What's new


Sep 25, 2015
Reaction score
United States / Thailand
Back again with another AES in a bad way... this one was stuck in the click of death (fast clicking/popping), then I started messing with it. :D

I immediately found a bunch of corrosion around the LSPC2; I cleaned most of it and ran a few bodge wires, and I was able to buzz everything out (most of the bodges were lines going to the 68K data bus). From there, I socketed the BIOS and installed a Diagnostic BIOS, but continued to get the same watchdog error. I shorted DOGE (ha; pin 94) on the B1 to GND to disable the watchdog, and now I get alternating forms of colorful garbage:


Here's what else I can remember doing:
  • Buzzed NEO B1 to the LSPC2 (PCK1/PCK2: Latch signals coming from LSPC2-A2) - checks out
  • Buzzed pretty much everything between the LSPC2 and 68K data bus (D0~D15: 68k data bus) - checks out
  • Buzzed the vias and lines I could see between the LSPC2 and the work (?) RAM (CXK58256 I think) - checks out
From reading around, work RAM seems like a possibility, as well as a bad 68K (oh no). I don't want to spend all day buzzing out every single thing on the board, so I'm curious if anyone has any ideas on where I might want to focus in on. I don't mind changing out the work RAM, but I'm hesitant to swap larger ICs since I don't have hot air or a desoldering gun at the moment. Eager to hear what others make of this.

UPDATE: I buzzed out everything I found on this post (post #5 to be specific): LSPC2<->G0/D0<->43256; 68K<->E0<->HC259 (or thereabouts), as well as BIOS pin 20 to E0 pin 59. Everything checks out. I also buzzed out the BIOS to the 68K, seems like everything's continuous.
Update 3: I went back to check the work RAM again, turns out I was checking the 32K VRAM at the top of the board, haha... buzzing out the Work RAM (43256; C2 & D2 on the lower left near the reset button) revealed that two pins did not appear to connect to the 68K (22 & 27) on C2. I ran two bodge wires from those pins to the vias at the end of their respective traces, and they started buzzing out on the 68K as well. So it seems as if that is resolved.

Put in a cart, powered up, still got a garbage screen. Removing DOGE from ground results in the return of the click of death.

One thing I did notice was pin 27 on D2 connects to NEO-C1 pin 7; I verified it with the meter, but neither pin 27 on D2 nor pin 7 on the NEO-C1 seem to connect to the 68K. C1 pin 7 is identified as "WWU" on the neodev wiki, but I can't figure out where else it goes from looking at the other pinouts.

I also buzzed out halt, reset & clock between the 68K and all the customs, everything seems to be continuous. I do have a diagnostic BIOS installed, but I can't get it to run at the moment since the system is stuck in watchdog. :(

Kind of hitting a wall here. Apologies for the SOS to @leonk & @ack, but would either of you have any ideas on what I might try next?

C1 pin7 does only go to RAM. the other command signals also ony go to Work RAM
WWU/L WRU/L are [W]ork RAM [W]rite enable [ U] pper / [L]ower and [W]ork RAM [R]ead (aka OE [O]utput [E]nable) enable [ U]pper / [L]ower
So they command RAM on when data is written and read
IMO it's never a good idea to troubleshot with DOGE on, you have to cure the watchdog first.
At the very least a logic probe is needed to see how address, data and command lines are behaving, a stuck line will most likely throw a watchdog
Unfortunatelly NEO-C1 is known to be a common failure point (at least in my experience), and it commands work (backup too on mvs) RAM :/

Oh and on AES you need to run the DiagBios with a cart in to get some Fix gfx to display (MVS have SFIX on board). It might manage to tell you you're stuck on watchdog... Which at this point is not very helpful lol
Last edited:
Thanks for the check on the C1 thing! Makes sense why I couldn't find it anywhere else. :) I'll keep C1 on the radar and look at it more closely. If it ends up being bad, I'm guessing I will need to buy a donor board to get another one? Which MVS boards have the C1?

The reason I'm trying to run the diagnostic with DOGE is because I can't seem to resolve whatever is causing it in the first place; my hope was that by disabling DOGE by grounding it that the diagnostic could run enough to give me some indication of what's wrong, but I'm not even getting that far. :( I'm aware that you need a cart to run the diagnostic, I mentioned in my other post that I had a cart inserted (but I can see how it may have been missed). Cart inserted + DOGE disabled = garbage screen in my first post.

I'll see if I can find a logic probe at a local shop and see what I can do, it's probably time I figured out how to use one.
I also had a giggle when I discovered DOGE - funny, clever, and adds to the functionality. A perfect engineering Easter-egg
IMHO you should try replacing the work RAM with fresh chips at this point. You should also tone out the connections on the cart pin connector and make sure that you don't have anything wrong there.
Swapped out the work RAM with chips I got from buyicnow; no change. These were sold as "used & tested", so it's unlikely that the work RAM is the problem. I buzzed out the pads and traces underneath and around the chips, everything is continuous.

I did pick up another 68K while I was at it, but really unexcited about the idea of desoldering that thing without a desoldering gun.

If anyone has any other ideas, do let me know!
Last edited:
I did not. Speaking honestly, I'm a little intimidated/put off by the idea, there's like 100+ things to buzz out? I might start with buzzing out the ones running to/from the CPU and the customs, then try the logic probe (it should be here sooner than later).