give consistent outputs. I also note that between the 75 ohm load resistors on the video outputs, and 75 ohm termination resistors on your display or JPAC, you get a voltage divider that will cut voltage by half anyway. If you want to try to reduce the signal output levels anyway, you should try to do it at the inputs rather than the outputs. But I would just give it a go first.

Finally, I suggest you get a few of these ICs and the little PCBs, so that you can scrap and start again if needed.
Depending upon your skillz, you might want to try the AC-coupled circuit (figure 3) as well. Hopefully you have a range of capacitors and resistors lying around too.

Thanks, i'll take that into consideration. I've ordered a couple of ICs, caps and resistors and will report back later. This will really test my skills as the most advanced solder work i've done so far is soldering a 3v PSU to pin 16 on a SCART cable..   

[Evewalker]

Great to hear you figured out a solution, Sublevel.  I've been playing with all sorts of applications for the same problem with no luck, so thanks for figuring this one out.

Let me know how your construction of the amp goes!

[Sublevel]

A more robust approach might be using a video-amp IC such as the THS7314 or THS7316. These have a low-pass filter built into them. You can also buy tiny little PCBs to mount them on!

It is very simple. Check out the retrorgb article on how to set them up and use:
http://retrorgb.com/thsamps.html

You'd need to get the IC, the PCB mounting board, a 0.1uF ceramic capacitor and three 75 ohm resistors. I'm not sure which IC (THS7314 or 7316) would be best for you. You can read the discussion on the site (and on shmups forum) and decide... or get both, and try them out.

I think the most appropriate place for this would be between PC VGA out and JPAC. You can get 5v to power it from VGA pin 9 (or from USB or make up a molex connector).

I'm gonna power it from an USB port on the computer. Can i use that same USB cable to ground the THS7314?

Great to hear you figured out a solution, Sublevel.  I've been playing with all sorts of applications for the same problem with no luck, so thanks for figuring this one out.

Let me know how your construction of the amp goes!

Well, the solution i know works for now is via the scart switch. Let's just hope a LPF IC like the THS7314/16 works just as well. Will keep you updated! 

[Zebidee]

I'm gonna power it from an USB port on the computer. Can i use that same USB cable to ground the THS7314?

Yes, you'll need to. They don't mention it on the retrorgb page, but I think you'll want to hookup your video ground directly to the PCB ground as well. If you have termination resistors on the inputs, ground it to video ground but connect video ground to the PCB ground. For the RGB signals to move through the circuit there must be a current return path to the PC, and forcing it to go through USB ground could cause unwanted interference.

[Sublevel]

I'm gonna power it from an USB port on the computer. Can i use that same USB cable to ground the THS7314?

Yes, you'll need to. They don't mention it on the retrorgb page, but I think you'll want to hookup your video ground directly to the PCB ground as well. If you have termination resistors on the inputs, ground it to video ground but connect video ground to the PCB ground. For the RGB signals to move through the circuit there must be a current return path to the PC, and forcing it to go through USB ground could cause unwanted interference.

So, not including caps / resistors something like this?

[Zebidee]

So, not including caps / resistors something like this?

Yes, just like that

[Sublevel]

Wow, this was way more difficult than i thought. The IC/board is so small and the VGA wires so thin! I do not have the skills/tools for this kind of soldering. I'll have to ask someone to make one fore me 

[Zebidee]

Wow, this was way more difficult than i thought. The IC/board is so small and the VGA wires so thin! I do not have the skills/tools for this kind of soldering. I'll have to ask someone to make one fore me 

I had hand surgery about 3 months ago and was back to soldering within a day, with my left hand still in a splint.

I was going to get a bunch of these ICs and give it a go anyway, mostly because I want to try them in a RGB=>YPrPb (component) circuit I've been building to use with modding TVs that can't be RGB modded. That is to say, I've been YPrPb-modding TVs.

My RGB=>component circuit already works. I have been using three C945 transistors for the signal amp stage, but these THS7314/7316 ICs would be much neater and the signal filter may improve my picture quality. I'm interested to see results with both kinds of the chip. The THS7314 may be better for Low-pass filter purposes.

That's my long way of saying... if you don't mind waiting, I'll happily solder up one of these retrorgb RGB amps, test it and send you one!

Unfortunately I don't have them on hand, so I'll have to order and wait for delivery (most likely from China). My last package of electronics from China, which arrived a few days ago, took about 7 weeks to get here. Normally packages don't take that long, but they commonly do take 2-3 weeks or so and we are coming into Chrismas. Just so you know what to expect.

I see that some people are selling these on ebay for N64 RGB mods. The one below looks nice and neat, SMD design on custom PCB using the THS7316:

https://www.ebay.com/itm/Nintendo-64-N64-RGB-Amp-THS-7316-7314-upgrade-mod-kit-NTSC/153258549660?hash=item23aeebe19c:g:W3AAAOSwS5JbcZvl:rk:21:pf:0

They would also work for PC or anything with more-or-less standard analogue RGB signals. Amplification might not be enough for arcade monitors, but that's not an issue with a JPAC. I'm interested to give it a go anyway (ie no JPAC etc) because I've been having issues with my Ultimarc video amp boosting signals *too much* for certain arcade monitors of mine.

UPDATE: Just found this on ebay, a THS7314 already soldered onto a tiny PCB chip:
https://www.ebay.com/itm/1PCS-THS7314DR-7314-ON-DIP-ADAPTER/321292514333?epid=0&hash=item4ace86b41d:g:xFoAAOSwFHZap5px:rk:7:pf:0

[Sublevel]

Wow, this was way more difficult than i thought. The IC/board is so small and the VGA wires so thin! I do not have the skills/tools for this kind of soldering. I'll have to ask someone to make one fore me 

I had hand surgery about 3 months ago and was back to soldering within a day, with my left hand still in a splint.

I was going to get a bunch of these ICs and give it a go anyway, mostly because I want to try them in a RGB=>YPrPb (component) circuit I've been building to use with modding TVs that can't be RGB modded. That is to say, I've been YPrPb-modding TVs.

My RGB=>component circuit already works. I have been using three C945 transistors for the signal amp stage, but these THS7314/7316 ICs would be much neater and the signal filter may improve my picture quality. I'm interested to see results with both kinds of the chip. The THS7314 may be better for Low-pass filter purposes.

That's my long way of saying... if you don't mind waiting, I'll happily solder up one of these retrorgb RGB amps, test it and send you one!

Unfortunately I don't have them on hand, so I'll have to order and wait for delivery (most likely from China). My last package of electronics from China, which arrived a few days ago, took about 7 weeks to get here. Normally packages don't take that long, but they commonly do take 2-3 weeks or so and we are coming into Chrismas. Just so you know what to expect.

I see that some people are selling these on ebay for N64 RGB mods. The one below looks nice and neat, SMD design on custom PCB using the THS7316:

https://www.ebay.com/itm/Nintendo-64-N64-RGB-Amp-THS-7316-7314-upgrade-mod-kit-NTSC/153258549660?hash=item23aeebe19c:g:W3AAAOSwS5JbcZvl:rk:21:pf:0

They would also work for PC or anything with more-or-less standard analogue RGB signals. Amplification might not be enough for arcade monitors, but that's not an issue with a JPAC. I'm interested to give it a go anyway (ie no JPAC etc) because I've been having issues with my Ultimarc video amp boosting signals *too much* for certain arcade monitors of mine.

UPDATE: Just found this on ebay, a THS7314 already soldered onto a tiny PCB chip:
https://www.ebay.com/itm/1PCS-THS7314DR-7314-ON-DIP-ADAPTER/321292514333?epid=0&hash=item4ace86b41d:g:xFoAAOSwFHZap5px:rk:7:pf:0

You, good sir, are too kind! I will happily let you make me one in exchange for moneys. There is no rush at all, after Christmas sounds just fine! Thank you so much!

[Zebidee]

You, good sir, are too kind! I will happily let you make me one in exchange for moneys. There is no rush at all, after Christmas sounds just fine! Thank you so much!

No problem at all  I just ordered what I need, will let you know when it comes in. Fortunately I have plenty of other things to keep me busy in the meantime 

[jimmyj1979]

I have just found this thread and I think it might be applicable to my issue

http://forum.arcadecontrols.com/index.php/topic,159278.0.html

I am yet to explore this solution in detail, but hopefully steers me in the right direction

[Zebidee]

I promise @sublevel I'll focus on soldering that THS7314 amp up soon.

Stuff has arrived, a little while ago, but while waiting everything got busy here. Many guests (paying and otherwise) over the New year period too (our part of the world is very popular this time of year). I have several projects at incomplete atm, including 3 arcade cabs and a swimming pool, but finding it difficult to focus on any of them. Fortunately, doing this basic video amp progresses more than one project so I'll work on it tonight.

Besides, I think Sublevel is probably pretty keen to see results. As am I.

[LazyAce]

I seem to be having some (similar?) interference issues, any chance you could take a look at my thread and lmk your thoughts? Many thanks: http://forum.arcadecontrols.com/index.php/topic,159526.0.html

[Zebidee]

Apologies to Sublevel and anyone else following this - I've been busy since mid-January mostly with a swimming pool project (large one). Also ...[edit: rant removed]

Wow, 4 paragraphs off-topic whinging. Back on-topic.

Good news is that I just cleaned up the electronics workspace in my shed, all the bits I need for the THS7314/7316 RGB amp are here now, I just need to do the darn thing.

LazyAce: a lot of people seem to have issues with the Nanao monitors in Astro cabs, just like you describe on your thread. It may be unfair to blame the JPAC, but would be good to definitively rule it out as a possible source of the problems too.

[LazyAce]

LazyAce: a lot of people seem to have issues with the Nanao monitors in Astro cabs, just like you describe on your thread. It may be unfair to blame the JPAC, but would be good to definitively rule it out as a possible source of the problems too.

The monitor seems fine on PCBs and Grant (gunblade) has looked at and recapped the chassis, but he confirmed it worked perfectly before a recap. I get the issue with 2 different jpacs across three different devices: 2 Pcs and 1 Pi, so I'm assuming it could be the JPAC. I still have the 5 and 12V lines connected to the Jamma edge, but the JPAC is powered separately. I'm wondering if this could be the issue? The weird thing is, the wobble and interference go away after 40 minutes.

[Zebidee]

You could always try powering the JPAC from the JAMMA 5v power. I guess you'd have to hack a cable as they normally get it via USB or PS2. Not sure Andy would approve 

[Sublevel]

Apologies to Sublevel and anyone else following this - I've been busy since mid-January mostly with a swimming pool project (large one). Also ...[edit: rant removed]

Wow, 4 paragraphs off-topic whinging. Back on-topic.

Good news is that I just cleaned up the electronics workspace in my shed, all the bits I need for the THS7314/7316 RGB amp are here now, I just need to do the darn thing.

LazyAce: a lot of people seem to have issues with the Nanao monitors in Astro cabs, just like you describe on your thread. It may be unfair to blame the JPAC, but would be good to definitively rule it out as a possible source of the problems too.

No apologies needed! Take your time with the amp, no worries. I'm rather busy taking care of my one-year-old right now so there's not much time for gaming. But i do still want to try this amp