Neo Geo Darwin Award runner up: Unibios fail and hopeful restore...

Tengugurl

Cheng's Errand Boy
Joined
Feb 9, 2017
Posts
111
Hey there Neo Geo gurus!

I've got an AES that had a failed unibios install. As Gadgetuk would say, this poor board got attacked lol ps love your vids!

Here is my worklog, let me know what you think could be the problem as I am unsure where to go next.

Background:
AES rev 3-6 JP - uses- POW PRO3 psu
Failed unibios. Working board prior to failed install.

Damage report:
Damaged traces near the bios socket (underneath)
2 damaged pads on bios holes
Wavy PCB (bent) as an air desoldering gun was used
Bent stock bios (prying and heatgun)


Worklog:
1) Inspected board
A. 4 traces damaged under bios from pry marks
Solution: Jump wires used and full continuity

B. Noticed lots of little scratches from prying under bios
Solution: Checked continuity and all passed

C. Damaged solder pads on bios
Solution: attempted to trace fix the board (aka check
continuity to all chips i.e 68k, neo DO, work rams...)
Continuity =passed after having to do 3 major jump
wires from 68k and E0
Installed sip sockets to drop in some diagnostic bios

2) Verified work
A. Turned on the console and still getting click of death

Odd things I am noticing:
1. Clicks are sporadic (sometimes) = trace cut?
2. With a game in no clicks (sometimes) =garbled screen
3. Fast clicks =something is missing major (68k etc..)
4. Screen shows blue (aka passes hardware test) but has wavy Yellow and magenta colors on the top part of the screen

***YouTube video if you want to see first hand what it's doing***


https://youtu.be/ThitMq_-8tM

Take away:
Continuity tests passed, bios fully installed and all traces should be good?
Still no luck!

I hope someone out there has some ideas, I am lost!
Looking forward to anyone's assistance.

Let's raise this AES from the dead.
It worked before
 
Last edited:

Neo Alec

Ned's Ninja Academy Dropout
20 Year Member
Joined
Dec 7, 2000
Posts
11,927
Check the last few pins on the bios -- which are connected and whether they are bridged or not. I've read differing advice on this and it did make a difference for me.
 

Tengugurl

Cheng's Errand Boy
Joined
Feb 9, 2017
Posts
111
@Neo Alec
I checked and I didn't have any of the pins bridged. Which pins did you bridge? (I am assuming pins: 38 - 40 as they are 5v on this board?)
 
Last edited:

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
It's really hard to know what might be the exact cause, but it's most likely a broken trace / pad connection related to the BIOS there, or perhaps a fix wire is going to the wrong place (that's an easy mistake to make when connecting to some of those tiny vias for sure). Double check that all address lines connect to the CPU from the BIOS, and do the same with the databus connections. Check OE and CS finally. If all those things are good and none are accidentally joined to each other (check pins next to each other on continuity), then it must be something around that area - perhaps another trace not related to the BIOS but passing by that area. Maybe something going to the E0.

What you seem to have experienced there is too much heat - did you use a temp controlled SMD rework station or just a cheap hot air gun? I've seen boards melted and warped when a cheap hot air gun is used - problem with them is there are 2 settings usually, 300 degrees C (sorry I know you use F), or 500 degress C. The lower temperature can struggle to melt some solder points, so often people switch up to the higher temperature which is way way way too hot. A PCB will often delaminate with that much temperature. Next time I would cut off the old BIOS - or piggy back the new socket on top of it and just cut its CS and feed it 5v via a resistor to disable the old BIOS. It can be messy but far easier to deal with! =D

The other thing to consider is if any pads are missing on the bottom side, solder wont flow to the top side. If the ROM socket has full connectivity to the CPU and no other traces are affecting clocks and stuff nearby, then it should boot the SMK diagnostic BIOS. If you cannot boot the diag BIOS that indicates there is still a missing trace around the BIOS somewhere. Even if the CPU cannot see the RAM it should still be able to boot the diagnostic BIOS.
 
Last edited:

Tengugurl

Cheng's Errand Boy
Joined
Feb 9, 2017
Posts
111
@ GadgetUK

Hey! thanks for helping :)

I will double check my jumper wires and traces. (You are so right @ easy mistake to attach to the wrong area)

I used an SMD Hot air gun - Smartxchoices 858D Hot Air Rework Station Digital SMD Soldering Station Kit Desoldering Welder W/Accessories (858D)
I hate that dang thing!
It wouldn't get hot enough to loosen the solder so I did turn it up as you had mentioned (Carefully though - yet I could see the board shape changing eeeeek!)

So I bought a Hakko De-soldering Iron to remove the stock bios but I think the damage was done.

I'll send some pics of the fixes I had to do and the board damage. (Hopefully it is salvageable)

The pads on the bottom were fine (Bios socket) but 2 pads were damaged on the top of the board.

Hmmm...
Lots of stuff to check as I think those pry marks might be impacting the E0 or something else? I checked continuity yet you never know.

I'll reply back soon.
Thanks for your help you 2! :)
 
Last edited:

Neo Alec

Ned's Ninja Academy Dropout
20 Year Member
Joined
Dec 7, 2000
Posts
11,927
I always advise to leave bios replacement to the professionals at this point. I've already failed and had to fix it on my consoles multiple times.

I have a professional desoldering station now and my wife is super good at cleanly desoldering with it, so that's the only reason I would ever doing it again.

I'm considering getting a new AES just for the NeoSD and keeping it stock because the cart has the Unibios built in.
 

Tengugurl

Cheng's Errand Boy
Joined
Feb 9, 2017
Posts
111
I am rechecking all the connections once more.:)
I bet it's the E0 itself as it's close to the bios area and could have been impacted

Thanks for the link to the schematics, more stuff to check :)

One thing I am noticing weird on my board which may just be me being a noob. I am checking for continuity on the 5v pad (aka pad 1) and I notice that I sometimes get a beep (using a fluke multimeter) from the ground for one beep then it stops.

Like it can't make up it's mind pin 2 is GND or 5v?
Do you think my board is shorted or there is something causing the 5v and the GND to touch? (As mentioned it's only a split second)

Just surprised me as I would not expect to hear anything unless they are some how tied together, you know?

Let me know your thoughts :)

Thanks!
 
Joined
Aug 30, 2001
Posts
349
You could consider using single, exposed pins for the socket. See this post.

Having them exposed on the top surface helps in checking of traces on it.
 

Tengugurl

Cheng's Errand Boy
Joined
Feb 9, 2017
Posts
111
Got the diagnostic BIOS screen to show now.
Learned how to convert the read/write diagnostic code from hexidecimal to binary to find if the WRAM issue is on an upper or lower RAM

WRAM issue: (Persistent)
WRAM_opt.jpg

This seems to be a 4th data line on the the lower RAM. Now I need to find a schematic (learn how to read one of those) to see which bit/ trace I need to look at as it seems like it's a bad trace giving me a bad address line.

Does anyone know how to find which bit or # leg on the lower RAM this is? And how one finds this out?

Got this far, now to figure out how to take my knowledge from reading the diagnostic code to a physical location.

Side issues I noticed:

BIOS ERROR: (Not persistent but showed up on first boot)
BIOS2_opt.jpg

WRAM A0-A7: (Not persistent but showed up on first boot)
A0_opt.jpg

WRAM FFFF: (Not persistent but showed up on first boot)
FFF_opt.jpg

This is how I learned how to convert my diag results to binary
https://wiki.neogeodev.org/index.php?title=Repair_help. (used a calculator online so nice!)


Kind of stuck unfortunately, if someone could give me an example on how to pinpoint which pad /WRAM leg to trace that would be so helpful. :)


Thanks!
Tengu
 

Attachments

  • Watchdog_opt.jpg
    Watchdog_opt.jpg
    23.1 KB · Views: 72
  • Pinouts-min_opt.jpg
    Pinouts-min_opt.jpg
    38.8 KB · Views: 58
Last edited:

Xian Xi

JammaNationX,
15 Year Member
Joined
Dec 1, 2005
Posts
27,748
Download the pinout for the ram, each ram has 8 data lines aka bits, depending on the data sheet, they usually start at D0, D1, D2 etc, some brands start at D1 and don’t use D0 as a tag. So if it’s the 4th bit, look at D3.
 

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
Looks like bad WRAM to me! Just based on the fact the addresses (and data bits) change between tests there. Unless you've got an intermittant connection on one of the data or address bits.

Just look at the 62256 pinout and the 27C1024 pinout (for the ROM) and check that D0 to D7 connect to one of the 62256, and D8 to D15 connect to the other.
 
Top