What's new

blonu

Student
Joined
Aug 1, 2018
Messages
78
Reaction score
37
Location
USA
I recently acquired my first Neo Geo MVS 1 Slot board. I have no experience with Neo Geo, but when i looked at the board, i saw that the battery has leaked and eaten up 3 components completely. The first one, i believe is an osciallator,, the second one looks like a tantalum cap and the third component only has legs remaining.

If someone can help identify these, i can try ordering them from digikey and see if the board will atleast show the cross hatch pattern.

I plan on taking the battery out along with the register that charges it, to replace with a CR 2032 battery holder and planning on socketing the 8X2 chips in socket as there seems to be considerable battery leak damage to the legs.

I traced a bunch of traces that look like were worn out, but i do get connectivity on majority on them but it seems some of them would need to be rebuilt.

Any help will be greatly appreciated regarding suitable replacement components.
 
There is also quite a bit of trace damage in that area - it may be worth keeping that one for parts and trying to find a non damaged board.
 
Thanks @avramce for the pictures. So that looks like the 22pf cap that was completely gone on mine. Is it correct that these are ceramic caps and not tantalum? Going to order the parts (20pf, 22pf and crystal) from Digi-Key and socket most of the chips. @xodaraP I’ve not got any other projects in the pipeline so it would be interesting to see if it can be brought back to life otherwise parts is an option. I like the look of this board compared to the other 1 slot variants so will see how it goes.
 
Removed a bunch of components around the battery to investigate the damage https://imgur.com/a/cGcZa1O

Next is identifying all the damaged traces and repairing them. Anyone knows whether 20pf and 22pf caps are ceramic or tantalum?
 
Minor update. I resoldered everything back and got the crystal and couple of caps back. However, when it boots, it watchdogs between blue and white screen. Occasionally, giving a corrupt cross hatch. I put the unibios 4 and this is what it shows up as https://imgur.com/a/47lCh0Y

Any suggestions on how to proceed? Anyone here who can burn a Neo Geo diagnostic rom? Not sure if using that i will be able to read the screen or it will still be corrupt graphics.
 
It's not able to display the data from the UniBIOS ROM. Most likely due to the damaged traces around that battery that's leaked. You need a multimeter and to start metering them out and replacing as required.
 
@xodaraP I've practically tested 80% of the traces on the full board. Around the battery, i've tested all traces multiple times. There were only 2, that did not show connected, which i fixed.

It seems there is a problem with the video sync circuit. Here's the video where it appears the pvm respond with a clicking sound when the sync is reached and a cross hatch is displayed. Practically lasts for just 1 second. https://imgur.com/41jhqP1

Any clue on where on the board or which chip to focus on the attention next.
 
ok, so i got the board to a state where there is just a calendar error. I believe the D4990 needs to be replaced. With Unibios i can get to the cross hatch and all the tests run fine now. I replaced all work and backup rams and 1 video ram. The palette rams were fine. No failures are reported in any of the tests with Unibios or the Diagnostic BIOS 0.19.

However, when i put a game in, i can do all CRC checks which report back fine, but i am unable to get past the mode setting screen. When i select exit, the board reboots with Unibios and back to the mode setting screen.
 

Attachments

  • MV1ModeSettingScreen.png
    MV1ModeSettingScreen.png
    25.8 KB · Views: 113
Check all DIP switches are turned off and check that the DIP bank isn't stuck on
 
yes i've verified that through multimeter as well as from the unibios hardware DIP test screens
 
Check your service button and the pin on the edge connector to see if it's jammed on
 
@xodaraP Thanks for the hint. That completely makes sense now why unibios was showing me the test screen. Checked Pin# 14 on the solder side (Service Button) and it was low. Compared it to the other buttons like up, down etc. They switched to low when pressed, so you were right, the service button was stuck low.

Traced it to the resistor pack just next to the clock chip D4990, and noticed that I had soldered the resistor pack backwards. I had originally removed all chips and components around the battery area when i was trying to access the damage on the board. I put a socket in and corrected the orientation of the resistor pack and now with Unibios, the game boots. All buttons work and there are no failures reported by the Unibios tests or the SMKDAN diagnostics 0.19.

For the Calendar error, i am waiting for the replacement D4990 chip, but i think i should test that one now as well. Maybe the chip is fine and the resistor orientation was breaking it.

@xodaraP Thanks so much for your help.
 
I don't remember if this model got the HC32 next to the D4990 but i suggest you to check that too. Mostly of mine calendar errors ended up being the HC32 with logic errors, at least for fz1 and 1b (i have repaired nearly 30 mobo, also, with this problem).
I also see Sony's 62256, if you want this to work for a long time i suggest you to change these, even if they are working at the moment.
From my experience, best to worst used on mvs:
1 Samsung 8o
2 UMC :thumbsup:
3 Sony :sleeping:
4 Sharp X(
5 Toshiba :cursing:
 
I don't remember if this model got the HC32 next to the D4990 but i suggest you to check that too. Mostly of mine calendar errors ended up being the HC32 with logic errors, at least for fz1 and 1b (i have repaired nearly 30 mobo, also, with this problem).
I also see Sony's 62256, if you want this to work for a long time i suggest you to change these, even if they are working at the moment.
From my experience, best to worst used on mvs:
1 Samsung 8o
2 UMC :thumbsup:
3 Sony :sleeping:
4 Sharp X(
5 Toshiba :cursing:
Very useful info. Thanks! :thumbup:
 
I don't remember if this model got the HC32 next to the D4990 but i suggest you to check that too. Mostly of mine calendar errors ended up being the HC32 with logic errors, at least for fz1 and 1b (i have repaired nearly 30 mobo, also, with this problem).
I also see Sony's 62256, if you want this to work for a long time i suggest you to change these, even if they are working at the moment.
From my experience, best to worst used on mvs:
1 Samsung 8o
2 UMC :thumbsup:
3 Sony :sleeping:
4 Sharp X(
5 Toshiba :cursing:
Very useful info. Thanks! :thumbup:
Ain't nothing special ;)
This also applies when you have a BRAM ERROR (unwritable). /W on both rams are connected to the HC32, thats the first thing to check when you got that message.
Don't even bother try to desolder rams first (unless they are Toshiba or Sharp, those are real SHIT).

And last applies for this is the other HC32 near the cartridge socket extender. That one ends in the video sync counterpart. If you have shitty sync or blank screen, check that one (just be sure on the third gate you got 100ohm from the sync jamma pin to the ic, pin 8.

I often never found out that D4990 is the real issue when this is happening. Also 32mhz xtal is always good, i have never changed one.

Other things to keep in mind while repairing these, a part from z80 error (mostly nearby 6116 ram failure, but also many defective z80s) is the shitty 102/1k pot, an unrealiable SOP Op Amp and the infamous 000-lo rom. This last one is a real bastard since many time you have graphic problems, at least to my end, never ended up being the VRAM (2k and 32K) or the 6264 palette rams. It was always the lo rom in sop format, very unreliable and he is capable of fuck the board very well.

For watchdog, other than working rams (mostly) it's because the 68000 have bad solder and doing so was a terrible idea for a plcc (they should have use sockets instead). 1B also suffer for bad 27c1024 in sop format, i have saved many 1b boards just by using the neobiomasta/unibios mod

Last, the NeoBuffer near the dipswitches. If you have a good toggle between low/hi transition but it keeps to stay turned on in the test menu, thats because the neobuff is bad. He can also cause visual issues sometimes (he is "capable" of doubling up the scanlines or make blocky graphics)

Obviously all of this was considered by having good connections on the board. Rusted lines or bad vias are another thing.
These are better infos, i think :)
 
Last edited:
Good to know. I have a 1B stuck in watchdog at the moment. I'll replace the work RAM and check the CPU soldering
 
Back
Top