CMVS Garbled Graphics

[bobrocks95]

I only got my Timeharvest CBox CMVS about a week ago, but it's already giving me trouble.

Graphics appear garble, specifically like vertical stripes/tiles are repeating themselves- lots of duplicated letters and weird half-sided SNK logos. This occurs as early as the Unibios screen so it isn't related to the cart that's in the system.

Last game I played was a pretty dirty copy of KOF '98. I cleaned it as best as I could before it ever touched the CMVS, but it's still the dirtiest cart I own. Next game I put in started showing the problems.

Any immediate ideas, or just send it back to Timeharvest? I figure that would take several weeks.

[thegreathopper]

Can you adjust the voltage? It may need adjusting up to 5.0volts or the carts may be dirtier than you think.

[bobrocks95]

Can you adjust the voltage? It may need adjusting up to 5.0volts or the carts may be dirtier than you think.

No voltage control, at least not externally.

I should also clarify that all my games were working fine before, all 11 of them. Until I cleaned KOF again and got it running finally.

[Josh128]

Bob, do you ever get any RAM error messages on startup? Sounds like it could be VRAM or work RAM related. Verifying the +5V would be a very easy check and because of that I recommend that as well, but its unlikely to be whats causing your problems.

If you boot the system with no cart inserted, do you get the crosshatch pattern (not sure if that applies with the Universe BIOS)? Crosshatch indicates all self-diags passed on a stock BIOS.

MVS systems from China are harvested from various sources and many non-working boards are repaired / attempted to be repaired. A close examination of the motherboard would reveal if any of the RAM chips were swapped out. Im 99% certain at least one of my backup RAM ICs was replaced. I currently get the same backup RAM error message intermittently during bootup and have some replacement RAM and quad OR gate ICs coming to attempt repair.

If you have a ROM programmer you could obtain a 27c1024 IC and flash the diagnostic BIOS on it and slap that in your system to help pinpoint the issue. I have programmer and would gladly flash it for you if you send me a chip.

[bobrocks95]

Thanks for the offer Josh, I may need to do something like that depending on what Timeharvest says about it.

Here is a short video I made showcasing the problem:

https://youtu.be/SMBCjjvUZAc

[Ryoandr]

https://wiki.neogeodev.org/index.php?ti ... ches#C_ROM

CA4 stuck / cut ?
I don't have a MV1C board so I can't tell you where it goes exactly.
If anything like MV1B, it's goes from CHA Board pin A18 (lower cart slot), through a 75R resistor and to NEO-GRC2 (all in one graphic chip) pin 8. MV1C uses another all in one revision, NEO-GRZ and CA4 pin isn't documented.

[Josh128]

While the glitches definitely look like what can be caused by a CROM issue, CROMs are located in carts, not the Neo MVS mainboard. Do the graphic glitches actually occur with no cart inserted? I see non-hex characters show up after you boot with no cart, but they dont appear graphically glitched. If they glitch with no cart, that seems to rule out a CROM problem, if they only glitch with a cart, it could very well be your issue.

If you are keen to it, open the CMVS and press around with a non-conductive object such as a pencil eraser or so. Press on and around different ICs on the board and in areas where traces connect to them and see if by pressing you can affect the graphics at all. You may be able to find a cracked trace or solder joint or identify the culprit IC by doing this.

How long did it take to get your Timeharvest unit? I would suspect as things stand now it will be quite a wait for a replacement and / or if you must ship back for analysis or repair.

[bobrocks95]

While the glitches definitely look like what can be caused by a CROM issue, CROMs are located in carts, not the Neo MVS mainboard. Do the graphic glitches actually occur with no cart inserted? I see non-hex characters show up after you boot with no cart, but they dont appear graphically glitched. If they glitch with no cart, that seems to rule out a CROM problem, if they only glitch with a cart, it could very well be your issue.

If you are keen to it, open the CMVS and press around with a non-conductive object such as a pencil eraser or so. Press on and around different ICs on the board and in areas where traces connect to them and see if by pressing you can affect the graphics at all. You may be able to find a cracked trace or solder joint or identify the culprit IC by doing this.

How long did it take to get your Timeharvest unit? I would suspect as things stand now it will be quite a wait for a replacement and / or if you must ship back for analysis or repair.

It took a long time to come in so I'm definitely going to poke around a bit myself first.

Those "hex codes" without a cart in are supposed to be like "Press B for soft reset, Press D for Test Mode" or something along those lines as I recall. Definitely not gibberish.

[bobrocks95]

Pressing on the graphics chip (NEO-GRZ as this is an M1VC) changes the output, though it seems a bit random. Might be something internal to the package unfortunately. I'll try traces next.

EDIT: Also, I should really know how to do this, but how would I verify 5V are coming in with a multimeter?

[Josh128]

Pressing on the graphics chip (NEO-GRZ as this is an M1VC) changes the output, though it seems a bit random. Might be something internal to the package unfortunately. I'll try traces next.

EDIT: Also, I should really know how to do this, but how would I verify 5V are coming in with a multimeter?

If pressing on the graphics chip does anything at all, the most likely cause is that there is a cracked solder joint or trace connected to that chip. Its *possible* that the chip itself has got issues and can exhibit them when pressed on, but thats less likely I think.

As far as checking the 5V, just set your multi-meter to Vdc and find the wires going from the power input jack to the motherboard. The wire connected to the outside of the jack should be negative, the wire connected to the center should be positive. Simply touch your positive lead to the positive and the negative lead to the negative. Do this while the system is powered on to ensure the voltage is read under load.

[Josh128]

I was able to replace the backup RAM chip connected to CPU address pins D7-D0 (RAM 2) on the CPU and Voila!! Fixed the backup RAM errors I was continually getting. Havent had a single one since replacing the IC. Stoked!!

Error


Old RAM


New RAM!

[bobrocks95]

I was able to replace the backup RAM chip connected to CPU address pins D7-D0 (RAM 2) on the CPU and Voila!! Fixed the backup RAM errors I was continually getting. Havent had a single one since replacing the IC. Stoked!!

Nice to see! I'll be replacing my VRAM chips as soon as "new" ones come in from Timeharvest and hopefully fixing my problems.

[Josh128]

lol crazy shit--after the RAM replacement, the system worked fine, my wife and I played about an hour of Puzzle Bobble, absolutely no issues. Later I booted SamSho2 and prior to playing I set the difficulty to 4 and then set the time and calendar. I then booted to game and played for about 10 minutes, then the game halted with some game graphics and some garbled error. I was able to make out "work", possibly as in work RAM I suppose.

I then powered down and powered back and now I only get a garbled pinkish screen for a couple seconds, then, if my 161 in 1 cart is inserted, the pinkish screen loops through a few frames of animation continuously. I have uni-bios incoming but other than that not sure. Calendar error maybe?
Here is a video of what happens when I power on with a cart in. If I power on with a cart out, I still get a pink screen, but it doesnt animate after a few seconds like it does with the cart inserted.

https://www.youtube.com/watch?v=GuBki4e ... e=youtu.be

Unbelievable!

[Josh128]

Well, I resolved "pink garbled screen" issue. Turns out the BIOS completely failed-- like mid-game. My initial attempt at installing a UniBIOS to diagnose it didnt fix it because I had some CPU socket solder joints not making good contact. I didnt know this until I tried to install the UniBIOS (via NeoBIOSMasta) on a known working MVS. Turns out, a completely dead or disabled BIOS, at least on MV1C models, will always give you some version of the "pink garbage" as seen below.

Its odd to have a BIOS just die mid game like that, but since installing the UniBIOS its all been 100% good, no RAM errors or anything to speak of.

[Josh128]

I was able to replace the backup RAM chip connected to CPU address pins D7-D0 (RAM 2) on the CPU and Voila!! Fixed the backup RAM errors I was continually getting. Havent had a single one since replacing the IC. Stoked!!

Nice to see! I'll be replacing my VRAM chips as soon as "new" ones come in from Timeharvest and hopefully fixing my problems.

Bob, how are you coming along on your Timeharvest unit? Did you hear back from them?

Im currently sitting on several NeoJamma (shout out to Cyril!) supergun kits for MV-1F units that I cant do anything with because the lot of 3 units I ordered from China a month and a half ago still have not arrived. Hoping they havent gotten lost.

[bobrocks95]

I received replacement RAM chips, but had to send them and the console off to a modder since I'm not really equipped to desolder the original ones. Maybe could've bought a cheap heat gun but I figured I'd screw up something anyway, so just waiting to hear back now on what the actual problem was (still guessing the two VRAM chips on the underside of the board).

Serves as a good reminder though that MVS hardware was commercial, and left on 24/7 until the boards ran into the ground (I am not someone who is used to working with old arcade boards obviously)... Wonder if anyone will ever make new hardware for it?

[Josh128]

Bob, this evening I just fixed (from what I can tell) another MV1C with a very similar issue to what you were having. The problem presented as the pics below:





After examining the board, I found where two traces had been previously patched and the BIOS IC had been replaced. Powered it up connected to the TV and pressed around the board and cartridge. I was able to to get the glitches to change and even clear by flexing the board in certain ways. I was then able to isolate it to when I would push or pull a bit on the cartridge. Figured there was a bad joint or trace so I looked carefully around the cartridge connector, top and bottom, but didnt see anything obvious. I then re-flowed all the cartridge connector pins (there are A LOT of them!) on the MVS and also reflowed the pins on the RAM closest to the connector. This fixed the issue and I was no longer able to produce it again.

Unless you get a video RAM error on startup, this kind of problem is MUCH more likely to be caused by a broken trace to or from the cart connector or bad/cracked solder joint. This unit was not giving any self-test errors. I'd be willing to bet the cause of your problem is something similar to what I found on this unit, because the glitches are very similar.

[aknot]

He Josh128!

Have you solved the issue with glitches?
I had a color pallet issue and was thinking that GRZ is dead as one pin was destroed. I fixed the pin and issue didnt go away. Figureout that issue is with the BIOS chip. Installed neobiosmasta over existing BIOS and board started to work but looks like you posted. Not sure that daughter board is neobiosmasta is fitting well.

[bobrocks95]

He Josh128!

Have you solved the issue with glitches?
I had a color pallet issue and was thinking that GRZ is dead as one pin was destroed. I fixed the pin and issue didnt go away. Figureout that issue is with the BIOS chip. Installed neobiosmasta over existing BIOS and board started to work but looks like you posted. Not sure that daughter board is neobiosmasta is fitting well.

Looks like Josh fixed his problem by reflowing the cartridge connector pins and some VRAM pins. I would try what he did and press on the board a bit while running (with something non-conductive! Pencil on the eraser end is good) and see if you can change what's on the screen.

I've also since gone over my MVS *very* closely with a multimeter to fix some other problem I was getting and there are some traces that are really hard to spot damage on! I've been through at least two MV1C boards myself.

[Josh128]

Hey guys! Its been several years since Ive worked with these Neo Geo boards and I am amazed just reading the knowledge I had when I posted about them back then. I had gotten pretty deep into the MV-1C and MV-1B during the COVID times, consolizing them to sell and doing repairs for a while, and learned a lot about them. Unfortunately, I've forgotten a lot of it now.

But yes, that last issue was fixed by reflowing every one of the multitude of cartridge connector solder joints on the bottom of the the mainboard. Solder joints and connections, bad RAM and BIOS ICs, and leaking batteries were the vast majority of the problems I ran across when working with the 1C and 1B boards. I think I still have one or two 1B and 1C boards and parts up on a shelf that I never finished. I need to get back to that one of these days....

[aknot]

Hi, I will try to solder cartridge pins. I already reflowed almost everything, except cartridge pins:)
BTW, this is how I fixed broken (rust) pin on GRZ.

He Josh128!

Have you solved the issue with glitches?
I had a color pallet issue and was thinking that GRZ is dead as one pin was destroed. I fixed the pin and issue didnt go away. Figureout that issue is with the BIOS chip. Installed neobiosmasta over existing BIOS and board started to work but looks like you posted. Not sure that daughter board is neobiosmasta is fitting well.

Looks like Josh fixed his problem by reflowing the cartridge connector pins and some VRAM pins. I would try what he did and press on the board a bit while running (with something non-conductive! Pencil on the eraser end is good) and see if you can change what's on the screen.

I've also since gone over my MVS *very* closely with a multimeter to fix some other problem I was getting and there are some traces that are really hard to spot damage on! I've been through at least two MV1C boards myself.