MV-1 turns on, now has the click of death

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,816
Hmm, okay. Simple enough I guess. I don't have a logic probe or oscillator, nor would I immediately know how to use one. I'm horrible at desoldering SMD ICs too, so I think I'll just pack this MV-1 away in it's box again. But I'm happy I did all this, at least I have a much better idea what's wrong with it. :) Considering how big this board is, I don't think it'd be worth anyone's time or money to buy this thing for parts. Perhaps my skill and toolset will improve enough down the road to facilitate fixing this PCB. Again, thanks for all the help everyone, you guys are awesome! :D

On a similar note, I found this post amongst Channelmaniac's repair logs. Boy, there sure is a lot that can cause the click of death. :(

channelmaniac said:
Model: MVS MV-4F
Symptom: Stuck in Watchdog (Click of Death)

Have a 4 slot board in to repair that was giving the click of death - stuck in watchdog - problem.

This can be caused by many different problems, including:

Bad CPU
Bad Work SRAM
Bad Backup SRAM
Missing control signals on Work or Backup SRAM
Shorted 74AS245 ICs buffering program ROMs on top board to CPU on bottom
Short on address bus
Short on data bus
Missing address or data signals to chips

First thing... Check for gouged traces. Found and fixed one. No effect.
Next... Substitute BIOS. Nothing.
Next... Check control signals on the Work RAM - Stuck high on pins 22 and 27.
Next... Check control signals on the Backup RAM - Stuck high on pins 22 and 27.
Next... Check control signals on the BIOS ROM - These were working.

So we had a problem somewhere in the system to where it was trying to read the BIOS but not able to initialize hardware.

Pulled the 74AS245 ICs buffering the data lines to the top board. Had no effect.

Checked for more broken traces. No more were found.

Checked the address and data lines going to the BIOS and SRAM ICs. All were good. Checked the data lines going to the NEO B1 chip. All were good. Checked the data lines. All were good.

Next checked the A22* and A23* alternate lines. Missing one one! Ran a jumper wire from Pin 55 of the NEO-E0 IC to pin 117 of the NEO-B1 IC and the board came up!

OOPS. Z80 Error. Reset line stuck high. On a Z80 since the reset line did not transition from low to high the Z80 never booted. Ran a jumper and the audio section worked. Replaced the pulled 74AS245 ICs and the board wouldnt' boot. Reset was stuck low.

Turns out the reset line for the Z80 CPU and chips on the left side of the board goes through the 74AS245 IC at E11. Removed the jumper for the reset line and board booted normally.

Plugged the top board back in, inserted some test carts and played a couple of games.
 

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
Most commonly on the 1 slot boards it's th connection between work ram to CPU to BIOS.

Just get your multimeter and make sure everything is connected. If then you have 100% connectivity then you can look at individual parts that may be bad.
 

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,816
I still don't have a logic probe or something, but I sat down and poked around that MV-1 again anyways. Again I cycled around the the main GFX chip which has corrosion on two sides of it and only found one line that wasn't making a connection, no matter how much corrosion I scraped away, so I jumpered it. It didn't seem to make a whole lot of difference, but fixed traces are always a good idea.

Quick question about that particular chip... Is corrosion conductive in any way? If not, would surface corrosion (not penetrating damage) cause limited conductivity on traces?

Also, due to the PCB's design I can't shine a light through it and check for traces that way. Are there other ways to do it? Like shine a light on the shiny trace? All the "good" traces (with bits of black crap on them) look okay when I angle a light on them...

Anyway, I check around the rest of the PCB (excluding the audio section and the area near the JAMMA edge and controller ports) and didn't find anything apparently wrong. Mostly just did visual inspection... I don't know enough about inter-connections between different ICs on the board, so trying to test between CPU/BIOS/Work RAM is pretty hard. The most I can do is just check surface traces and I didn't see anything wrong...

Oh yeah, one more thing. The bad looking traces around the big square GFX chip, had this black stuff that actually scraped off, so clearly just surface damage. Overall I'm still very puzzled over this thing. I've found maybe one bad trace but it's still got the same problem... Any other dodgey looking traces all test good. If the board's problem was one or more dead chips (like RAM), wouldn't it display an error? ... Then again, it doesn't finish posting, so even if it had one, it can't get far enough to show it... Where could the problem be?
 

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
Whenever doing board repair it's always good to have a mag glass. I have a 10x and a lighted 40x, helps when checking QFPs as well.

If you are getting the click of death it only has to do with the Work RAM, BIOS and CPU connection. Do a search on the forums or google for watchdog and you'll find a pinout to test with.
 

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,816
Whenever doing board repair it's always good to have a mag glass. I have a 10x and a lighted 40x, helps when checking QFPs as well.

Well, I have a "helping hands" soldering thing with a magnifying glass. Not sure how powerful it is, but it really makes following traces and seeing QFP pins much easier.

If you are getting the click of death it only has to do with the Work RAM, BIOS and CPU connection. Do a search on the forums or google for watchdog and you'll find a pinout to test with.

Well, earlier in the thread I removed the BIOS and pretty much nothing changed. It's probably an issue with CPU to work RAM? Would a damaged CPU also result in the click of death?

I dug around and found this thread. Does it matter if it's AES related? Aren't the BIOS/CPU/RAM connections the same?

http://www.neo-geo.com/forums/showthread.php?207382-Neo-Geo-AES-Work-Ram-Error

I hate to be a bit helpless, but I don't know how to read MKL's connectivity lists... Well, maybe I do, but I don't want to be misinformed. In the case of C2 (work RAM right?), the RAM pins are on the left and 68000 is on the right? Sounds kinda right, since the 68000 on my MVS board is DIP64.

(CPU >>> Work RAM?)

C2

1 68000-12 (43)
2 68000-12 (41)
3 68000-12 (36)
4 68000-12 (35)
5 68000-12 (34)
6 68000-12 (33)
7 68000-12 (32)
8 68000-12 (31)
9 68000-12 (30)
10 68000-12 (29)
11 68000-12 (5)
12 68000-12 (4)
13 68000-12 (3)
14 GND
15 68000-12 (2)
16 68000-12 (1)
17 68000-12 (64)
18 68000-12 (63)
19 68000-12 (62)
20 GND
21 68000-12 (39)
22 NEO-C1 (92)
23 68000-12 (40)
24 68000-12 (38)
25 68000-12 (37)
26 68000-12 (42)
27 NEO-C1 (93)
28 VCC

D2

1 68000-12 (43)
2 68000-12 (41)
3 68000-12 (36)
4 68000-12 (35)
5 68000-12 (34)
6 68000-12 (33)
7 68000-12 (32)
8 68000-12 (31)
9 68000-12 (30)
10 68000-12 (29)
11 68000-12 (61)
12 68000-12 (60)
13 68000-12 (59)
14 GND
15 68000-12 (58)
16 68000-12 (57)
17 68000-12 (56)
18 68000-12 (55)
19 68000-12 (54)
20 GND
21 68000-12 (39)
22 NEO-C1 (6)
23 68000-12 (40)
24 68000-12 (38)
25 68000-12 (37)
26 68000-12 (42)
27 NEO-C1 (7)
28 VCC



The green via is on the trace that connects NEO-B1 (86) to LSPC2 (106).

SP1

1----VCC
2----GND
3----68000-12 (54), LSPC2 (21), NEO G0 (50), 43256 (19) D2
4----68000-12 (55), LSPC2 (20), NEO G0 (49), 43256 (18) D2
5----68000-12 (56), LSPC2 (19), NEO G0 (48), 43256 (17) D2
6----68000-12 (57), LSPC2 (18), NEO G0 (47), 43256 (16) D2
7----68000-12 (58), LSPC2 (16), NEO G0 (34), 43256 (15) D2
8----68000-12 (59), LSPC2 (15), NEO G0 (33), 43256 (13) D2
9----68000-12 (60), LSPC2 (14), NEO G0 (32), 43256 (12) D2
10---68000-12 (61), LSPC2 (13), NEO G0 (31), 43256 (11) D2
11---GND
12---68000-12 (62), LSPC2 (10), NEO G0 (18), 43256 (19) C2
13---68000-12 (63), LSPC2 (9), NEO G0 (17), 43256 (18) C2
14---68000-12 (64), LSPC2 (8), NEO D0 (3), 43256 (17) C2
15---68000-12 (1), LSPC2 (7), NEO D0 (2), 43256 (16) C2
16---68000-12 (2), LSPC2 (5), NEO D0 (1), 43256 (15) C2
17---68000-12 (3), LSPC2 (4), NEO D0 (64), 43256 (13) C2
18---68000-12 (4), LSPC2 (3), NEO D0 (63), 43256 (12) C2
19---68000-12 (5), LSPC2 (2), NEO D0 (62), 43256 (11) C2
20---NEO E0 (59)
21---68000-12 (29), NEO E0 (64), 43256 (10) C2 D2, HC259 (1)
22---68000-12 (30), NEO E0 (1), 43256 (9) C2 D2, HC259 (2)
23---68000-12 (31), NEO E0 (2), 43256 (8) C2 D2, HC259 (3)
24---68000-12 (32), NEO E0 (3), 43256 (7) C2 D2, HC259 (13)
25---68000-12 (33), NEO E0 (4), 43256 (6) C2 D2
26---68000-12 (34), NEO E0 (15), 43256 (5) C2 D2
27---68000-12 (35), NEO E0 (16), 43256 (4) C2 D2
28---68000-12 (36), NEO E0 (17), 43256 (3) C2 D2
29---68000-12 (37), NEO E0 (18), 43256 (25) C2 D2
30---GND
31---68000-12 (38), NEO E0 (19), 43256 (24) C2 D2
32---68000-12 (39), NEO E0 (20), 43256 (21) C2 D2
33---68000-12 (40), NEO E0 (21), 43256 (23) C2 D2
34---68000-12 (41), NEO E0 (31), 43256 (2) C2 D2
35---68000-12 (42), NEO E0 (32), 43256 (26) C2 D2
36---68000-12 (43), NEO E0 (33), 43256 (1) C2 D2
37---68000-12 (44), NEO E0 (34)
38---VCC
39---VCC
40---VCC
 

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
C2 and D2 are the Work RAM.

1 68000-12 (43)

Means, Pin 1 of the RAM is connected to the CPU pin 43. 68000 is the name of the CPU and 12 is the speed of the CPU, it's just extra info.

If all the connections are good then check Work RAM to backup RAM.
 

smkdan

Galford's Armourer
Joined
Dec 30, 2009
Posts
452
If the board's problem was one or more dead chips (like RAM), wouldn't it display an error? ... Then again, it doesn't finish posting, so even if it had one, it can't get far enough to show it... Where could the problem be?

That's right. The original BIOS depends on some functioning/accessible RAM to get far enough to print any error text. A logic probe would help a ton in narrowing down the causes especially on something like this the list of causes is pretty long. It goes beyond just CPU/BIOS/WRAM and having one helps you rule out lots of possible causes easily.

Well, earlier in the thread I removed the BIOS and pretty much nothing changed.

If 68k can't read the BIOS for whatever reason and crashes early, it'll look the same whether you have the BIOS in or not so it doesn't rule anything out.

In the case of C2 (work RAM right?), the RAM pins are on the left and 68000 is on the right? Sounds kinda right, since the 68000 on my MVS board is DIP64.

The only system specific stuff for WRAM/BIOS are the enables which come from NEO-C1 in that pinout but the address/data connections are the same everywhere. MV-1 doesn't use a NEO-C1 but has the older PRO chipset instead so those few pins don't apply. Someone else might have the pin numbers but info on the oldest h/w is really lacking in general.
 

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,816
C2 and D2 are the Work RAM.

How do I tell between them? There's no such markings near the RAM chips on my MV-1. There's 6 of them, 28 pins each and they all look alike.

For instance, I tried testing 68k pin 1 to C2 pin 16 and it only connected on two of the six RAM chips. Is it supposed to do that?

1 68000-12 (43)

Means, Pin 1 of the RAM is connected to the CPU pin 43. 68000 is the name of the CPU and 12 is the speed of the CPU, it's just extra info.

If all the connections are good then check Work RAM to backup RAM.

Thanks for the explanation. I realized a lot of MKL's list terminology is silk screened onto the PCB, except for C2, D2 and HC259. I can't find any of those.

That's right. The original BIOS depends on some functioning/accessible RAM to get far enough to print any error text. A logic probe would help a ton in narrowing down the causes especially on something like this the list of causes is pretty long. It goes beyond just CPU/BIOS/WRAM and having one helps you rule out lots of possible causes easily.

In your experience, what all can cause the click of death? Are broken traces and bad ICs included in that list?

If 68k can't read the BIOS for whatever reason and crashes early, it'll look the same whether you have the BIOS in or not so it doesn't rule anything out.

Okay then, thanks. I'll keep trying continuity tests, though I have a couple questions.

1) Do the lists I quoted from MKL apply to the MV-1? I'm concerned some parts aren't, like the LSPC2-A2, which my board doesn't have. It has the older LSPC-A0, along with the PRO chipset you mention further below. Are their pinouts different? (the LSPCs I mean)

2) How do I deal with jumpering connections on both sides of the PCB? Like say a QFP pin to the solder side of a DIP IC?

The only system specific stuff for WRAM/BIOS are the enables which come from NEO-C1 in that pinout but the address/data connections are the same everywhere. MV-1 doesn't use a NEO-C1 but has the older PRO chipset instead so those few pins don't apply. Someone else might have the pin numbers but info on the oldest h/w is really lacking in general.

That's rather unfortunate, sure makes trying to test for IC to IC continuity harder. :(

BTW, are traces under the battery foam and QFPs ever at risk of being broken?

Also, I've gone over some of the BIOS/CPU/LSPC/RAM inter-connections. Some do connect, but I've also come across lines that either don't connect between one chip (LSPC) and anything else on the board (BIOS/CPU) and some lines that only connect between a few chips and not everything. The worst culprits seem to be the LSPC and RAM. The immediate traces surrounding these chips are fine, but the IC-to-IC testing says otherwise... I'm just wondering if it's safe to try and patch lines like this or if using a somewhat different pinout list (that of an AES) isn't a good idea. Just kinda feel like I'm doing this wrong... :(
 

smkdan

Galford's Armourer
Joined
Dec 30, 2009
Posts
452
How do I tell between them? There's no such markings near the RAM chips on my MV-1. There's 6 of them, 28 pins each and they all look alike.

For instance, I tried testing 68k pin 1 to C2 pin 16 and it only connected on two of the six RAM chips. Is it supposed to do that?

That's normal. They're all the same type but a pair of RAMs will go to LSPC and the other 2 pairs go to 68k for work and backup RAM. Ignore the LSPC ones since they can't cause this issue. If they are bad you'll find out when the system actually starts working. There's a pattern with all 4 68k RAMs that you can use for any system. Find any 62256 RAM and 68k pinouts in google image search to check against. Then check if

A1~A15 from 68k connect to A0~A14 on all work/backup RAMs
D0~D7 from 68k connects to D0~D7 on one RAM in each pair (lower byte) (this includes the 68k pin 1 you touched which is D4)
D8~D15 from 68k connect to D0~D7 on the other RAM in each pair (upper byte)

You'll get the same results as the 68k part of the pinouts posted up there. For the BIOS, the 68k parts of the list should be fine including the bits on NEO-E0 and HC259 (just below SFIX on MV-1). Don't worry about the LSPC2/NEO-D0/NEO-G0 parts. They don't cause this problem assuming they don't have shorted pins and aren't damaged somehow. The two LSPCs have different pinouts anyway.

what all can cause the click of death? Are broken traces and bad ICs included in that list?

Sure, I just meant certain chips and certain traces from them. Here's a rough summary off the top of my head

-68k RAMs and BIOS + their traces like already mentioned here
-the custom chips that output the enables for these chips (NEO-C1 on newer systems, guessing PRO-C0 here but I don't know). A bunch of 68k address lines and other signals lead into them and a bunch of /CE and /WE signals for various components come out. The NEO-C1 labelled pins in the list will lead to the PRO chips on your MV-1.
-the watchdog IC itself and all the 68k traces to it (NEO-B1 on newer systems). It's the chip that will lead to the reset lines the 68k,Z80,YM2610 etc. If it can't be written to then you'll get this problem even if the system is perfect in terms of actual gameplay.
-74HC259->NEO-E0->PRO-?? These are needed to enable the BIOS at reset or else 68k will crash instantly.
-Reset lines that lead to various components (including HC259,68k).
-there's some short on a bus somewhere that is stopping the 68k from running properly
-the 68k itself is bad but this should be extremely rare

I can't give any more specific info than that. The newer chipsets have far better documentation. Anyway unless you get lucky in diagnosing this will probably take a while to fix especially without a probe. You do have lots of info to work from with just the RAM/BIOS connections to the 68k though.
 

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,816
Back with an update. I started with the BIOS/CPU/RAM pinout list from MKL, checking off each line. I got up to BIOS pin 3 and found it wasn't connected to CPU pin 54, nor WRAM pin 19 (on either of the ICs). I traced the line and the break was probably near the BIOS itself (the pin and socket checked out fine). I soldered a jumper wire directly between those two pins and connection was restored. (earlier I made a jumper fix on the LSPC, maybe it helped, maybe it didn't). I went and powered it up and the board cycled through it's POST successfully and hit the cross hatch. I put it all back together and plugged in Metal Slug 3, plays great! No graphic or sound problems and the hardware test works fine. I'll have to get a battery and a holder for this thing!






Thanks so much for all the help guys! I feel like a million bucks right now. Feels so good to get this thing runnning! :D
 

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
Pretty simple right? Your next one will be even easier since you already know what to do. Awesome job.
 

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,816
I wasn't expecting to find the problem so soon, much less just one problem and not running into any other issues. I got really lucky with this thing. There could have been more problems down the list too. Absolutely astonishing the old battery didn't actually wreck any components or traces! I'm not sure if I'd buy another broken Neo Geo, but I certainly do know way more about the Neo now than I did even a few months ago. You guys continue to amaze me with your knowledge of the Neo Geo, I've read lots of other repair threads. Thanks again! It's very sweet having a back up MVS. :D
 

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
MKL and smkdan are like Neo Geo encyclopedias.
 

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,816
I powered that MV-1 up again it was stuck on green. I've read threads about this before so I just dug them up and went from there. I went and did voltage testing on the calander IC. It's usually between 4.88 and 4.93 volts. However, sometimes the PCB would reach the crosshatch, stay on green, or take an abnormally long time to reach the crosshatch.

I didn't see an answer to this kind of issue on those threads. Would a dying oscillator or PSU cause this? If not, what would? I was thinking of getting a fresh oscillator anyways, considering there's corrosion in that area of the PCB (and the oscillator too).

EDIT: I reflowed the solder joints on the oscillator and a nearby ceramic cap and topped them up with fresh solder. It's not staying on green for as long as it used to. I think I'll replace the oscillator anyways...
 
Last edited:

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,816
I must have the worst luck ever. I was just playing Metal Slug 3 on it, during Mission 2 the graphics started getting messed up. I tried other games, it kept getting worse, so I took it apart and cleaned all the riser card and slot connections. Tested it with Last Blade without the cart guide, was working fine, then started looking real bad, just figured it needed the cart guide. I go and put it back on (and foolishly felt like reflowing the bad looking side of the LSPC). I tested it again, now it's got a fucking video ram error! :blow_top:

Address: 000000 Read: AAAAA Write: 0AAAA

The last time I tested the PCB, the video was going apeshit with a garbling looking gray overlay and the error message was all screwed up (saying stuff like Read: 5555 etc).

:oh_no: :help:
 

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
You probably bridged or lifted pins trying to do that. If you have a mag glass double check all the pins by pushing down on them slightly. If they move then you need to reflow them.
 

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,816
Yeah I'll do that, but what about the video ram? Which one is it? Is it dead or are there disconnected traces? I recall reading depending on the read/write error, you can tell which is which.

I'll do it later though, I've got an awful headache...
 

smkdan

Galford's Armourer
Joined
Dec 30, 2009
Posts
452
That VRAM pair needs to be working to get any sort of text on screen even with the garbling so they're still live. One of the bad LSPC traces to VRAM (same pair mentioned earlier) can cause this so it might just be a bridged or bad trace like mentioned already. The "bad side" of the LSPC was probably also causing the VRAM glitching you saw when playing so that would've needed work either way.

"Address: 000000 Read: AAAAA Write: 0AAAA" <- this is the exact message you saw including the extra '0'? It doesn't fit the format or make much sense but the VRAM problem itself can make it appear like that. It's also inconsistent and pops up with different values too? If that address and data text is atleast partially right then it's most likely a bad data line from LSPC to VRAM. A pic of the bad error screen might help too.
 
Last edited:

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,816
That VRAM pair needs to be working to get any sort of text on screen even with the garbling so they're still live. One of the bad LSPC traces to VRAM (same pair mentioned earlier) can cause this so it might just be a bridged or bad trace like mentioned already. The "bad side" of the LSPC was probably also causing the VRAM glitching you saw when playing so that would've needed work either way.

Well that's a relief. I hate having to desolder SMD stuff. Do you, or anyone else here, know which SRAM chips are connected to the LSPC? Or is it as easy as just testing for continuity? I know which two are for the 68k, but not the backup or VRAM.

"Address: 000000 Read: AAAAA Write: 0AAAA" <- this is the exact message you saw including the extra '0'? It doesn't fit the format or make much sense but the VRAM problem itself can make it appear like that. It's also inconsistent and pops up with different values too? If that address and data text is atleast partially right then it's most likely a bad data line from LSPC to VRAM. A pic of the bad error screen might help too.

I didn't pay attention to how many zeros were in the Address part, but the it was all zeroes... I guess I'll take a picture then, or a video if the screen is still going apeshit. But yes, the last time I tested the numbers were inconsistent. One time I think the read/writes were all 5s.

EDIT: Here's a pic:



The screen is flickering and garbling like crazy, so there's probably cold joints in all sorts of places... :( When I feel better I'll try to resolder the LSPC (I only did one side of it) and make sure all the traces there are working.
 
Last edited:

smkdan

Galford's Armourer
Joined
Dec 30, 2009
Posts
452
Well that's a relief. I hate having to desolder SMD stuff. Do you, or anyone else here, know which SRAM chips are connected to the LSPC? Or is it as easy as just testing for continuity? I know which two are for the 68k, but not the backup or VRAM.

You can rule them out using continuity tests. All the WRAM and BRAM chips have their address inputs tied to 68k. Do a test with your meter from pin 1 of a WRAM (A14) chip to pin 1 on the all the other RAMs. The 4 WRAM and BRAM chips will have continuity. The remaining 2 chips won't. Those are the VRAMs and only connect to LSPC. If certain data traces from the LSPC are bad for whatever reason I can see how all that garbage can flicker on top of otherwise normal text. Getting 5555 instead of AAAA sometimes means that it's intermittent. If the entire VRAM passes the 5555 test, it moves onto the AAAA test and a few others. It just stops at the first fault it finds. When the trace(s) are fixed and stable that hopefully goes away.
 

Mike26

McWow,
Joined
Sep 1, 2001
Posts
389
there is some amazing info in this thread, i am going to work on my dead MV-1 board today using this thread. HMG you're so close man don't give up.
 

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,816
Got some much needed rest and went back to the board. I started with testing for bridged pins and I found a pair by themselves that were making contact together, so I fixed that up. Then I found about 4 pins all touching together (in a chain effect). That took quite a while to fix up, there was some flux surrounding quite a few pins which took some time to burn away. I used some more standalone flux to get the legs and solder pads straightened out. This time I used a little pin for applying the flux so it wouldn't get everywhere. Dropping a big blob of flux was a REALLY bad idea!

Eventually I got it all cleaned up and there weren't any more bridged pins. I powered it back on and there's no more VRAM error and the graphics look fine. It boots to green but it's stuck there again. Without the riser board it eventually hit the crosshatch. Tried it with a game and it wouldn't get there, I think the oscillator has finally quit. They are sensitive to physical shock so it wouldn't surprise me if one bump did it in. I'll buy some replacements.

Still, pretty nice to get it going... Again. Thanks a lot for the help guys! Hopefully the funky graphic errors won't be there anymore when I get this board going fully. :)
 
Last edited:

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,816
It's clicking again... :crying:

I checked my one repair wire between the BIOS and CPU, it's fine. I also checked off everything else on MKL's pinout list, all the data lines on the BIOS to whatever else it connects to. They all make connectivity, but the MV-1 is still clicking. The only odd one was pin 20. MKL's list says it goes to pin 59 of the NEO-E0 chip, but on my PCB it doesn't connect there. Instead it goes to pin 59 of the NEO-G0. Should I add a jumper between BIOS pin 20 to NEO-E0 pin 59 anyways?

Other than that, I don't know what's wrong with this MV-1... BIOS to CPU to WRAM/BRAM seem to all be good. What should I check next?
 
Last edited:

smkdan

Galford's Armourer
Joined
Dec 30, 2009
Posts
452
Instead it goes to pin 59 of the NEO-G0. Should I add a jumper between BIOS pin 20 to NEO-E0 pin 59 anyways?

No that sounds normal actually. On other systems it leads to NEO-E0 instead but the pinouts say that pins 59~61 work the same on both G0 and E0 so it's fine. The same pins on the E0 are probably unused then. That means pin 60 + 61 on G0 lead to one of the custom chips for the BIOS enables. Either PRO-B0 or PRO-C0 will be connected to them. It should be the same chip for the WRAM/BRAM enables (the ones with "NEO-C1" in the lists).

Apart from that, just by looking at the board are there any other shorted pins on LSPC or any other chip? Including ones that haven't been brought up here. I can give you starting points for various things (like the G0 BIOS pins up there) but you'd have to figure out where they're supposed to lead instead of having a convient list to go off. You might be spending a fair while longer on it with the calendar issue on top of it. Do you still want to salvage this?
 
Top