My Timeharvest CMVS seems fried for a second time

bobrocks95

n00b
Joined
Dec 16, 2021
Posts
18
So I figure best is short story first, long story second. This is a Timeharvest consolized MV-1C board.

My CMVS crashed while playing Metal Slug, and now gives me pink/cyan/neon green garbage on boot, cart or not. This is the type of screen I was getting:
Emailing JNX to ask about purchasing a diagnostic BIOS, he and someone on the shmups forum both told me pink garbage meant it was likely a BIOS issue. I have a NeoBiosMasta VMC that I installed, and the system was working before I installed it and after I installed it.

I took off the VMC, popped out the UniBios chip from its socket, reseated it and checked continuity between the chip's pins and the socket, cleaned the CPU pins with some CRC electronics cleaner and Qtips, cleaned up whatever fibers were left, and reseated the VMC a few times, same thing.
I removed the VMC, undid the pin 10 & 12 lifts and wire on the original BIOS chip, and got different garbage each boot (a solid blue screen, solid gross shade of yellow screen, some flickering 1-pixel dots flying up the screen), but still nothing.

So I'm kind of at a loss now. I can take pictures of what I'm seeing with the original chip reconnected if that's helpful. Maybe I didn't get the pins back down properly, but there are vias they go straight to and a continuity test checked out. I'm suspicious about this still being just a BIOS issue, which takes us to the long story:


This is the second MV-1C board I've had in this Timeharvest unit. The first one started to show graphical glitches on games (sprites being like one off, d's becoming e's and all that), then rapidly got worse and worse until finally it wouldn't boot at all. The modder I sent it to said they hooked up diagnostic BIOS and told me the CPU was fried, so I bought a whole new MV-1C for it.

I thought at the time that it was because I used a 151 multicart for a whole 10 minutes on it once... I'm wary about 3.3V flash and I figured it was just my luck that others use those carts all the time, but I pop one in literally to test it before selling it and my console dies.

But now the Timeharvest may have killed another MV-1C, and that's where I get officially suspicious. Is the cheap power brick it came with supplying the proper voltage to the unit? Is the step-down converter circuit I see near the JAMMA connector getting anywhere close to 5V? I don't even know how to start testing these things, but I have a multimeter and at this point am so desperate for answers I want to find some smoking gun because no way I'm that unlucky.

Do Timeharvest units have any well-known problems I should look for?


Thank you for reading, sorry to make this my first post and for it to be so long-winded.
 

bobrocks95

n00b
Joined
Dec 16, 2021
Posts
18

BIG BEAR

SHOCKbox Developer,
20 Year Member
Joined
Dec 14, 2001
Posts
8,230
I remember repairing one of these a couple of years back and I believe it was a power issue(I don't remember if the unit was getting too much juice or not enough) I know I ended up replacing the power jack and the power brick though.
BB
 

bobrocks95

n00b
Joined
Dec 16, 2021
Posts
18
I remember repairing one of these a couple of years back and I believe it was a power issue(I don't remember if the unit was getting too much juice or not enough) I know I ended up replacing the power jack and the power brick though.
BB

I need to look up and learn how to probe voltages anyway, but should I focus on probing voltages on CPU 5V lines and such? Maybe the JAMMA edge 5V? I've never had to test a power supply before.
 

Neo Alec

Ned's Ninja Academy Dropout
20 Year Member
Joined
Dec 7, 2000
Posts
11,927
Super easy. Get a multimeter, put it in voltage mode, ṽ. Start with the power supply - put the positive probe inside the barrel, with the negative touching the outside.

Buying a better quality supply of the same voltage and polarity will probably help.
 

bobrocks95

n00b
Joined
Dec 16, 2021
Posts
18
I think the squiggly line is for AC voltage but thank you it was easier than I thought!

Getting 12.61V on the barrel jack, though it's dropped to 11.8V in-circuit. See the picture at the bottom, but given the 5V line is available right there I don't know if it's going through a step-down on the underside of the board, or just oddly passed down to what kind of looks like a step-down that's on the front I/O board and then straight back up to the back board.

Saw 5.13V on the 68HC CPU, 5.12V on the YSA2, and around 5.11V on the JAMMA pins. Oddly the -5V JAMMA pin had the same polarity as the normal 5V one as far as I could tell, and the 12V pin also had 5.11V. I don't know if the MV1C board would use the 12V line or not- it's connected to a via so it goes somewhere I guess.

I don't know how tolerance works for VCC on an IC. I would think 5V +/- 5% would be well and fine, but I'm not sure.

https://media.discordapp.net/attach...9986/20211218_020113.jpg?width=686&height=915
 

SignOfGoob

Butthurt Enthusiast
20 Year Member
Joined
Sep 18, 2003
Posts
2,857
I would also wonder if the 5V was correct. Two PCBs dying in the same setup in a short span of time isn’t great. They can live for decades in filthy arcades…but in those situations they are running legit carts, locked away in wooden boxes, and have PSUs that don’t suck.

I’ve used little wall adaptors like that but I’m trying to think of I’ve ever run any JAMMA system on a SG that did the 5V conversion. I’m pretty sure I’ve always already had it in the PSU. It’s been a while though…
 

SignOfGoob

Butthurt Enthusiast
20 Year Member
Joined
Sep 18, 2003
Posts
2,857
I think the squiggly line is for AC voltage but thank you it was easier than I thought!

Getting 12.61V on the barrel jack, though it's dropped to 11.8V in-circuit. See the picture at the bottom, but given the 5V line is available right there I don't know if it's going through a step-down on the underside of the board, or just oddly passed down to what kind of looks like a step-down that's on the front I/O board and then straight back up to the back board.

Saw 5.13V on the 68HC CPU, 5.12V on the YSA2, and around 5.11V on the JAMMA pins. Oddly the -5V JAMMA pin had the same polarity as the normal 5V one as far as I could tell, and the 12V pin also had 5.11V. I don't know if the MV1C board would use the 12V line or not- it's connected to a via so it goes somewhere I guess.

I don't know how tolerance works for VCC on an IC. I would think 5V +/- 5% would be well and fine, but I'm not sure.

https://media.discordapp.net/attach...9986/20211218_020113.jpg?width=686&height=915

Some arcade PCBs are super sensitive to 5V. Having only 0.13 over isn’t too much for a Neo though, I don’t think. Arcade PSUs are usually adjustable which allows you to aim for 5.0V.

The 12.6 is probably fine.

Given the lack of a evidence I would look for another PSU and maybe replace the reg on the SG PCB.
 

Neo Alec

Ned's Ninja Academy Dropout
20 Year Member
Joined
Dec 7, 2000
Posts
11,927
I would get a new PSU regardless. The voltage might appear fine, but if it's cheap it still could be providing noisy, inconsistent power.
 

bobrocks95

n00b
Joined
Dec 16, 2021
Posts
18
Fair enough, couldn't hurt to do so.

Is the +5V on the -5V and 12V JAMMA pins harmful? I don't know how the MV1C uses them.

Also should I pick up a diagnostic BIOS to see if my CPU is fried again on this unit?
 

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
Is that 5v reading with a cart inserted, if it is then the power should be fine as it would dip way down if the power supply wasn't keeping up.

So when you undid the 10 and 12 pin lift and soldered them back where they should be, the console still doesn't boot?
 

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
One thing you can try, turn on all the hard dips to clear the backup ram then power on the unit in case it's causing some issues.
 

bobrocks95

n00b
Joined
Dec 16, 2021
Posts
18
Is that 5v reading with a cart inserted, if it is then the power should be fine as it would dip way down if the power supply wasn't keeping up.

So when you undid the 10 and 12 pin lift and soldered them back where they should be, the console still doesn't boot?

Did most of my testing without a cart, but with KoF 98 in (figured a beefy large cart was best) I only retested the JAMMA edge but it only dropped ever so slightly to 5.09V or so.

Pins 10 and 12 have been soldered back down, not my prettiest work but fortunately two vias are right there on the board and continuity checks out:
The garbage was more varied each boot after doing it, here are some examples:
https://i.imgur.com/Ya9huG0.png
https://i.imgur.com/LYUyYXh.png

One thing you can try, turn on all the hard dips to clear the backup ram then power on the unit in case it's causing some issues.

Huh, didn't know that. Didn't do anything though. I should point out when I got this board it had the backup ram battery removed and what looks like some trace repair done to it. Which is why I figured my BIOS settings never saved between boots.

https://i.imgur.com/eF36F29.jpg
 

bobrocks95

n00b
Joined
Dec 16, 2021
Posts
18
Looks like I should probably go ahead and order a diagnostic BIOS after Christmas. Only other idea I have at this point is maybe checking traces between the BIOS and CPU, though that wouldn't explain why the NBM VMC also wasn't working since I'm assuming it interfaces directly with the CPU.

EDIT: Redid VMC install/BIOS disable, results are back to the stable pink and cyan at least. While inspecting the board for any possible trace damage I did find two spots near the CPU address pins that looked odd. On the top the board had a bit of copper exposed- I've never seen a trace lift like that so I'm not sure where it came from. It was very nearly shorted to a ground via nearby.

The top, with the problem spot near a via in the middle (conveniently located between two partially corroded Ground vias):
Wanted to check continuity with the bottom of the board but it's got something hard covering those pins??? Almost like the solder mask melted and bubbled up, or some sort of epoxy was applied or something, I didn't want to damage traces so I only tried to scrape with my fingernail with no luck.

Bottom:
 
Last edited:

Misos

New Challenger
Joined
Aug 17, 2018
Posts
57
Looks like I should probably go ahead and order a diagnostic BIOS after Christmas. Only other idea I have at this point is maybe checking traces between the BIOS and CPU, though that wouldn't explain why the NBM VMC also wasn't working since I'm assuming it interfaces directly with the CPU.

EDIT: Redid VMC install/BIOS disable, results are back to the stable pink and cyan at least. While inspecting the board for any possible trace damage I did find two spots near the CPU address pins that looked odd. On the top the board had a bit of copper exposed- I've never seen a trace lift like that so I'm not sure where it came from. It was very nearly shorted to a ground via nearby.

The top, with the problem spot near a via in the middle (conveniently located between two partially corroded Ground vias):
Wanted to check continuity with the bottom of the board but it's got something hard covering those pins??? Almost like the solder mask melted and bubbled up, or some sort of epoxy was applied or something, I didn't want to damage traces so I only tried to scrape with my fingernail with no luck.

Bottom:
Previous trace repair, i think that is just nail polish, probably to stop corrosion.
Corrosion could have been spreading trough the board and affecting the traces and causing the watchdog.
Traces are specially delicate in those MV1Cs.
 

bobrocks95

n00b
Joined
Dec 16, 2021
Posts
18
When you have the VMC installed, check continuity between the Unibios and stock bios pins.

How many pins are supposed to have direct continuity???

Going by the original BIOS pins, I only saw pins 6, 7, 8, 9, 16, 17, 18, and 19 on the JAMMA side and pin 32 on the CPU side as having direct continuity with any UniBIOS pins.
 

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
Did you clean the pins on the CPU with a scratch pen or 1500 grit sandpaper? That sounds like your issue if you only have that many connections.
 

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
If you use the sandpaper just lightly sand it to get the gunk off the pins not take off any plating. The scratch pen would work best.
 

bobrocks95

n00b
Joined
Dec 16, 2021
Posts
18
Okay that one's my bad, snapping the VMC back on to check continuity I must have not had it very firmly down on one side. I did at least learn some info about the CPU <-> BIOS connection.

Looking at the CPU pinout (https://wiki.neogeodev.org/index.php?title=68k), it looks like the 16 D/Data lines, VCC, and GND are connected to the BIOS directly. Connection is good between original BIOS chip and the UniBIOS chip on pins 13, 15, 17, 19-30, and a couple others I missed marking down. I didn't map exactly but I know pretty much one whole side of the UniBIOS chip maps to the D lines on the CPU, and they all seemed to be accounted for.

EDIT: Triple-checked and D0-D15 are all good.
 
Last edited:

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
It should be almost 100% except for /CE and /OE and the 4 extra pins on the left side of the bios, first two and last two afaik.
 

bobrocks95

n00b
Joined
Dec 16, 2021
Posts
18
Wow alright I don't know how I can be that bad at checking continuity. Doing things actually methodically all but two pins on the UniBIOS connect to the original BIOS. I would go ahead and use a fiberglass pen on the CPU if I had one, but I've already cleaned it with alcohol and don't want to bust out sandpaper on an old chip unless I have to.

Does that finally rule out CPU connection to the UniBIOS?
 
Last edited:

maki

Edo Express Delivery Guy
Joined
Jan 1, 2022
Posts
332
Does that finally rule out CPU connection to the UniBIOS?
I'd say so.

The pics show that there is corrosion spreading, very common for Neo Geos :(

Look for specs/dots on traces and check their continuity.

I'm assuming you're getting the same problems with and without a cart, this means its not booting and a diag bios won't help IME
 

bobrocks95

n00b
Joined
Dec 16, 2021
Posts
18
I'm back at poking around with this again to see exactly where things went wrong. MV-1C boards are a bit over $100 these days so even if I got a new one it isn't touching this timeharvest circuitry unless I find a smoking gun. I hold out some small shred of hope that I can fix this one instead.

A couple notes/corrections that I have after probing for a bit:

1) The board has a solder bridge underneath the fuse. So I guess the fuse was blown and rather than replace it, just scrape some solder mask and short circuit the other side straight to the 5V line... Not like a fuse is there for a reason or anything, right??? God that's stupid.

2) I tried another power supply and got the same results, though it was another cheap Chinese one and was actually 15V. 5V was still just a bit over so the step-down circuitry was still doing its job, but I didn't check 12V line. Not a great idea to keep this running for very long of course.

3) The JAMMA edge connector on the Timeharvest is labeled backwards somehow, so I was incorrect when I said earlier that the 12V line was getting 5V. Probing on the MVS connector and referencing its properly-labeled JAMMA pinout, I actually read about 11.6V or so as I recall. I forgot to check polarity on all 5V lines but it looks like the MVS is expecting +5V on all the 5V lines and getting it.

4) 12V is on pin F in that picture and goes into the 2 black components on the right, D1 and Q2. Not sure what those are doing with the voltage exactly. One leg of Q2 goes to a via with "125V" labelled next to it which eventually feeds into pin 87 on the NEO-DCR-T chip but I didn't read any voltage when the unit was powered on and that's not a VCC pin. Just curious what 12V was used for.

5) Is this DCR-T chip used during the boot process in any way that could be affecting things?

Sorry for word vomit and thanks for continued help so far.
 

Attachments

  • 20220316_035908.jpg
    20220316_035908.jpg
    9.1 MB · Views: 6
Top