C64 Keyboard
This project is still work in progress!
Actually, I never wanted to develop a C64 keyboard.
1. There are people, who have/had indigogo crowd funding campaigns about such keyboards and I don't want to interfere with their business.
2. There will always be a keycap problem.
3. It is a hell lot of work.
I would have backed both crowd fundings, indigogo didn't let me, because my credit card requires a security procedure, that they do not provide.
A few days before Christmas 2024, Retro Fuzion from Australia contacted me and asked me, if he can donate me a case for my "merits for the Commodore community". I feel pretty and grateful for that. Since I have some experience with sending parcels to Australia and didn't expect the cases before end of January 2025. But they arrived on new years eve. Now, I had a need for a keyboard. I might eventually get a CBMStuff Blingboard64, which I would really like. I have already backed his keycaps.
Actually, I never wanted to develop a C64 keyboard.
1. There are people, who have/had indigogo crowd funding campaigns about such keyboards and I don't want to interfere with their business.
2. There will always be a keycap problem.
3. It is a hell lot of work.
I would have backed both crowd fundings, indigogo didn't let me, because my credit card requires a security procedure, that they do not provide.
A few days before Christmas 2024, Retro Fuzion from Australia contacted me and asked me, if he can donate me a case for my "merits for the Commodore community". I feel pretty and grateful for that. Since I have some experience with sending parcels to Australia and didn't expect the cases before end of January 2025. But they arrived on new years eve. Now, I had a need for a keyboard. I might eventually get a CBMStuff Blingboard64, which I would really like. I have already backed his keycaps.
The wonderful new Retro Fuzion breadbin cases. I have got a clear one and a blue one.
And that is the next problem: I would really like to have C64 keycaps for Cherry MX switches. Those are so popular and wide spread, that it is just the way to go for a new C64 keyboard. I am not so sure, if I can 3D print adapters for the old keycaps with my FDM printer. A resin printed is just nothing, I want to get, due to the nasty resins and the smell.
Anyway, I just started with the work for a C64 keyboard, not really sure, how or if I was able to solve the keycap problem. I was not even sure, what the keyboard was going to be. I just knew, that there will not be a light show on it. That is business of theindigogo guys. First, I got out the plastic cover of a C64 keyboard. and measured every switch position with a caliper. Great, that I have a 30cm caliper, not just the standard 13cm one. It is not long enough for the complete keyboard, but it is working with most of the keys. The keyboard is 385,6mm wide. The keys, that I could not reach with the caliper were measured from the right end of the keyboard. My method was measuring the distance from the left side of the board to the right side of the little cylinder and then subtracting half of its diameter. For the vertical distance I have measured from top of the plastic cover to the lower side of the cylinder, subtracting half the diameter.
I entered all the measures in an excel table. After doing so, I have calculated the average distances, which are pretty close to the standard distance of 19.05 mm. That is a 3/4" and in the world of keyboards, it is called 1U. Then I have used the average distances to calculate an evenly spaced keyboard and determined the most off measures. I then repeated the questionable measurements. Some were indeed off, some were just what I had measured, no matter how often I measured them. I got a working delta x of 19.04 mm and delta y of 19.05 mm. I have kept the 19.04mm for the delta x, since I had a better overall standard deviation with that value. the top row has 16 keys, so being off 0.15 mm compared to the standard 19.05 mm is less than the usual tolerance of 0.2 mm. Some of the switches on the original keyboard were even 0.3 mm off and I didn't notice it visually.
The standard and the extra long vernier caliper.
After calculation the position of every switch, it was time to start the Eagle design. First, I have drawn the keyboard matrix. I used the "Attribute" of every of those keys to keep the information of the actual letter, that is generated by the switch. For the keyboard matrix, I have used the commonly known schematics.
In the Eagle board editor, I have drawn the outline of the board and then coarsely positioned the switches. The exact positions were entered in the "device properties". After having a properly placed keyboard, which is quite a bit of work, I measured the outlines of the key caps. The normal 1U key is 18mm x 18mm. Most of the wider keys (except RETURN) are 1.5U. That means, they are 0.5U (9.53mm) wider. The said RETURN key is 2U (37mm) and the space bar is a 9U space bar.
I did not find any 9U space bar, so why not add to more switches left and right from the space bar (1U + 7U + 1U = 9U)?
Since the Cherry MX compatible 1.5U keys have the switch connector in the middle and the Commodore key caps have two of those, I had to have alternative (overlapping) positions for the switches, so one can install original keycaps (with an adapter) or modern keycaps. I always want to leave the user as much as possible freedom with my projects.
Eagle design of the C64 keyboard
I have noticed, that there would be space for an Arduino Pro Micro and a 16bit I/O (PCF8575) module with I²C interface. That would connect to the keyboard matrix. The pro micro was chosen, because its CPU provides a USB interface, which then can be configured as a keyboard (HID device). This way, my keyboard could be used as a USB keyboard for VICE.
Some years ago, I have made a keyboard controlled kernal switcher for the C64, which is pretty unknown, I think, but it is one of my favorite gadgets. It lets the user select, which kernal (1 of 8) the user wants to select, by pressing RESTORE + number key. It can also reset the C64 by holding the RESTORE key for some seconds and issues an EXROM-reset by holding the RESTORE key even a bit longer. This feature can be incorporated in the keyboard.
The programmable keys could access the keyboard matrix via the said I/O-module and produce typing the programed key.
Oh, I forgot the shift lock key. Of course, it is not latching, but it should work that way. First, I had to determine, how bouncy such a Cherry MX key is (all mechanical switches produce bouncing, when they change their state). There are some different ways to eliminate the bouncing I could have done it in software, but I wanted to keep the Pro Micro optional. And I had the ambition to get a 1 IC solution (I have seen, most C64 keyboard developers have a 2 IC solution).
Bouncing of the switch at release
The most bouncing occurs on release (a pressed button produces 0V, a released button +5V). In the previous image, it is two divisions, which is about 100 µsecs. I am using an R/C circuit to eliminate the bouncing. Then comes a toggle flip flop circuit (from a 7474 D-FF). For switching the left shift, I am using a transistor. That is inspired by the circuit of the SX64 (I/O-Board), which also uses an NPN transistor for the shift lock.
Shift lock circuit on a bread board.
I haven't tried out any circuit on a bread board in my job, but since the keyboard PCBs will be pretty expensive, I thought, it might be better. It was fun, too.
Actually, that would be a complete circuit for a C64 keyboard, the said Pro Micro can actually be optional, but it would add the following features:
● Kernal Switcher
● USB Keyboard (like for VICE)
● programmable switches
● A WS2812B RGB-strip interface (there was a GPIO pin left, so why not).
The keycaps are a major problem. The best keyboard is hard to use without. Generally, it will be possible to use the keyboard with original keycaps plus a wobbly adapter pin, but that is not the way of my choice.
So, I have researched methods of printing blank Cherry MX keycaps at home. Well... some investments are required.
Dye sublimation is a way to get the letters on the plastic. That requires an inkjet printer, that is capable of printing dye sub ink. The ink will be printed on a special paper. With heat and pressure, the ink will be transferred to the keycap. I have seen people doing that with a hair iron with adjustable temperature. I think, this method is nice, especially, since you can use the printer for normal printing, too, but it will take several minutes per keycap.
The next candidate for solving the keycap problem is laser engraving. This seems to be faster and probably more accurate, since you don't have to place a paper upside down on the keycap, but have to adjust the position of a laser projection of the outline on the keycap. The downside is the requirement for a 1300€ investment, the laser engraver (IR laser, which makes it capable of engraving plastic and metal).
Since I see myself engraving my logo or whatever on everything I have in the future, I have decided for buying a LaserPecker LP3. This is working with a galvanometer and mirrors for positioning the laser beam, so it is faster and provides a higher resolution than the laser engravers, that have a moving laser head, like a 3D printer or CNC machine.
Before printing comes drawing. And before drawing comes scanning the keycaps.
Actually, that would be a complete circuit for a C64 keyboard, the said Pro Micro can actually be optional, but it would add the following features:
● Kernal Switcher
● USB Keyboard (like for VICE)
● programmable switches
● A WS2812B RGB-strip interface (there was a GPIO pin left, so why not).
The keycaps are a major problem. The best keyboard is hard to use without. Generally, it will be possible to use the keyboard with original keycaps plus a wobbly adapter pin, but that is not the way of my choice.
So, I have researched methods of printing blank Cherry MX keycaps at home. Well... some investments are required.
Dye sublimation is a way to get the letters on the plastic. That requires an inkjet printer, that is capable of printing dye sub ink. The ink will be printed on a special paper. With heat and pressure, the ink will be transferred to the keycap. I have seen people doing that with a hair iron with adjustable temperature. I think, this method is nice, especially, since you can use the printer for normal printing, too, but it will take several minutes per keycap.
The next candidate for solving the keycap problem is laser engraving. This seems to be faster and probably more accurate, since you don't have to place a paper upside down on the keycap, but have to adjust the position of a laser projection of the outline on the keycap. The downside is the requirement for a 1300€ investment, the laser engraver (IR laser, which makes it capable of engraving plastic and metal).
Since I see myself engraving my logo or whatever on everything I have in the future, I have decided for buying a LaserPecker LP3. This is working with a galvanometer and mirrors for positioning the laser beam, so it is faster and provides a higher resolution than the laser engravers, that have a moving laser head, like a 3D printer or CNC machine.
Before printing comes drawing. And before drawing comes scanning the keycaps.
Scanned C64C/G keycap
Unfortunately, the surface of the keycaps is not flat, hence the resulting images are distorted. A simple image processing to get the print is not possible.
Fortunately, I have fount out the font, that the keycaps use: ITC AvantGarde LT Medium and AvantGarde CE. The PETSCII characters are comparatively simple and it did not take a ton of work.
Inkscape drawing of the S keycap.
The keycaps are not perfect clones, but close to the original. The (DSA) keycaps, that I have chosen, have a different shape, than the original. Maybe with more experience, I will find a profile, that is closer to the original keycaps.
The complete keyboard drawn in Inkscape.
Blank DSA profile keycaps (for Cherry MX switches)
When reading about the keycaps, I had to learn quite a couple of things:
1. There are many different shapes/profiles of keycaps.
2. some profiles have different shapes depending on the row, some have equal shapes in every row.
3. For dye sublimation, you have to have a more heat-resistant type of plastic. PBT (polybutylene terephthalate) is suitable for that, ABS is not.
I did not decide, which "printing method" I would be using, when I have ordered the blank keycaps, so I ordered PBT ones, which I found ot that I liked them. The surface feels nice.
1. There are many different shapes/profiles of keycaps.
2. some profiles have different shapes depending on the row, some have equal shapes in every row.
3. For dye sublimation, you have to have a more heat-resistant type of plastic. PBT (polybutylene terephthalate) is suitable for that, ABS is not.
I did not decide, which "printing method" I would be using, when I have ordered the blank keycaps, so I ordered PBT ones, which I found ot that I liked them. The surface feels nice.
Some keycap profiles
The previous image shows some of the available profiles for keycaps and their approximate heights. It is a compilation of multiple sources on Internet. Further, it is not an accurate drawing, but a graphic approximation of the actual keycaps.
I have ordered the keycaps some time before I decided for the printing solution, that I will make use of. DSA keycaps to have a good availablity, the top surface is not very curved, so there is neither a great distortion of the text on a curved surface, nor a focus problem for the laser.
I have ordered the keycaps some time before I decided for the printing solution, that I will make use of. DSA keycaps to have a good availablity, the top surface is not very curved, so there is neither a great distortion of the text on a curved surface, nor a focus problem for the laser.
LaserPecker LP3 laser engraver
When the laser engraver arrived, I have tried to engrave some sample objects, which came with the LP3, but I could not wait until the next morning to try it out on the actual keycaps, so I have done a couple, before I went to bed.
The first keycaps, that I have engraved
Actually, I have worried a bit about a consistent positioning of the text on the keycap, but that is not really hard. I have made the size of the PNGs, that I use for engraving identical to the size of the top of the key cap and there is a laser frame projected in the size of the actual picture. If it is any misalignment, this frame goes down the sides of the keycap, quickly.
Aligning the print on top of the keycap. It is easy, but don't forget it!
There are some dangerous aspects to laser engraving. Maker's Muse on Youtube has made a good video (do not click the URL if you are not ok with Youtube collecting your data) about it, which I would recommend. Mainly, I see a risk for your eyes and those of the people and pets around. Don't engrave without proper eye protection and the safety hood. It is an IR laser, which is invisible. The next aspect are the fumes of the things you actually burn with the laser. I have noticed a pretty decent amount of precipitate on the inside of the safety hood. The exhaust fumes are transported out of the box without filtering. There is only a fan.
Laser engraving a keycap. The red dot is just a marker.
Engraving one keycap only takes a few (maybe 15) seconds (excluding the time for selecting and aligning the text). Actually, engraving is not, what it really is. It does not have any depth on the keycaps. It is just darkening the plastic.
A laser engraved keycap
The result is a dark grey, but not black text. I have tried to mix in some kind of toner, like pencil dust, crayon or putting a flimsy paper on top, but that only caused problems and did not produce a darker "print".
Comparison of the shift lock key with different print parameters
Finally, I have found some parameters, that produce a slightly darker text on a white keycap. Actually, I have even used less power for engraving.
Grey keycaps with different speeds and power
Grey keycaps can be laser engraved as well. The result is a relatively light text, which is also readable. More power does not always help, as can be seen from the previous image. The text looks charred and even boils up.
Lasering transparent keycaps did not produce a decent result, yet.
The complete set of keycaps vs. the original keyboard.
I have improved the keycap text, that was not accurate enoug:
● pound
● up arrow/pi
● all f-keys
Final keycap text
Work to do:
● Constructing a keyboard holder/frame that can be 3D printed
● finishing the layout (done!)
● ordering the prototype PCBs
● assembling a first prototype
● writing the software for the Arduino Pro Micro
● testing
● documentation or iteration