Sammy Atomiswave SD Restoration

[mourix]

Another day, another questionable purchase: a Sammy Atomiswave SD that spent a decade in a very damp shed.

...But it looks pretty nice right?


Hmm that is a lot of rust at the bottom.


Let's open it up.


Crap...


To be continued...

[NASguy]

**** vile. but does it run?

[sven666]

Oh man how bad is it? Clean up enough or will you need to change the metal plate?

[o1s1n]

On the positive side, It looks like most of the rust is just on that bottom panel.

I've never disassembled an AWSD, but from the bottom photo it looks like that panel just screws off?

Worst case scenario you could take the panel off, use it as a template to recreate it in a piece of wood and replace it with that.

[geotrig]

On the positive side, It looks like most of the rust is just on that bottom panel.

I've never disassembled an AWSD, but from the bottom photo it looks like that panel just screws off?

Worst case scenario you could take the panel off, use it as a template to recreate it in a piece of wood and replace it with that.

definitely doesn't screw off ,your eyes deceive you. they are spot type welds or the like i'd say. a bit of wire brush and rust treatment and it'll buff up ok !

[Pete]

Looks ok to me compared to some of the recent cabs on eBay.

[sven666]

The panels are riveted so theyre fairly easy to replace.

The sheet metal is pretty thin so going after the rust might punch some holes in it and compromise the integrity of the whole cab.. the design kinda hinges on everything being snug and tight together.

[mourix]

Thanks for the replies, guys. I've been getting to the bottom of this. First I need to get it working, but I am weighing the options on how to handle the rust.

Diagnosis time.

You know, all in it does not look that bad. Just little bits of rust around the entire thing.



Some missing stickers and bonus grime, but the CPO does not even need to be replaced really.



I actually got this machine after a tip from a discord member who went to look but did not take the machine. He got the monitor working during his visit but said it had sync issues and the neck board was loose before it worked. When I tried the cab at home, the monitor had the famous click of death with the monitor trying to start but failing. The power supply rails did seem fine but I guess I can not make any assumptions there.

Time to take it apart. Looking inside shows remarkable a remarkably clean machine, not counting the obvious rust.



Again, the PCBs do not look high hour really. I guess the coin counter at 10933 confirms that.



Some investigation found the two most probable causes for my monitor issues.
First off, the plastic of the CRT neck is loose. Secondly, it was a miracle the flyback even started once with those solder joints...



Status right now:

• Ordered electronics safe silicone glue for the CRT.
• Added some first stuff to the monitor wiki and started the PSU wiki.
• Began making cap lists/maps for the monitor/PSU.
• Am approaching some contacts for a possible complete automotive respray for this thing.

[mourix]

81 euros worth of caps from Mouser came in, so let's get start with the PSU.

The Wei-Ya P271 came with "quality" caps from the likes of Giant, TK and S. Replaced them all with good 105C Panasonic, Nichicon and Wima. Also resoldered everything while I was at it.



It came right back to life, even though it technically was not broken in the first place. I also made this wiki page with everything I learned plus a cap map.

[mourix]

So you want to recap a Wei-Ya M3129D, huh?

Let this be a first warning. This marvel of tri-sync engineering is made from mostly analog circuits, which means the sheer amount of caps is nuts.
I spent around 4 afternoons soldering and the thing is not even 100% working yet.

To begin, I made this cap guide for the project.

On your path, you will encounter hilariously broken solders joints and caps in barely reachable places.



Also inspect glue corrosion and replace toasted resistors.



When done, reflow EVERY solder joint and fix the few solder pads you'll inevitably have torn off.



And then I turned it on...Click.Click..Click...and no picture. Not a single CGA source worked.

Eventually I tried my new VGA tester and looked up:



After dialing it in, I did get an amazing VGA picture.


So it seems I have created a perfect VGA monitor here, but all my standard res boards just keep the monitor in a 1Hz relay click loop.
Can anyone confirm the monitor is always in this relay click loop in case of no signal? Or is this the problem already?

I did not replace a few caps due to availability but I am going to order them just in case.

[mourix]

Built a quick CRT-emudriver setup to confirm the monitor works on 31K en 25K, but not on 15K. Strange.

[mourix]

Developments and monitor frustrations..

Still trying to get 15KHz working on the Wei Ya monitor. Thought I had it when I found out the daughterboard relay measured 5 ohms terminal resistance. Put in a new relay but no cigar..



As for the developments. I happened to own a Japanese VA1 Dreamcast, and the Atomiswave has a dual Japanese power outlet on the inside.... Are you thinking what I'm thinking?

One Aliexpress GDEMU, battery holder and PTC fuse later:



To get it going in the cab, I am designing some some custom open source PCB's though.

First off, this great project turns any raspberry Pico into a Dreamcast controller.
Annoyingly enough it seems no one made open source PCBs for it as his license is very kind/permissive. Let's change that:



Secondly, the stereo amp for Egret 3 / Atomiswave is unobtainium at this point, so lets make a converter PCB for an Aliexpress TPA3110 module with the same power output as the original:

[mourix]

Oh my gosh, the monitor issue was a royal pain in the rear.....

I must have spent 50 hours on this troubleshooting and even ended up in a crisis as I had put my step down transformer in the wrong setting (but only blew up the transformer itself in the end, phew).
Even invested in new test gear like an LCR meter. Eventually Grant was a great help in getting it going, see here. I also may or may not have reproduced IC5 and IC9 modules for the Wei-Ya 3129 / Rodotron 666.

So what was the issue? This frigging thing:


This monitor changes the B+ voltage depending on 15k, 24k, or 31k input. Q12 was shorted, which is part of this switching. Basically, we were stuck at 31k voltage. Hence it going in protection on 15k.

You have no idea how happy I was to see my faulty Neo Geo on this monitor...



On with the restoration.

[mourix]

Removing decals and prep for reproduction

Before respraying a candy cab, it's important to make sure there are replacement stickers for everything you rip off.
In the case of the Atomiswave SD, side art and control panel art are well available, but all the small decals are hard to come by. Today is step one of changing that.

Step 1: Heat the decal with a hair dryer for about half a minute or more. You want it to be hot enough to hurt your fingers when touching the panel.



Step 2: Using a razor blade, slice through the molten glue. DO NOT PULL on the sticker as you'll stretch it..



Step 3: Flatten the stickers onto a friendly material. I used CD sleeve plastic.



Now scan it all on at least 600 DPI in full color. After all this effort you want as much usable pixel data as you can. Also look around if there already is any vector data available.

Here is my collection so far, so feel free to use this in any way possible.

[mourix]

Time to get back to the audio part. I assembled the custom TPA3110 class D amp.



It actually sounds great!...But keeping up with the theme, Mr Wei Ya is unhappy



I had put extra filtering on the class D amp but its clearly too noisy. Sound is 10/10, but all my games have snow-like distortion.

Time for me to design a class AB amp I guess..

[cools]

Try disconnecting the PSU ground on the amp: https://wiki.arcadeotaku.com/w/Taito_Stereo_Amp

[mourix]

Try disconnecting the PSU ground on the amp: https://wiki.arcadeotaku.com/w/Taito_Stereo_Amp

That is a bit of a dirty trick to reduce ground loops, with the result being that all the 2x10W of the psu would run over audio cable ground.

Unfortunately not the issue I am having as sound is fine here.