Commodore CHESSmate Reproduction

After complaining about how hard these devices are to come by, I actually managed to win an eBay UK auction for a working Commodore CHESSmate. I’m pretty pumped. Now I won’t have to impose on Peter for a loaner.  

Imagine my surprise when I tried using one of the world wide eBay search engines and this listing popped up 

 https://www.ebay.ca/itm/186196149458

So happy. 

My CHESSmate arrived. It took about eight days from the UK. Not bad. The package was very well wrapped. Kudos to the seller.

After removing the packaging, I was surprised to see how good the box looked. Sure it's a bit banged up, but remember this CHESSmate is 45 years old.  

No I did not rip the box apart ;-) I took pictures of all sides of the box and put together this exploded view:

As you can see the features are listed in English on the top and bottom, German (where CHESSmate was very popular at the time) on one side, and French on the other.

More surprises when I removed the CHESSmate from the box.

Both the original cardboard insert and users manual were intact. Bonus. And it keeps getting better.

The CHESSmate itself is in great shape. 

Next Steps

Well I would like to power up the unit, but before I do I will need to get a 10V 600mA power supply with a North American plug since it looks like the one it came with is 240V only. Also need to find a matching plug to connect to the CHESSmate (wish me luck).

Of course before I apply any power to the unit I will take it apart and check the PCB for corrosion and bad electrolytic capacitors. Stay Tuned.

To open the case remove the 4 screws marked in the photo below. 

You should then be able to separate the top and bottom pieces of the case. 

Setting aside the bottom for the moment, remove the four screws (red circles), two nuts (blue circles), and the two bolts holding the power connector on (green circles) then lift out the PCB, being very careful not to damage the ribbon cable between the PCB and membrane keyboard.

To detach the membrane keyboard from the top of the case, remove the 8 nuts highlighted in red.

And that's about it. Only four major parts, case top, case bottom, PCB, and membrane keyboard, and one minor part the attached power connector. 

Now with the case pieces completely free of components, they can serve as the basis for the 3D models I will need for my reproduction.

In virtually every other reproduction I have done I have never had an original to work from. This is a rare treat for me. I get to use my calipers for a change. Nice.

Rather than try and cobble together a North American power supply for the CHESSmate, I ended up getting a 110-120V to 220-240V adapter. Since I had the machine apart and the electronics checked out, no rust or leaking caps, I reassembled the CHESSmate, plugged it in, and it powered right up.

Since there is no screen, you really need to have a physical chess board to keep track of the moves when you play against the machine. I dusted of my old chess set (literally it has been on a shelf for decades) and my equally dusty chess skills (which are pretty anemic to begin with). You can see my game in progress above. Notice I was even able to castle.

So not only am I now able to make precise measurements when modelling the case in Fusion 360, when the time comes I will be able to use the CHESSmate to validate that my emulation is running correctly. Bonus!

... having an actual physical object to model from is a joy.

I'm going to have to work on the colors, but I'm pretty happy with my first crack at the CHESSmate case.

I've simplified the model somewhat. I may go back in to add some cosmetic details later.

I did make one outward change, to the power connector, opting for a more modern micro USB connector (because I had some of these lying around).

From the inside I modelled a "shelf" to hold the power connector. While I still like the idea, it turns out that once you have soldered the header to the PCB, it's pretty much impossible to insert the combined unit onto the shelf, not to mention the pain of removing support material that is required to print the shelf. I'll rework that in the next version. 

On the inside I have removed all of the "posts" that were used to hold the CHESSmate PCB in place and to attach the top and bottom case pieces together. I have other plans for doing he same.

I haven't figured out all the details for how the "membrane" keyboard will work , but I do know that I will be needing the artwork that rests on top of the switches, 7-segment displays, and LEDs. So I spent a couple of days re-learning Inkscape (since I don't use it enough to be truly proficient), and produced this:  

I started with a scan of the originals keyboard:

which I brought in to Inkscape at a 1:1 scale as the bottom most layer and basically traced the various elements using Inscapes vast array of drawing tools on "higher" layers. I was also lucky enough to find a font, "Christ 100", that was a pretty close match to the originals which saved a lot of time.  Not so lucky with the A-H 1-8 button characters which had to be carefully traced and rendered as lines and Bezier curves.

As a quick check of the fit:

• I printed the artwork on a transparency (to avoid ink bleeding into paper),
• laminated the transparency backed by white cardstock using 3 mil film,
• cut the keypad membrane artwork to size,
• and attached the result to the inside of the CHESSmate top case with two sided tape.

A perfect fit.

I continued to work on the membrane. Specifically the white backing for the art work (which is printed on a transparency) needs to have cutouts for the seven segment displays and LEDs.

I created a DXF file for the cutouts and for this version laser cut the membrane backing from a piece of "photo" paper. Then I lined up the art work and backing and laminated them together with 3 mil sleeves. The membrane created feels pretty good. I'll have to try it out for a bit and see if it needs adjusting, like maybe using a white vinyl sheet instead of paper.

I used an old soldering tip on my iron to melt away the laminate blocking the mounting holes. Speaking of mounting, I created a prototype for the mounting "block" that I will be using to bridge between a PCB and the membrane.

Without going into a lot of detail now, on one side of the PCB I will mount the seven segment displays, LEDs, and buttons. The other side will have all of the other components.  The thickness of this block will be adjusted so that the displays, LEDs, and buttons will sit just below the membrane, and at the same time the block will support the rest of the membrane. 

For now I just printed a thin "prototype" block that will validate the placement of the cutouts and hold the membrane I just made in place. 

You will also notice in the above picture that I flipped the power connector holder to make it easier the install.

I designed and printed a snap-on clip to hold the connector in place. (NOTE: There was a thin "wall" above the connector that was damaged when I removed supports. I fixed the problem on this print with my trusty 3D Printer Pen, and adjusted the model for a more permanent fix going forward.)

So the case is pretty much finished. I created a “commodore” logo in Inkscape, printed it on transparency, and attached it to the case with a white paper backing. 

I found some better colors for the case but because the color choices for PLA filament is limited I think this might be as close as I get.

I sure hope that the "insides" turn out to be as good as the outside did.

With the case done I started thinking about getting the electronics going. Now I don't usually start making a PCB by adding some of the silkscreen layer first, but in this case it makes sense. I started by creating a new Inkscape design base on the full membrane from which I removed all of the color fills, and some of button icon details.

Then I imported the resulting SVG file into the KiCad 7 PCB Editor.  To do this select File->Import->Graphics... and you get the following dialog:

I imported the file into the Silkscreen layer with no scaling, with mixed results. The good news was that the graphics were perfectly scaled (see right hand picture above). The PCB will be the exact size of the membrane.  Outlines of the seven segment displays, LEDs, and buttons in the silkscreen will allow me to place the footprints for their components precisely onto the PCB.  

Unfortunately the silkscreen text did not translate well. Bold text was too "fat" with the letters barely recognizable. Smaller text was missing in action. Annoying but not a problem, I removed the text from the SVG file, re-imported, and added the text manually in KiCad. 

I wanted to make sure the text was in place so that the CHESSmate could be run as just the PCB without a case.

A good start.

As I was working on the PCB I realized that aside from poking around on the original CHESSmate to make sure it was working, I had not actually played a game against it. 

Since the CHESSmate does not have a display to show the board you need to have a physical board to keep track of the moves. To that end I purchased the small travelling chess set you can see below (https://a.co/d/i9xmQcL) This was one of the least expensive sets, but I'm very impressed with it. Solid build. Pieces are smallish but have a nice feel. There is even an extra Queen for pawn promotion. Also a plus, the rows and columns are clearly labeled, which you need when entering and reading moves on the CHESSmate. 

Now I'd like to say that I was playing black in the above game, but I wasn't. I was never a strong chess player, and I did make a few unforced errors, so I'm not too surprised. What I was surprised about is how much fun I had. 

Unlike modern AI based chess programs, there is a good chance that an average person (if they concentrate and don't make stupid errors) can beat CHESSmate. This is not because the programming is lacking, it's actually very clever. The reason is that the programming is constrained to only 2K of memory. 

I'll do better next time.

Update 2024-01-14

I brought the CHESSmate to a monthly Retro Computer Club meeting this afternoon that takes place at my local makerspace, Kwartzlab. It just so happens that there was another gentleman there who brought a TIMEX sinclair 1000 (ZX-81) with a chess program, so we tried to have a match between the two machines.

Unfortunately we could not seem to make it through a whole game without messing up when keying in the moves. Neither program had a way of "taking back" a move once it had been entered. Oh well, it was fun none-the-less trying to guess what the next machine move would be. We didn't get far enough to decisively determine which machine had the upper hand, but if I had to guess, the 16k program on the ZX-81 might have been stronger.

I just finished my initial cut at a PCB for the CHESSmate. This first PCB is an IO board only holding the buttons, 7-segment displays, and LEDs in the precise positions so that they line up perfectly with the membrane.  I also threw in a piezo speaker circuit.

I ended up having to make my first 4-layer PCB primarily because the components were in fixed positions and could not be altered to obtain a more favorable layout for putting down the traces.

Having IO separate from compute will make it easier to test. In addition this configuration will support multiple "back-ends". For instance an emulation implementation that I am leaning towards and/or a "pure" solution replicating the original hardware.

I still have a lot of testing to do before I commit to fabrication. I discovered that 4-layer PCBs are quite a bit more expensive than 2-layer so I want to be sure of a good result.