Mvs1 graphic issues - any tips/tricks?

[Tengugurl]

Hi!
I am on a neo geo fix athon and I am onto my next fix which is an mvs1 (old school type) and I am needing a bit of guidance before I hop into things.

TL;DR graphic issues, even with diag bios still missing image/ most text.

From what I have read/researched this evening, could be a few things.

1) dirty or corroded traces (cleaned the pcb already and inspected, seemed in great shape)

2) ram issue (vram maybe? Thinking it might be something in this area)

3) leaked battery (battery is clean and I’ll be doing a battery mod tomorrow)

Hopefully the pic attached will help in providing an area to focus on but, I am probably going to look into ram to see if I can get a pic first then see what the diag bios show if and when I get a pic.

Love your input/ideas anyone who replies

Thanks!

Ps with unibios in the background color changes to black so I have colors but graphics = missing

 

[hatmoose]

I'm sure that one of the more experenced folks will comment shortly, but I've just been through a number of similar repairs.

First thing is to figure out exactly what hardware it is there is a big list here
https://wiki.neogeodev.org/index.php?title=MVS_board_types

What kind of MVS1 is it?

 

[maki]

I am on a neo geo fix athon

I noticed that you're on fire recently, you go!

I'd check each signal that its present at the slot, easy to do, just turn it on and you can test them rather fast, easy to access.

If the diag bios can't find problems its often connections to the port that are broken, or the connections between the big ICs.
RAM problems should be discovered by the Diag bios (for testing the Z80 you'd need an extra cart, but its not a Z80 problem at the moment).

 

[BIG BEAR]

Check your zmc2
BB

 

[ack]

The screen shot appears to be an original bios trying to say

VIDEO RAM ERROR

ADDRESS    WRITE  READ
0???0????   ??5?  ?7??

You will want to look at the 32k vram chips. I would start with the lower 32k vram chip as it contains the tile number for displaying s1/sfix data (text on that screen).

 

[Tengugurl]

@hatmoose its an MV1-1 https://github.com/jwestfall69/neogeo-diag-bios/blob/master/docs/ram_locations/mv1.md

@BIG BEAR awesome! I will inspect that today thanks!

@maki thanks haha, I am so fortunate to have the support of everyone on this forum <3
I love my Neo Geos and I got a bunch of parts consoles and I aspire to have almost every revision PCB (I am missing a few)

@ack ahhhh! Talk about having super skills, I am so glad someone could somewhat distinguish what it was trying to say.
I posted the pic thinking someone would be able to see some of the text and say it might be X

nice! You all are amazing! Going to do some major inspection today and check out:
1) VRAM ICs for sure
2) ZMC2


Yay! More to follow

 

[BIG BEAR]

The C0 may most likely be the culprit instead of zmc2 so check C0 first.
BB

 

[Tengugurl]

Nice, got it to boot to the diagnostic bios today.
It was just a jumble of crap before, how I got this far was reflowed the lspc as I noticed there was some slight bridging from corrosion. (Sooooo small)

Reflowed the vram 5814 chips (both) socketed those bad boys as it’s only wise to do that when possible.

After those both were replaced I got something on the screen. Yay!

Getting the vram data 0000 error now
Replaced the upper and lower vram and still an issue persists
Latest is the vram 0004 error
Thanks

 

[Tengugurl]

So I am perplexed now… replaced upper and lower vram and I am still getting a vram error. Install went nice

Do you think it’s the upper 2k? that’s what I read here:

neogeo-diag-bios/docs/ram.md at master · jwestfall69/neogeo-diag-bios

Disassembly of smkdan's Neo Geo Diagnostics BIOS with added features - jwestfall69/neogeo-diag-bios

github.com

Didn’t see any Bad traces and it’s a new cx….35-L ic so not sure?

 

[maki]

So I am perplexed now… replaced upper and lower vram and I am still getting a vram error.

Bummer, did the address change? If so, its usually the IC, if the address hasn't changed, it wasn't the IC.

Apart from rotten traces (only visual inspection isn't reliable IME, had plenty of great looking traces without continuity), quite common the situation was that the continuity between the pad and the trace was gone, but I had to measure it.

https://wiki.neogeodev.org/images/7/74/Mv1fs-page2.jpg

The way I see it, the VRAM is directly connected to the LSPC2, try checking continuity between the actual pins on the LSPC2 and the actuals on the RAM ICs, this way you can catch the "pad and trace aren't connected" errors.

 

[Tengugurl]

Yeah the address changed from 06 to 04
Okay I’ll trace it out.
Looking at the link.

I hope I can find an easy mapping that is something like

Vram pin 1, LSPC2 (20)
Vram pin 2, LSPC (21)……..

I know that the bios chip has such a mapping (used it before) but I’ll have a look at the link @maki ++++++ thank you

Oh should I check the 2k or 32k vrams? Aka the long cx…35L socketed ones or the smaller surface mount ones?

 

[maki]

Yeah the address changed from 06 to 04

Address is the same: 8000h, the value changed
from the wiki link:

Addresses 0x008000 to 0x0087FF are the 2K VRAM

IMO the smaller 2K are the ones to focus on.

 

[ack]

Thought I replied to this thread yesterday but guess I forgot to hit the post button.

The error is pointing to the lower 2k ram chip. Since you already have the 2k chips in sockets I would swap the chips between them. If the "actual" goes from 0060 to 6000 then it points to the issue being the ram chip. If the "actual" remains the same it points to some issue between your socket and the LSPC.

Also if you use the updated bios from https://github.com/jwestfall69/neogeo-diag-bios it will tell you the exact chip that is bad. In your case the error would have been "VRAM 2K DATA (LOWER)"

 

[Tengugurl]

So lesson learned today…. Don’t tackle things aren’t needed.


Exhibit 1:
Me: “oh I guess I could reflow the lspc real quick “
*air soldering buzz buzz

Me 5 min later: *ffffff it’s now booting in windings rather than normal text
Haha

Me 30 min later: reflowed and microscope inspecting for errors - fixed but what a pain

Haha
Now the issue is showing

Same error but address is 4000 rather than 0040