NeoVGA: Lagless, pixel-perfect line doubler for Neo-Geo MVS and AES

mikejmoffitt

mikejmoffitt

Mickey's Coach
Joined
Feb 6, 2014
Posts
578
Argh - Rev 2.0 boards fixed issues of the first one, but the longer clock distance and more complicated path is definitely causing some minor issues. Just in time for a new revision:

554b58c67a73c0783fecba833dc116b9.png


I got it SERIOUSLY small now.
 
Xian Xi

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,750
Nice, also I read in the other thread that the board does both VGA and component video?
 
mikejmoffitt

mikejmoffitt

Mickey's Coach
Joined
Feb 6, 2014
Posts
578
Yep. Through two switches, you can output in one of these four modes:

-480p VGA
-480p YPbPr
-240p RGB
-240p YPbPr
 
mikejmoffitt

mikejmoffitt

Mickey's Coach
Joined
Feb 6, 2014
Posts
578
The extra-advantage is that the YPbPr isn't generated from an analogue conversion, so it's not susceptible to additional noise from an extra amplification stage, or inperfectly-tweaked pots, old caps, etc. that can give off color conversion.
 
aha2940

aha2940

AH, A, COLUMBIAN!,
Joined
Dec 15, 2013
Posts
2,528
Man, this really looks to be a very nice add-on for all of use neo lovers. Thanks for your efforts Mike, the extra output modes are very appreciated. From what I understand, the VGA mode is suitable for a computer monitor, am I correct?

Thanks!
 
mikejmoffitt

mikejmoffitt

Mickey's Coach
Joined
Feb 6, 2014
Posts
578
Arrgh, I'm going a little crazy here. The good news is that the new PCB has more or less eliminated all jailbar interference. The bad news is that the clock doesn't seem to be stable! In some circumstances, it'll end up skipping clocks, resulting in weird artifacts and a skewed image. I'm seeing what I can do to stabilize it on my MV-1C, but it's acting strange. I will see if my MV-2F acts the same way.

If I can't solve it for this iteration of hardware, I'm not sure what to do. This project just keeps getting more and more expensive every time I have to iterate on it. I think I'm going to have to sell off some of my extra Neo systems soon...

EDIT: Well, it's perfect on my MV-2F. I have a late generation 4-slot I can try it on too, and I'll be getting my MV-1FZ soon. I'm willing to believe my MV-1C is a bit wonky, as it's the one I Had to repair and deal with the latent vblank interrupt issues on. Really glad to see the MV-2F run it without any issues, I was getting pretty worried.
 
Last edited:
DaytimeDreamer

DaytimeDreamer

Southern Pounce.,
Joined
Jul 22, 2005
Posts
747
mikejmoffitt said:
Arrgh, I'm going a little crazy here. The good news is that the new PCB has more or less eliminated all jailbar interference. The bad news is that the clock doesn't seem to be stable! In some circumstances, it'll end up skipping clocks, resulting in weird artifacts and a skewed image. I'm seeing what I can do to stabilize it on my MV-1C, but it's acting strange. I will see if my MV-2F acts the same way.

If I can't solve it for this iteration of hardware, I'm not sure what to do. This project just keeps getting more and more expensive every time I have to iterate on it. I think I'm going to have to sell off some of my extra Neo systems soon...

EDIT: Well, it's perfect on my MV-2F. I have a late generation 4-slot I can try it on too, and I'll be getting my MV-1FZ soon. I'm willing to believe my MV-1C is a bit wonky, as it's the one I Had to repair and deal with the latent vblank interrupt issues on. Really glad to see the MV-2F run it without any issues, I was getting pretty worried.
Click to expand...

Good news. Have you done any tests on AES hardware?
 
Xian Xi

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,750
Some MV1Cs require a sync cleaner. I've had a few like that some also just wonky even if sync is fine. I'm guessing it's one of those. If you need, I can send you an MV1C to double check. Just shoot me a PM.
 
mikejmoffitt

mikejmoffitt

Mickey's Coach
Joined
Feb 6, 2014
Posts
578
Well, I did just now realize that I've been using the JAMMA edge CSync for the MV-1C. On others, I'm finding the output from the graphics chipset before it goes through resistors or any buffering transistors (if any). I can hunt down the original signal on the MV-1C and see if it helps the situation.

Next I'll try wiring up my 4-slot.
 
G

GTRetro87

Kabuki Klasher
Joined
Aug 20, 2014
Posts
127
This is looking AWESOME. Hope the MV-1C issue is figured out cause I want to install this in my CMVS!
 
mikejmoffitt

mikejmoffitt

Mickey's Coach
Joined
Feb 6, 2014
Posts
578
Wiring constraints can be tough. It works nearly perfectly on my MV4FT3, but on some scenes the large changes in pixel data seem to interfere with the clock signal, causing skewed points in the image. This is not due to the sync line having issues, but is in fact the clock line. The MV-1C problems are an ultra-extrapolated version of this problem. It doesn't always happen, though. Power rail stability is important too.

RBJLrwG.jpg

vWHq3mb.jpg


EDIT: WIth some shortening of wires to the clock and better organization, it's been rock solid on the MV4FT3. I am going to see about giving the board its own stable 5V regulator and see if that helps at all.
 
Last edited:
Xian Xi

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,750
I remember someone saying that Altera chips are very picky if they don't get their required voltage so a regulator like you said might help a lot and would be ideal since MVS mobos fluctuate on the main rail.

Also, What are the dimensions of the small NeoVGA?
 
mikejmoffitt

mikejmoffitt

Mickey's Coach
Joined
Feb 6, 2014
Posts
578
The new board is 2.61x2.01 inches (66.32x51.05 mm).

The 5V only powers the DAC, and then my 3.3V regulator. I have a lot of filtering, but even still my own stage may help.
 
B

buaku

Hardened Shock Trooper
Joined
Dec 3, 2007
Posts
437
So very awesome!

mikejmoffitt said:
EDIT: Well, it's perfect on my MV-2F. I have a late generation 4-slot I can try it on too, and I'll be getting my MV-1FZ soon. I'm willing to believe my MV-1C is a bit wonky, as it's the one I Had to repair and deal with the latent vblank interrupt issues on. Really glad to see the MV-2F run it without any issues, I was getting pretty worried.
Click to expand...

Think it should work fine on a 6-slot?
 
mikejmoffitt

mikejmoffitt

Mickey's Coach
Joined
Feb 6, 2014
Posts
578
Well, by shortening my clock wire a LOT I've gotten it to be completely stable on my MV4FT3. Hopefully that's the end of that! Once I get more parts I'll populate another board and try the MV-1C once more.

Let's have a picture of it working well:

yAMexJw.jpg


The 6-slot should be fine. PRO-B0, NEO-B1 and NEO-GRZ chipsets are all tested and have support.
 
Last edited:
B

buaku

Hardened Shock Trooper
Joined
Dec 3, 2007
Posts
437
Nice! Now I just gotta figure out how to power that thing again.
I built a supergun once, and then gutted it when I was trying to build a better one. Never quite got around to finishing that.

Will the wire length matter much going from the NeoGeo to your board?
 
You must log in or register to reply here.
Top