Yes! Yes I would like to buy a MV4-FT2 for $20 - that sounds like a perfectly sensible idea.

[hatmoose]

"Untested - assume broken" they said and they were not kidding.
In this thread I shall spend hundreds of hours and hundreds of dollars trying to get this thing working before giving up and buying a new one

You dont get much packaging for twenty bucks, but other than the smashed up legs it arrived reasonably undamaged


I guess it's hard to do any testing when the slots are full of leaves, and bugs...


Off to a strong start - at least it does something...

Stock BIOS complains of bad backup ram, universe BIOS agrees about the bad backup ram


Time to dust off the test bios and z80 test cart I made by following these instructions http://smkdan.eludevisibility.org/neo/diag/ and see what's going on

 

[hatmoose]

The board was in much better physical shape that I could have possibly hoped. Not sure if someone has been in here before me or not. It was only moderately foul

On closer inspection this turned out to be mouse piss - preferable to pretty much anything else...


There is a bug in the code...


And the battery does not seem to have leaked. My very good friends at buyicnow have produced some test bioses and roms for a z80 test cart for me - will see what they have to say


Overall - off to a less disastrous start than I dared hope

 

[stj]

"On closer inspection this turned out to be mouse piss - preferable to pretty much anything else..."

let me school you on that.
mouse piss is highly concentrated acid, one of the most corrosive agents i have ever seen.
seen the film "alien" ?? - well mouse piss is worse!

i have seen mouse piss eat right through heavy chrome plating to expose the steel base!!!!!

 

[stj]

as for repair - it looks very good.
i would remove the battery and then wash the boards in the sink with dish washing liquid and boiling water.

after drying them, inspect the cartridge ports - i already see broken edge connectors that need replacing - dont test it till you know that none of the bent pins are shorting to anything else.
and replace the bios socket with a turned-pin type.

 

[hatmoose]

i have seen mouse piss eat right through heavy chrome plating to expose the steel base!!!!!

Thanks for the heads up! I havent done any cleaning yet but it sounds like I should probably do it right away.

In checking if mouse urine was acid or base and so how to neutralise I came across this little gem. Somewhere on earth someone is working right now to study how the acidity of mouse urine is changed over time by the ingestion of bicarbonate of soda
https://www.researchgate.net/figure...icarbonate-Urine-from-mice-was_fig2_255738009

 

[stj]

it doesnt matter - your going to wash it off with soap anyway

 

[hatmoose]

It was all going so well

With the Unibios I was able to force AES mode to skip the backup ram error


with a bit of pushing and pulling and reseating and cleaning I was able to get it to boot a game - some graphics glitches, but those are normally just the cart slot needs cleaning


And after hitting the bottom board with the hair drier to warm it up I was even able to clear the backup ram error.
That sweet, sweet "all tests passed" from the test bios.


And then I finished for the night and left it in the unheated garage overnight, which is where things started to go wrong.
Stay tuned for the next chapter - in which I should have taken @stj s advice when I had the chance

 

[hatmoose]

And then It was not going so well

After I left this in the unheated garage overnight I was greated with this the next day - it's winter in NZ at the moment, not sub zero outside, about 7 degrees c

Garbage graphics and what looks like no sync(?)

By removing the top board and pressing on some of the big video chips I was able to get a stable picture

And with some more pressing and flexing of the board, different colors

My test equipment is a bit limited - I do have an oscilloscope and a multimeter, I dont have a logic probe yet (on order now).
I'm really not sure what I'm looking at here.

Are these glitch patterns common? do they normally point to a fault in a particular area that I shoud start checking?
Thoughts and guidance would be gratefully received

 

[stj]

so i'll start with what i do with every board i get - that will be most helpfull to people

first, if it's not clean - really clean, i pull everything in a socket out and wash the board - so i can see it properly.
either way, next is visual inspection.
looking on top for scratches through tracks, bent/damaged/missing components and for signs of liquid around capacitors etc.(like every fucking namco sys11 i ever recieved!!!)
look on the underside for bent pins touching each other or smashed down flat against other tracks.

next pull everything socketed out and check for bent pins and damaged sockets - including insects living under sockets!!!!!!
next i dump all the roms incase it's a set not in mame yet - sometimes you find a bad rom doing this - which saves some time.
and test any ram that got removed while i have the programmer set up - testing sram in programmers is not perfect but catches 99.9% of bad ones.
reasemble the now clean looking board.

AND NOW YOUR READY TO PUT POWER INTO IT.


if it wont start then scope or probe the clock and reset pins on the cpu('s)
most reset problems are caused by a dried up startup capacitor.
a small - usually under 100uf cap connected in series with a resistor across the power that generates a pulse when you power the board up.

 

[ic3b4ll]

Cold solder joints are an issue that arises frequently on boards that are stored on less than ideal conditions.
If the solder in the SMD chips is dark grey it's likely you'll have contact issues between chip pins and their pads. The solder has oxidized superficially and likely also beneath the pins. It's a really insidious problem because they are mechanically connected but with poor electrical conductivity.
A reflow usually takes care of it. The LSPC2-A2 generates sync IIRC. Start with that one.
The fat NiMH battery may not have leaked per se, but check the traces near it. I'm starting to believe that the minute current from the battery circuit causes galvanic corrosion. Corroded traces should look much darker than normal and may be cut in points.
Bluish speckles are usually also found in boards where the battery corroded nearby traces and components.
Source: many boards fixed.

 

[stj]

battery current does attract corrosion,
i'v had boards that were clean other than the backup power track all the way to and including the pin!! of the ram chip had green corrosion on it!

 

[hatmoose]

Thankyou @stj and @ic3b4ll - really appreciate this!

I'm a bit restricted by test gear, so in order of what I have on-hand I'll probably do the following

1) Clean it up so I can see whats going o
2) remove/replace the battery and check for damage
3) visual inspect everything
4) start checking continuity with the multimeter
5) test/replace those electrolytic caps, starting with the startup capaciator
6) replace the BIOS socket with a new forged pin one
7) Start reflowing, Begin with The LSPC2-A2
start probing - this will have to wait for the logic probe to arive
9) start testing anything that probes funny on the oscilloscope

This was working until it got cold overnight, so I might also give it some time in a warm environment to see if I can coax it back to where it was

 

[stj]

cold effects capacitors and temperature change also causes expansion/contraction that can indicate broken solder joints

 

[hatmoose]

Spent a very pleasant day working on this - then got sidetracked and re-capped the whole thing - Other than the one

Anyhoo Progress as follows
Cleaning went well, combination of IPA and Acetone
Removed the battery as suggested - it had just started to leak, hidden under the battery, thanks for the heads up - I have microscoped and multimetered this area carefully and there is no trace damage that I can find

Visual inspection didnt show any damage - will keep inspecting
Warmed it up with the hair drier for a bit to see if I could coax some life out of it - sadly no
Started checking continuity with multimeter which is where I left it for the day

At this point I realised I was going to have to narrow down my search a bit rather than just randomly beeping at stuff all over the board, thats going to take weeks.

I did a bit more reading on watchdog errors and the click of death - the best explanation that I found was here
http://newlifegames.net/nlg/index.php?topic=152.30;wap2

Bad CPU - unlikley
Bad Work SRAM - possible
Bad Backup SRAM - highly likley - given that both the stock bios AND the Universe bios complained about this 2 hours before it stopped working
Missing control signals on Work or Backup SRAM - somewhat likley
Shorted 74AS245 ICs buffering program ROMs on top board to CPU on bottom - top board is off so I cant see that being a problem
Short on address bus - possible
Short on data bus - possible
Missing address or data signals to chips - possible

So bad backup ram, or the traces/control signals associated with it are my prime suspects

This thread is absolutely halarious by the way - I'll roll the HC32 into the mix as well
https://www.arcade-projects.com/threads/mv1c-backup-ram-error.13969/

So I did some digging on bad SRAM and found this about diagnosing 4-slot SRAM locations
https://www.arcadecomponents.com/repairlogs.html#84069
SRAM is expressed as UPPER UPPER LOWER LOWER
So if backup RAM written = 55 55 and BACKUP RAM read = D5 55
55 is not equal to D5 so upper is bad
55 is equal to 55 so lower is ok

Backup RAM locations are given as the following on the MV4
Upper backup RAM = D9
Lower backup RAM = F9

Unfortunately the board layout on the MV4 is not the same as the MV4FT, but I'm guessing its the one at the bottom


@xodaraP Thanks for the pointers that you gave over on the "post new stuff" thread - this is starting to look a lot like a RAM problem, exactly as you predicted

So my next steps from here will be
0) bridge J2 to disable the watchdog to see what happens - this may be unwise.
1) Probe around those RAM traces some more to see if I can find any breaks
1a) either wait for my logic probe to arrive or learn how to use my osilliscope as a logic probe
2) reflow that LSPC2 (I'm not super hopeful about this, but it's good practise)
3) order parts to replace the Backup RAM and HC32
4) potentially order parts for the Palette ram, and Fast video ram just in case, depending on how much they cost

 

[xodaraP]

Backup RAM is a very common fault on MVS boards. I’ve also had the HC32 die but mostly on MV1B boards

Fast video RAM is a very common failure. If it’s working now, I’d recommend socketing it. I did on all the 4 slots I repaired because it is a very common failure. It’s 2x 6264 SRAMs, very common part, not expensive.

 

[stj]

i just pull and test all the ram on neogeo boards.
i use a 28pin soic to dip adapter i got from china to test the smd parts real fast in my programmer.
anybody doing more than a few boards with smd ram should get one - they are only about €5-8 if i remember right.

 

[ack]

Unfortunately the board layout on the MV4 is not the same as the MV4FT, but I'm guessing its the one at the bottom

https://github.com/jwestfall69/neogeo-diag-bios/blob/master/docs/ram_locations/mv4ft.md

Looking at some of your original pics I would examine the underside of the cpu board. Almost all of the feet have been broken off, so it took a beating at some point. Look for possible scratches that could be causing a cut trace and bent/touching pins (I can see a bunch of bent pins on one of the board to board interconnects). Likewise one of the pics show bent pins in the memory card IDC connectors, verify none of those are touching.

This may sound stupid, but verify you have the bios plugged in the correct way.

 

[ic3b4ll]

Spent a very pleasant day working on this - then got sidetracked and re-capped the whole thing - Other than the one

Anyhoo Progress as follows
Cleaning went well, combination of IPA and Acetone
Removed the battery as suggested - it had just started to leak, hidden under the battery, thanks for the heads up - I have microscoped and multimetered this area carefully and there is no trace damage that I can find

Visual inspection didnt show any damage - will keep inspecting
Warmed it up with the hair drier for a bit to see if I could coax some life out of it - sadly no
Started checking continuity with multimeter which is where I left it for the day

At this point I realised I was going to have to narrow down my search a bit rather than just randomly beeping at stuff all over the board, thats going to take weeks.

I did a bit more reading on watchdog errors and the click of death - the best explanation that I found was here
http://newlifegames.net/nlg/index.php?topic=152.30;wap2

Bad CPU - unlikley
Bad Work SRAM - possible
Bad Backup SRAM - highly likley - given that both the stock bios AND the Universe bios complained about this 2 hours before it stopped working
Missing control signals on Work or Backup SRAM - somewhat likley
Shorted 74AS245 ICs buffering program ROMs on top board to CPU on bottom - top board is off so I cant see that being a problem
Short on address bus - possible
Short on data bus - possible
Missing address or data signals to chips - possible

So bad backup ram, or the traces/control signals associated with it are my prime suspects

This thread is absolutely halarious by the way - I'll roll the HC32 into the mix as well
https://www.arcade-projects.com/threads/mv1c-backup-ram-error.13969/

So I did some digging on bad SRAM and found this about diagnosing 4-slot SRAM locations
https://www.arcadecomponents.com/repairlogs.html#84069
SRAM is expressed as UPPER UPPER LOWER LOWER
So if backup RAM written = 55 55 and BACKUP RAM read = D5 55
55 is not equal to D5 so upper is bad
55 is equal to 55 so lower is ok

Backup RAM locations are given as the following on the MV4
Upper backup RAM = D9
Lower backup RAM = F9

Unfortunately the board layout on the MV4 is not the same as the MV4FT, but I'm guessing its the one at the bottom


@xodaraP Thanks for the pointers that you gave over on the "post new stuff" thread - this is starting to look a lot like a RAM problem, exactly as you predicted

So my next steps from here will be
0) bridge J2 to disable the watchdog to see what happens - this may be unwise.
1) Probe around those RAM traces some more to see if I can find any breaks
1a) either wait for my logic probe to arrive or learn how to use my osilliscope as a logic probe
2) reflow that LSPC2 (I'm not super hopeful about this, but it's good practise)
3) order parts to replace the Backup RAM and HC32
4) potentially order parts for the Palette ram, and Fast video ram just in case, depending on how much they cost

I've been thinking about the fault happening literally over night and the board giving a wobbly picture. The issue may be with Neo-D0 which outputs all clocks to the system. Try touching its pins or pressing it hard and see if it has any effect on the wobbliness. If it does, this would suggest a cold solder joint on one of its pins. Reflow it.
A bad clock signal would also explain why the system doesn't boot at all. The CPU doesn't like that...
Work and backup RAM and the 74HC32 that connects to them is also a common point of failure so I'd check them next.
The 245 buffers that pass the signals between the boards can stop the board from booting. They are connected to the CPU's address and data bus. When they fail they can cause the signals from the ROM and RAM to be corrupted. This is easy to check with a logic probe.
The oscilloscope is a better logic probe. You check if the signals high and low voltages are within the normal range for TTL or HC logic and the pulses transition cleanly from one level to the other.
Good luck!

 

[hatmoose]

Well, I spent several hours on this tonight randomly trying “stuff” but because I didn’t have a plan I did not accomplish very much.

Still watchdoging, and unstable sync

Very occasionally I could get a stable picture out of it. Normally those stable pictures complained of the same backup ram error I had before.

So my new plan is as follows
1) replace bios socket with new one with forger pins, confirm that bios is in round the right way and etc - complete and confirmed, every now and then it will boot to the test bios for a moment or two before the watchdog reboots it

2) reflow LSPC2-A2 and see if i can get the sync stable

3) reflow the Neo D0 a d see if I can get stable clock

4) use the microscope and multimeter to inspect for broken traces

5) learn how to use the logic probe and start probing

6) order new backup ram chips, and fast video and etc just in case

Will take a more disciplined approach tomorrow and update.

 

[stj]

can your programmer test sram?