As the title implies, even though I stated in my previous blog post back in October of 2025 that I wasn't going to be doing any new development in the retro computer arena, that changed. In fact a lot has happened since that posting.

A month later I had a heart attack which put me in the hospital, and I walked out days later with two new stents installed in the arteries feeding the heart. I was also put on a lot of meds to prevent damage to the stents.

At first after leaving the hospital I was feeling much better than I had for at least a year prior, probably because I was finally getting good blood flow.

However a few months later while watching TV at night with my wife, I suddenly experienced a change in my vision to my right eye. It was like someone had dropped a heart shaped piece of something into that eye, thus obscuring my central vision. Later it was confirmed that I had experienced a retinal bleed likely brought on by Macular Degeneration (AMD). I had what was called "WET" AMD which is the type that can bleed out into the retina. So as can be imagined my life changed a lot.

At first it seemed like my vision issues were not going to be a big problem because my brain quickly adapted and learned to ignore all the "bad pixels". But it wasn't long until the issue progressed, and in a relatively short period of time my vision in that eye became completely obscured, and all I could see were shadowy remnants of whatever was in front of me through that bad eye. It had now become a distraction. So I got some new glasses fitted with a dark lens on that side thus allowing my brain to focus on my good eye. I had also tried the only treatment presently available for WET AMD, but that was both a bad experience as well as ineffective. So now I've accepted my situation and hold out hope that an alternative treatment is found.

So on the subject of getting back to retro computer development, I did embark on a new project in January of this year. Mainly my change of mind came about because I needed something interesting enough to take my mind off of my problems. That project is what I've called the 5200XEGS.

Here's a link to AtariAge where I've been discussing it and reporting on its progress.

- Michael

One thing I've always experienced in life, is that it's absolutely guaranteed to change on you. And that's exactly what has happened with my interest in developing retro hardware. I'm no longer feeling the buzz and excitement I once had, and of course something else rushed in to fill the void.

Here's my back story...
-------------------------
A little over 10 years ago, I had rediscovered the Atari 8-bit computer after stumbling upon AtariAge and was shocked to still see a lot of activity and development being done with these systems. It had been a few decades since I left the scene, with my last endeavor being to design a color video genlocking system called the Studio II. Something that I had tried to market, but after receiving only a small amount of pre-orders was soon abandoned. However I did manage to fulfill and deliver all of those pre-orders.

In reality there were two versions created, one being the Prism Studio, of which only two units got built, and was soon followed by six Studio II's getting produced. The Prism Studio marked the beginning of this project which was built into an off the shelf metal utility project box, and used a hand etched circuit board. It had a pot that allowed fading the overlayed Atari video in and out over the external video source, and had a rotary switch which picked a particular luminance bit from the Atari that could selectively disable the fader, thus staying present when all other luminance levels were faded.

Prism Studio was the prototype, with only one being sold. This quickly got replaced by the more refined Studio II version, which replaced those manual controls with memory registers that could be poked directly by the Atari.

At the time I was still working for a company that produced video camera systems for inspection of sewer pipelines to look for cracks and other faults. While I was there I learned a lot about how analog video worked, and also the techniques behind genlocking that to a computer, which I quickly adapted to the Atari 8-bit system. After all the Atari 8-bit computer system was my personal computer from 1981-1994.
  

• Here's a LINK to a discussion at AtariAge about the Studio II.

So getting back to my discovery of AtariAge in May of 2015. I was very excited to see all the interest and development in modern day upgrades for the Atari 8-bit platform, and found myself very quickly jumping into the conversations in the 8-bit forums. It was there that I caught a thread about the TransKey. Specifically someone was looking for one. Well that turned out to be my very first Atari product that I developed and sold back in 1990, and quite frankly I had many times envisioned a rebirth using modern technology. So one thing led to another and TransKey II (TK-II) was born. The idea really began to take shape, and then got fleshed out in real hardware with the first target being the Atari XEGS. This later morphed into a few other iterations, all having the ability to allow a PS/2 keyboard to be plugged in and replace or augment the stock keyboard.

As time passed I found myself making all kinds of upgrades for these retro machines beyond the TK-II, and ended up taking it all the way to the creation of three unique Atari motherboard projects based on the integration of the original 40-pin Atari VLSI chips. I even set off in an entirely new non-Atari direction with the development of a ColecoVision Game Console clone.

All of these projects can be seen on the pages of this website. And in most all cases the PCB gerbers, BOM, schematics, and firmware are freely available as downloads.
-------------------------

So even though I was quite prolific in the retro computer arena, I eventually came to the conclusion it was time for a change. This resulted in closing down any new development about five months ago. And to be truthful I'm not feeling any urge to get back into it again. Of course I'll never say never, but I seriously doubt that I'll ever pick up where I left off.

It certainly has been a fun ride!

- Michael

Its been a very good ride, but I've pretty much done all I can think of for the Atari, and or anything else I wish to do for any of the other 8-bit computing platforms. So after 10 years this month of mainly doing Atari upgrade projects in my spare time, I've made a decision to redirect all of my creativity in an entirely different direction.

And that will be concentrating more of my time to my Low Temperature Refrigeration roots, which I've been contracting out my services to another company for over 12 years now. I'll be specifically looking at a very small desktop chiller that will hopefully blow a -50 C compressed air stream over electronic components in a lab setting for test purposes, and be nearly 1/2 the size of an existing product.

This should be a lot of fun since it also incorporates almost all of my creative talents. Yes there will be a custom PCB, and yes that board will have an embedded PIC micro-controller, and there will be a 3D printed case as well.

I'm not at liberty to reveal much more about my chiller development, since I'm under an NDA.

Until next time,
- Michael

The 1200XL-1mb SRAM upgrade was just released last week, which was very much a personal project as well as the RAMBO XL-II which is also seeing an official release at this time.

The RAMBO XL-II initially had only been discussed and talked about on AtariAge, and also saw an initial release on that platform back in 2022. At the time I needed a way to increase my 1200XL's RAM up to 256K so as to run some applications with a higher memory requirement. The original ICD RAMBO board looked like a good way to do that, but was no longer available, thus the reason I cloned their 1980's era product.

A few years later I encountered some noise issues in the analog video output on that same 1200XL, which after some investigation appeared to be originating from the DRAM and it's support circuitry. Thinking that swapping the DRAM out for SRAM might be a good way to eliminate the noise issue, the 1200XL-1mb board was born. And after installing one, I'm glad to say the noise is no longer present.

So manufacturing files for both upgrades is now downloadable at the RAM tab in my website's navigation bar.

Enjoy,
Michael

After the integration of a MIDI Player to the former FujiNet only plug-in module, I decided it was finally time to release all the changes I had been working on for the last several months. This included fixed firmware for both PS/2 sub-processors, a corrected pad hole size for the CR2032 Battery Holder on the main NUCplus4 PCB, a new combo FujiNet/MIDIplay PCB, new improved Version 2 firmware for the TK-II controller on the 576NUC+, and a new updated datasheet covering many of the changes. These downloads are all provided at the NUCplus4 webpage on this site.

The image below shows a completed board assembly with the new FujiNet/MIDIplay module, as well as the Cooling Fan and Shroud that got introduced a while back. So what you see is a fully packed system with all the bells and whistles.

1. 1088K RAM
2. A8PicoCart
3. Choice of SDrive or FujiNet (selected one at a time via keyboard control)
4. R-Time 8 Battery backed Real Time Clock and Calendar
5. Dream SAM2695 MIDI Synth for MIDI playback

Items shown in RED are the original "4" aspects incorporated into the NUCplus4 name, and a standard part of the Main daughter board.

And this complements the 576NUC+'s Multi OS and Language Slots, PS/2 Keyboard input, and high quality Composite and S-Video Output with VGATE technology.

And even though there's a lot of stuff that's been added via the NUCplus4 daughter board on top of the 576NUC+ base system motherboard, it all still squeezes into a fairly small case that easily fits in the palm of your hand.

Case Dimensions:    5.38" wide x 5.25" deep x 2.25" tall
                                      (147 mm x 133 mm x 57 mm)
Assembled Weight: 17 ounces
                                      (482 grams)

In its present form, this has become my preferred Atari computer system, and as close as it comes to a daily driver for me.

I think what I like best is not only the small size, but also the incredible versatility of having multiple solid state drive possibilities as well as wireless network access. And at least for me, incorporating the A8PicoCart was the optimal way to have cartridge based software that would run at full speed, without the bulk associated with hauling around a bunch of real physical carts. After all this was first and foremost meant to be an extremely portable system that could be tossed in a bag and taken to a meet-up.

This journey first began back in 2020, and after 5 years has matured into what I consider not only one of the smallest Atari computers out there, but also an extremely powerful one. And aside from the rare and somewhat hard to get Atari VSLI chips, is practically built with off-the-shelf parts and based upon open source designs (thank you Atari community).

I think the journey that began so long ago has finally arrived at its intended destination.

- Michael

I got the final boards about a week and half ago, assembled one, and gave it a test. Everything worked as it should.

So my plan now is to build up a 2nd one and drop that off with someone I consider to be the utmost authority on the 1200XL, and who has gotta have the largest collection in the known world, and perhaps the entire universe - Bob Woolley.

Currently I'm waiting for some SRAM chips to show up from China (both the 512K and 64K chips). I'm hoping that I get a good batch, although if not and they turn out to be fakes I won't be out much in $$$.

The 512K chips are in current production and readily available from the likes of Digi-Key and Mouser, but I'm looking for a deal so I thought what the heck I'd give AliExpress a try to see if I get lucky.

The 64K SRAM is a different story, having been out of production for years. However I've bought these in the past from multiple electronics surplus electronics sellers without even one failure, so I'm hoping for the same on this go around from an eBay seller

The great thing about switching from DRAM to SRAM, was that I got rid of the last remnant of video noise producing hardware on my 1200XL, and the results were fantastic!

New Project

So as the 1200XL-1mb project gets wrapped up, I'm off on another. This time I've decided to integrate a MIDI Player aspect into the NUCplus4's FujiNet module.

Since the NUCplus4's FujiNet module already plugs into the entire SIO bus, it has audio input capabilities via the SIO's Audio Input pin. This was already taken advantage of by FujiNet with the Disk Rotate being called out by the emulated S.A.M. voice in the ESP-32's firmware. And it's a simple matter of mixing in additional audio devices through this same pin (e.g.; MIDI Synthesizer).

Now it just so happens that there is a very inexpensive encased module that incorporates the SAM2695 MIDI Synthesizer chip that goes for cheap, and is manufactured for M5STACK in China.

• M5STACK P/N: U178
• Digi-Key P/N: 2221-U178-ND
• Mouser P/N: 170-U178

For our purposes the SAM2695 carrier PCB will get extracted from the plastic case it comes in, and with a small modification be incorporated into a brand new NUCplus4 FujiNet/MIDIplay board.

The original miniature 8 ohm speaker will get dispensed with and I'll be routing the Class D amplifier's mono output directly into the SIO Audio Input pin via a resistor mixing circuit.

At first I wasn't sure that this would work and actually provide good quality audio output, but after connecting it up to my LG 55" HDTV and Sony Sound Bar via a RetroScaler-2X video converter I was soon won over by the sheer clarity of the highs and the dynamic bass that this system combination was producing.

Luckily there was still a small fragment of real estate left inside the 576NUC+/NUCplus4 system that allowed the MIDI player to be added as an extension of the original FujiNet Module PCB.

PCB TOP VIEW






At the same time, an independent 3.3V regulated supply was added for the FujiNet side of things in order to correct some drop-out issues in FujiNet's operation that had been seen with the previous version.

Using the MI-PIC chip (PIC12F1572) incorporating the MIDI-CNTRL firmware to interface with the M5STACK MIDI-Synth module, adds a full MIDIMATE compatible connection scheme, as well as the ability to play an embedded start-up MIDI sound file if so desired.

1. 31,250 Hz Synchronized MIDI-CLK tied into the SIO-CLK-IN pin.
2. SIO-MOTOR enables the MIDI device to be on the SIO bus.
3. Power-Up MIDI Intro tune is possible, same as other Mytek SIO2MIDI devices.

On item 3, currently three modes are offered in the firmware download, Silent, Atari, and MULE. I'll be experimenting down the road with the possibility of adding a more traditional computer start-up sound, to be both short in duration and subtle (with a low volume level).

New FujiNet/MIDIplay module schematic:

EDIT (2/1/2025): Until such time that the official online FujiNet firmware has been updated past v1.4, the zipped file: fujinet-ATARI-v1.4.1_CKI-TEST.zip which is available in the NUCplus4 downloads section needs to be flashed for proper MIDIMATE usage. This file is part of the total firmware download file: NUCplus4_firmware_all.zip


Well that wraps it up for this post.

- Michael

Well lot's of things happened over the past 3 months.

• The NUCplus4 Project got released
• The UGV got a new, better, revised version released called the UGV-II
• I was in the hospital for 3 days because of massive blood clots in my lungs
• The 1200XL-1MB Project (a 1088K SRAM upgrade) got started

NUCplus4
This got released on September 21st 2024 nearly 3 months ago, and there have been several people that jumped right on it and are building their own (Marek, LarryL, remowilliams, Panther, Jfcatari).

There were a few small issues that popped up, such as the solder pad holes for the battery holder being too small, and the FujiNet module's 3.3V supply needing a larger capacitor. But all in all things went pretty smoothly. I also introduced a new beta version of the TKNUC1MB firmware on November 15th, with a download at AtariAge in the World's Smallest Atari 8-bit forum thread (LINK).

EDIT: The new firmware now works very similar to the 576NUC+ original firmware, and only brings up the A8PicoCart aspect when you specifically want it via the F6 key. The 576K vs 1088K RAM selection is still automatic, being wholly dependent upon whether one of the internal Language/Game Slots is active. Yes I'll need to update the operating manual datasheet, and that will happen eventually.

I haven't updated the files yet on this website, but that'll hopefully be happening soon.

And probably worth mentioning, the SD Push-Push Socket P/N: RSMT-2-MQ-WF went obsolete soon after the project's release. Unfortunately there is no drop-in substitute that I've been able to find, and when it went obsolete it completely disappeared from the likes of Digi-key, Mouser, and Newark. So for those contemplating this build, you'll need to source this as a surplus item from eBay or AliExpress, with the cost being higher.


UGV-II
I wasn't 100% happy with the original UGV, and finally decided to do something about that. Thus the UGV-II was born.

Improvements...

1. Better quality video than even the 576NUC+, being on par with the latest UAV.
2. Non-interference with stock video outputs.
3. No modifications to Atari motherboard, other than the GTIA needs to be in a socket (W2 jumper is a non-issue).
4. Simpler and easier build for a DIY person than the previous original UGV, being mostly THT components this go around.

This got pre-released at AtariAge only 3 days ago (LINK), and will eventually make it's way over here (I gotta lot of changes to the UGV webpage in mind).


I Was in the Hospital for 3 Days
Yep you heard that right, due to a genetic deficiency my blood began clotting at an unrestricted rate and got to my lungs making it very difficult to breathe. My brother has a similar condition and went through this 23 years ago. For some unexplained reason my susceptibility to this took much longer to kick in. Anyway I got on top of it in the nick of time, and I'm doing fairly well now without any permanent damage.


1200XL-1MB Project
During my recovery I was getting a little bored, so I started a new Atari upgrade project for my 1200XL. That being to replace the noisy DRAM and its associated glue logic with SRAM instead. I called upon Juergen van Radecke (AtariAge: tf-hh) to help me out with this, utilizing a variant of his NUCplus4 1MB EMMU jed.

The new memory upgrade will piggyback 4 chips on the 1200XL: PIA, SALLY, MMU, and the 74LS138. From these it will get all the required connection points, and thus be a totally solder-less design. The prototype layout is complete, and sample boards are on their way from JLCPCB in China.

Oh did I forget to mention that this is a switchable 64K, 576K, 1088K upgrade. It's based on a RAMBO banking scheme, so the only downside is that it doesn't have true 130XE compatibility like the CompyShop banking alternative (more info on that HERE). However most programs, demos, and games never supported the separate Antic access of a 130XE, with the only examples being counted on one hand.

So that wraps it up for now.

Hopefully see all of you in the New Year.

- Michael

NUCplus4 Brief Description...

This daughter board is meant to expand the basic 576NUC+ miniature Atari computer system with a multitude of new abilities. Four at any given time to be exact. Go here for more INFO.

Assembly...

This is pretty much a through-hole board in most regards, with the only exceptions being the the Push-Push SD Card socket and the optional FujiNet Module's ESP-32 which has a castellated surface mount configuration.

So just like most such boards, it's normal to start out with the low profile components first such as the resistors followed by the capacitors, and then move into the larger ones. I'm not going to go into a lot of detail regarding this, since I'll be assuming that whoever undertakes this project from a DIY aspect already knows how to assemble and solder a PCB.

Let's get into the more critical stuff, with the first being the installation of the pin-to-pin male headers for piggybacking the PLDs on the 576NUC+ board.

In order for the NUCplus4 daughter board to sit parallel to the 576NUC+ board, these pin-to-pin headers need to be absolutely flush as possible with the underside of the NUCplus4 PCB and perpendicular to it (specifically the shoulder on the pins should be flush with the PCB). To insure perpendicularity, temporarily plugging in one machine pin socket will maintain proper alignment. These will get removed after the pin headers have been soldered in place on the top side of the PCB.

Here's what that'll look like.

After soldering the male pin headers, trim all the protruding pins inside the IC socket footprints as flush as you can with the top of the board (good quality side cutters should be used for this).

Now insert the sockets for the PLDs, and after removing the temporary alignment socket on the bottom side, solder those in as well. This will be a bit tricky, and a soldering iron with a long skinny conical tip really helps.

Now you will be ready to solder in all the remaining sockets and headers. I suggest doing it in this order, so that you don't block your access.

1. Machine Pin socket for U1 (make sure U1 is pushed all the way in with its shoulder slightly protruding from the bottom of the board).
   
2. Now insert and solder all the rest of the sockets.
   
3. Insert SIO extended male pin header (P1) from the bottom side of the board and solder it from the top.
   
4. Insert remaining headers and components and solder in place. Wherever possible insert devices (Pico Module and FujiNet Module) with their male pin headers in place to hold alignment of female headers before soldering.
   

With all the soldering done, it's now time to install the necessary number of piggyback machine pin sockets for the PLDs (2 sockets each) and U1 (3 sockets). The final stack up should look like this.

The completed assembled board should look like this when you are done.

Three chips will need to get moved (U3 SRAM, U4 MMU, U5 EMMU) from the 576NUC+ to the NUCplus4. The EMMU will need to be re-flashed with the 1024K jed.

Mods...

There's a modification that needs to happen before plugging the NUCplus4 into the 576NUC+, and that has to do with the top clearance of the 576NUC+'s 8-Pin DIN A/V jack. It needs to have some of the plastic filed off (mainly the bump towards the rear).

It'll also help to trim off the pins on the rear most row of the PLD socket with side cutters after soldering.

Plug in all the required chips after flashing any that require it, and be sure to re-flash the embedded TK-II chip in the 576NUC+ with the NUCplus4 specific firmware (consult the Datasheet for more information about this)

You should now be good to go.

- Michael

I got side tracked once again the other day, and found myself studying some posts on various websites about MIDI-CAR and it's player. I had seen some things about this a couple years back, but at the time it appeared to be focused on a PBI based hardware device only. And because my focus had been centered around SIO based MIDIMATE devices, I wasn't all that interested. However when I was having a discussion with AtariAge member flashjazzcat and the subject came up about how it would be nice if there were a new and better MIDI player on the Atari 8-bit, AtariAge member PeBe popped in and said MIDI-CAR now had an SIO driver designed for MIDIMATE and its clones, and that I should check it out.

Well long story short, I did check it out and I was quite impressed at what I saw. So much so, that I decided to package it in an ATR and put it on my MIDI page as a download to take the place of that less than desirable MIDIPLAY program.

MIDI-CAR is a full featured player. It accommodates multiple drives, and works very well with SpartaDos and SDX, as well as other DOS'es. The interface is very intuitive to use, and it has an easily accessible help screen that can be viewed anytime for specific command shortcuts, and brief descriptions of what they do.

This player also fully supports the MIDI-CLK output as generated by MIDIMATE and compatible SIO2MIDI devices for proper syncing and timing, something that MIDIPLAY did not do.

However in my testing I ran across a problem with that last aspect when trying to launch the player from FujiNet. It seems that an issue has existed with FujiNet's firmware which has been pulling the SIO CLK-IN line low during its idle state, and holding it there. This defeated the MIDIMATE's MIDI-CLK output which also shared that same SIO pin. The end result was that MIDI-CAR was not able to actually send MIDI data to the end device, thus there was no music being heard from a connected synthesizer. Luckily AtariAge member mozzwald got right on the problem and very quickly provided a test version containing a fix. Since that worked, we'll be seeing an update coming soon in the next FujiNet firmware build to come (V1.4 and before will exhibit this problem). Until that update occurs, here's a link to the patched version of the firmware.

---------------------

This MIDI player diversion gave me renewed interest in MIDIMATE devices, and I finally found time to upload some of the MIDI related stuff from previous contributors like AtariAge members Montezuma and Fierodoug5 in the form of new firmware and a new 3D printed case. Montezuma provided a SILENT "no intro" and MULE theme music version for the PIC MIDI Controller, and Fierodoug5 customized one of Mr Robot's 3D case designs for the SIO2MIDI-S2 system. Check out the MIDI page for these updates.

That covers the MIDI changes and related innovations for now.

- Michael

The original 576NUC+ case design was created by Steve Boswell (AtariAge: Mr Robot) who later released the STLs to the open source community, which I was able to modify to suit this new incarnation. This was also made possible through the use of a free online 3D CAD application called TinkerCad, a simple to use but very powerful 3D design tool.
 

New 3D Printer

I recently graduated to the Bambu Lab A1 mini 3D printer, and can now do some seriously good prints.

Previously I had been using an AnyCubic Kobra Plus that I picked up for ultra cheap as a customer return via eBay. It was good for getting my feet wet in 3D printing, but I got tired of constantly having to tweak all the settings in order to get good prints, and quite frankly the Bambu blows it away. And now I can do multi-color printing the easy way.

So the cases you see above are of my final design made on the Bambu printer.


3D Printing Station Update

I had to do some rearranging to make things fit into the allocated 3D printing station space in my shop, but in the end I think it came out quite well. And I also got serious with my filament storage, opting to use sealed containers with desiccant packs inside.

HeatSets are Out

In the previous 576NUC+ case design, the upper and lower halves depended upon heat set threaded brass inserts for the screws to screw into. However I wasn't thrilled about using these since they are very dependent upon correct insertion to avoid cross-threading issues. So in this new case design I've gone with screws and nuts, with both being recessed into the case halves, and black in color to make them virtually invisible.

I sourced these from McMaster-Carr, but probably something equivalent could be gotten at your local hardware store, not requiring a minimum purchase of 100.

P/N 91249A618  Black-Oxide 18-8 Stainless Steel Pan Head Phillips Screw, 4-40 thread, 1-3/4" long, pack of 100

P/N 90328A101  Black-Oxide Zinc-Plated Steel Locknut, 4-40 thread, pack of 100


Got Feet?

On the bottom of the case there are 4 circles showing the location of the rubber feet (Clear Vinyl actually). These reference circles were sized at exactly the same diameter as the feet produced by EverBilt which can be purchased either from Amazon or Home Depot.

P/N 158 724 1/2 in. Self-Adhesive Vinyl Surface Bumpers, 16-Pack


- Michael

Previous Post