Help please - getting garbled graphics

bacteria

n00b
Joined
Oct 23, 2010
Posts
28
Ok, thought i'd see if my MVS-1B single slot NeoGeo MVS board worked. I understand from the manual that this board is rated at 3 amps at 5v and 1 amp at 12v; so to power the system i'm using a power supply from mains that outputs 4.16 amps at 12v, taking the 12v line directly to the NeoGeo and getting 5v from three 7805 regulators in parallel (which can output up to 1.5 amps each, so should be able to provide around 4 amps).

I know 7805's aren't efficient, I think they work at about 60% efficiency. The NeoGeo is probably the only system i'm going to have issues with in regards to amperage requirements.

The issue is that i've tried two game carts and both give the same results:

24faaeq.jpg


As you see, garbled graphics and also a band of black scrolls too.

I don't know what the issue is - I figure it might be one of the following scenarios, however I figured you guys would probably know the answer and help poor ol' Bac get his NeoGeo board working...?

* Not enough amperage, if the 7805's work at 60% efficiency then it might mean that i'm not getting 3 amps out of a 4.16 amp 12v supply. As the 12v line is for the audio requirements, I tried the system without the 12v line and got the same result, so might not be that - ?

* The MVS board I got from a chap in the USA, the cart from the UK and the PSone screen i'm using for testing (RGB mode): my understanding is the game carts are region free, and if the output was NTSC instead of PAL then i'd expect an image, squashed and scrolling.

* Something dead on the board - checked the capacitors, none are brown on top or misshapen to indicate they are dead, board visually looks fine.

* Dead board?? It was sold to be as "untested" hence cheap, although from a source that seemed ok

This is a pic of the system rigged up, the crocodile clips are soldered directly to 6 amp wires so not using the normal thin cable in crocodile clips.

e87biv.jpg


Wired up as follows: (numbers are the traces on top, on the side with all the components, alphabetical letters are underside):

1: Gnd (pins 1 and 2 are shared)
3: 5V (pins 3 and 4 are shared)
6: 12v
12: Rgb
13: rgB
14: Gnd (video)
N: rGb
P: C-Sync

Didn't rig up the start, coins, direction or action buttons, or audio; all I wanted was a screen image initially (and not getting it at the moment).

Bottom line is if I can't get this board to work i'll abandon the NeoGeo in my project as i'm not paying about £80 for a replacement board on ebay and then a multicart for another £80 - I don't mind getting a multicart if my current board works, but not to pay double that!

Being the board it is, all the chips are soldered onto the board, the battery is a circular one raised off the board (so not a leaky battery), no traces appear to be broken or damaged on the boards.

So guys, i'd welcome your comments to help me get this board working, if in fact it is able to work. What does the screen image above indicate - faulty board, insufficient amperage, etc?
 
Last edited:

FA-MAS

Kula's Candy
Joined
Oct 3, 2009
Posts
299
Looks wired up correctly. So the PSOne display will take pure RGB? All the mods I've seen that use taht screen are for consoles with composite video.

I believe you can also flip dip 1 on the MVS board and boot it without a game to get into the hardware menu.
 
Last edited:

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
Why don't you ditch the audio power to check if the board works first. If it still has those blocks then it's a ram issue. Bios issue would be mostly pink blocks iirc.
 

Hewitson

Metal Slug Mechanic
Joined
Mar 6, 2008
Posts
2,198
You're not providing enough current. I suggest you ditch that dodgy setup and invest in a proper arcade power supply which can easily be bought for $20.

And please, don't solder directly to the JAMMA connector.

if the output was NTSC instead of PAL then i'd expect an image, squashed and scrolling.
PAL and NTSC don't exist in RGB.
 

blueraven

n00b
Joined
Aug 16, 2009
Posts
24
My initial guess would be a RAM issue, I've had similar visual problems with NEC hardware and this has been the culprit.

If this IS the case, I would recommend talking to channelmaniac for new RAM...

Good Luck with the power issue.
 

bacteria

n00b
Joined
Oct 23, 2010
Posts
28
FA-MAS - PSone screen can input composite and RGB, which is handy.

Xian Xi - I did ditch the audio power (12v line) on last attempt, same result. I noticed on a different thread you mentioned about RAM issue, what if anything can be done on the basis that the chips are all soldered into place and not push-in chips? If there isn't anything I can sensibly do, then i'll just sell my MVS carts and put it down to experience - pity, would have been nice to have NeoGeo in my setup.

Hewitson - Soldered to the JAMMA connector for convenience, this board isn't going into an arcade machine after all.

The power setup isn't "dodgy" as you put it - as i'm powering many different console systems I need a lot of voltages off one supply: 1.9v, 3.3v, 3.43v, 5v, 8.5v, 12v, 15v, -3v, -5v; these are achieved off various regulators, step-up and step-down. I don't want a stack of different transformers and no point throwing money at getting an arcade power supply if the board is dead!
 

bacteria

n00b
Joined
Oct 23, 2010
Posts
28
Traces all look fine.

Tested with and without the 12v line, no difference. 7805's providing solid 5v and are cold to touch so not getting hot (have 3 in parallel and all sat on a thick and large heatsink). Screen still gives the same image.

Read on the forum here to also try all the dip switches as "on" to see if it gives a RAM message, didn't - no difference in fact. From what i've read, looks like bad RAM, yes? ... in which case, the board is toast, yes?
 

bacteria

n00b
Joined
Oct 23, 2010
Posts
28
I've tried with the dip switch 1 to "on", tried with a cart inserted and without, makes no difference, get scrolling small blocks of colours with strange characters in them as per photo above. Nada.
 

Hewitson

Metal Slug Mechanic
Joined
Mar 6, 2008
Posts
2,198
It is dodgy when the maximum current it is capable of outputting is less than the device needs!
 

channelmaniac

Mr Neo Fix-it
15 Year Member
Joined
Nov 29, 2005
Posts
4,275
What is the voltage to the RAM chips on the board? Measure it with a voltmeter. If the 7805s can't supply enough current then your voltage will drop.

If they are solid then it's time to start checking for gouged traces and other physical signs of damage before you go replacing chips.

RJ
 

bacteria

n00b
Joined
Oct 23, 2010
Posts
28
With the system running, getting 5v off the 7805's and they are cool to touch too; which proves the 7805's are not overheating and cutting out and are also providing 5v to the circuit under load.

Can't see any gouged traces on the board.

Looks like the board has dead/faulty RAM from what i've read.
 

Kyuusaku

B. Jenet's Firstmate
Joined
Mar 26, 2003
Posts
419
-It's not a good idea to put regulators in parallel (same as batteries), you can think of it as a short between the slight voltage differences. For increased current you should use a current bypass transistor, circuits for this will be in the 7805 datasheet.

-With a linear regulator current in = current out. If you require 5V @ 3A (15W), it will draw 12V @ 3A = 36W, 16W burned up as heat (a shitton). Using a switching regulator will transform POWER thus give you a better (and much needed) efficiency. BTW, a MVS does NOT require such current and your supply while inefficient is sufficient. There is very little chance your board isn't operating due to current or voltage.

-There's obviously a CPU issue, that's the first thing to repair. It could be bad components (not likely) or it could be traces/dry solder joints/bad vias (very likely). Fix that first then diagnose/treat any other problems.
 
Last edited:

bacteria

n00b
Joined
Oct 23, 2010
Posts
28
Indeed, as you've indicated, the 7805's are very inefficient indeed, however as this project is to run off mains power and not batteries, it's not a big issue. I'll look into seeing if I can find a 3 amp or 2.5 amp 5v switching regulator from 12v DC if I can find one (ideas/link to one?). As you say, the voltage isn't an issue, it's a faulty board. Given the fault could be RAM or could be CPU, it's toast as far as i'm concerned unfortunately.

I can get the NeoGeo 16 in 1 gamepack for PS2 which has a couple of the games I really want (Neo Turf Masters as one), and can also get some of the Metal Slug series on that format / original Xbox. The only NeoGeo games i'd be wanting for are either repeated on other systems (eg Puzzle Bobble) or mid-interest games. I would have paid £90 for a 138-in-1 multicart (you can take off the games you don't want from the menu, I checked with the seller), however don't fancy paying for another MVS board, at £90 off ebay; that's £180, and I can play about 2/3 of the games via the SNK packs for the PS2....cheaply. Pity. I'll keep my eyes open for a cheap MVS board, otherwise i'll sell my faulty board as "faulty/spares" and the games too.
 
Last edited:

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
Ram is cheap and simple to install. It's a good learning experience when you have to replace ram.
 

Kyuusaku

B. Jenet's Firstmate
Joined
Mar 26, 2003
Posts
419
Not to mention it's probably not bad RAM XD Components aren't so fragile. This is almost certainly a really easy fix.

It's hard to find high current switching regulators, I wouldn't bother.
 

bacteria

n00b
Joined
Oct 23, 2010
Posts
28
Yeah, found a 3 amp 5v switching regulator from Texas Instruments and then discovered it needed two 200uf ceramic capacitors - ceramic not electrolite, 200uf is far higher than you normally get and they would have been very expensive to buy, so didn't bother. Found a circuit to make one using other components, will give it a go sometime.
 

jew90

Geese's Thug
Joined
Nov 6, 2010
Posts
273
Strange output.

With audio connected and volume turned up full can you hear it clicking?

Sometimes with a bad connection the cpu will try and reset the machine constantly.. (watchdog circuit) this scenario would be a hardware fault...

Either way (sometimes watchdog doesn't work) it looks like a hardware fault.but It'd be good to test it on a proper setup to eliminate your possibility.

What is the battery like on the board? Has this leaked and eaten traces away?

I'll go with Kyuusaku on this one and say it is probably an easy fix if you know what your looking for! (all of which pointed out)

If you switch all of the dip switches to on it should boot into 'work ram test'
 

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
In some cases a NeoBiosMasta can remedy this type of thing since it mounts onto the CPU itself so if there are any broken traces from the bios to the CPU the NBM will not need them.
 

distropia

SouthTown StreetSweeper
10 Year Member
Joined
Apr 28, 2010
Posts
1,075
Looking at the power requirements for the mv1b1, I have made a quick test to verify the data, and it needs less power than SNK claims:
mv1b1-realpowerreqs.jpg


As you can read in the small tester, it draws only 1.21 amps, including the RGBs to VGA adapter.
Definitely your problem is not caused by amps.

You can also search for some welding rests between chip legs. Recently I repaired one mv1FZ that had the same problem, and it was a microscopic welding rest on the graphic chip.

As Xian Xi says, the universe bios can help you troubleshoot bypassing some hardware tests.
 

bacteria

n00b
Joined
Oct 23, 2010
Posts
28
"Welding rest"? - what's that please?

I've kinda given up on the board unfortunately.

Interesting on the amps.
 

distropia

SouthTown StreetSweeper
10 Year Member
Joined
Apr 28, 2010
Posts
1,075
With "welding rests" I mean very small rests of tin that can be left floating around the PCB after you solder/desolder something if you're not careful enough. Those rests can make a merge between two chip legs and give you that blocky screen.
Sometimes the rests are so small that is difficult to see them.
I would reccomend dismantling the board, taking a deep look between every chip's legs, brushing it with non-electrostatic brush and using a PC power supply to test it again.
Tell me if it helps you.
If not, maybe the backup ram is defective. Keep in mind that is very very sensitive to electrostatic discharges.
 
Top