progress to Sept 1


I finally got the coloring close enough on the yellow and red light cubes, these are now labeled along with the four white ones. Once the latest try at forming two green cubes was complete, I pronounced them acceptable and slapped on their labels.

The method of drilling the slots on the back of keycaps worked fine - using a 1/16" drill bit on the press, it was fairly easy to form the slot into which a metal tang will insert when the cap is pressed onto the keyboard. Having finished all the text decals for the key-caps on Monday, I began applying them to all the caps, finishing with a spray of clear enamel as protection. They were all drilled and pressed into position on the keyboard.

I have a full complement of buttons, but am not satisfied with the quality of many of the blue colored buttons so will recast them during the week, one per day in my best mold. Fortunately, I have enough to complete the 1130 main keyboard/button panel, the remaining set are for the black plate on the front of the typewriter which I haven't started building yet.

A few decals still had to be produced, due mainly to failures in the processing of earlier decal fabrication. These needed a new sheet printed at Kinkos, which I took care of on Tuesday while traveling between home and Techshop. By Tuesday evening, they were all created and most were already rubbed onto their respective buttons or keycaps. All that is remaining are the buttons and handles of the console printer (typewriter).

Buttons, light cubes, keycaps, toggle switches etc 
In the picture above, the reason there are only eight key-caps is because all the rest of the keys on the keyboard are already titled the same as the 1130. It was a clone of an IBM 029 keypunch, which is the mechanism that IBM borrowed to put into the 1130. They changed eight keys - so have I.

Observant readers may notice that I only have seven light cubes along the left - the eight position could have a toggle switch like the ones just under the logo plate, but the machine I am building has another blankwhite cube in the eighth position. I currently have it inserted into the black plate for tests of the illumination from LEDs that sit underneath each cube.

There was a setback in the project, due to a missing key-cap. The cap for @ and # had been removed to provide access to the adjacent KB REST key, but misplaced. After hours of searching in vain, I poured an additional brown key-cap that will be labeled #/@ . Of course, once that was setting in the mold, I did find the key-cap.

The three typewriter push-buttons were labeled (SPACE, TAB and RETURN) as they were each taken out of the mold, sprayed with clear enamel and set aside. By the end of the week I had all buttons complete and ready for insertion. That is, once I make the black plate for the front of the typewriter enclosure and make the enclosure itself.

All the key-caps were drilled and slots cut in the back, but those were not precise enough to guarantee perfect alignment of the key-cap against the original key-caps on the mechanism. I therefore decided to epoxy these new caps into place, which gave me the opportunity to shim them into the desired alignment while the glue was setting. It was a slow process, done a few caps at a time, but gave me an acceptable keyboard at the end.

Key caps set onto keyboard mechanism but not yet aligned and epoxied

I spent some time on the 3D printer (Makerbot Replicator), spitting out a model of the tab switch to use in forming a mold. The part has a cylindrical hole open from the rear, into which the bat handle of the toggle switch will fit.

Because of the shape of the handle with its curving wings, it can only be printed in one orientation, but that causes the cylindrical hole to be up in the air, parallel to the printer table. The printer could not span the top properly, thus there is a hole vertically in the printed object where all the deposited ABS filaments just fell into the cylindrical hole.

I ran to a workstation and tried to close up the hole, planning to drill it after casting the parts in blue resin anyway, but didn't get the plug deep enough. Now I have a partial hole at the top, sagging into a blind cylindrical cavity inside the handle.

Since I need to choose a "gate", the mouth of the mold, I picked the side of the handle that had the sagging hole. I hot-glued that face down inside a small plastic box, then poured silicone mold solution around it. When I pour blue colored casting resin into the mold, I will have the handle recreated just the way it sat on the 3D printer except that the sagging hole will be gone, filled in by the top of the resin poured into the mold. The surface won't be very smooth and polished, but it will likely be good enough for this infrequently activated control on the typewriter faceplate.

With the handle successfully out of the mold and overflash trimmed off, I am now evaluating ways to increase smoothness. The first trial was the use of acetone to partially melt the surface, to remove some of the digitalization that comes from the stacking of plastic layers. Another method would be to sand down the part with a series of increasingly fine wet sandpapers until it is smooth enough. A third method would be to apply a solution made by dissolving some ABS plastic in acetone, the slurry can be wiped or brushed onto the parts to form smoother top layers. The same techniques may be needed to make the typewriter knob domes smoother.

My handle was smoothed out enough to pass, brushing on acetone after sanding down the raised flash and wiping it down with Kimwipes to finish off. I placed the lettering (CLR/SET) on the handle and covered it with a clear enamel spray coat.

I produced the tab set/clear handle with my new mold, then need to drill a hole in the rear to place it over the bat handle of the toggle switch it will activate. However, the bat handle is too wide at its tip to safely fit inside the handle - the wall clearance would be very slim and the precision of the hole location is a bit beyond what I can reliably achieve.

Original bat handle, too large to safely fit into hole behind tab handle - right
It occurred to me that I could grind down the top and bottom of the bat handle, so the clearance was more manageable. It would require me to drill a few smaller diameter holds in the rear to form a slot. The tang I created by grinding the bat handle would fit into that slot. I also had to consider the required depth of the handle on the switch - since it has the two wings that will move behind the black plate, and the front of the handle will protrude through a slot on the plate, which sets the depth of the drilling in the handle.

Bat handle after grinding to make it an easier fit into handle on right
I successfully ground the handle to a slot shape and then drilled the backside of the handle to match. Here is the handle, in position in the typewriter switch plate.

TAB SET/CLR switch handle in place on typewriter switch plate
As you can see from the picture above, the blue 'wings' of the tab switch handle are not far enough back. The console entry switches, of which you see two on the right in the picture, must stick out of the slots far enough to be manipulated by the system operator. I will need to grind the switch to make the flat part extend further back on the handle toward the switch body, so that the tab handle will slide on deeper.


The typewriter knobs are taking longer - they need a two part mold which I haven't done before, and the top (dome) surface is not smooth. It displays the steps of the 3D printer as it deposited layers of plastic to make the knob, thus there are bands of concentric flat spots that together roughly match the dome shape.

The first attempt at building the two halves of the knob were fairly successful, but they were done with standard resolution. I think I should switch to high resolution, making the ridges and texturing smaller. That should make the smoothing easier to accomplish. Alternatively, since the remainder of the knob is smooth enough for my purposes, I just need to smooth the dome or cover it with something smooth.

I decided that if I could glue layers of plastic atop the dome, it might smooth out the surface enough to make the mold. I experimented with a plastic bag that was packaging for a USB cable - holding it in place by hand suggested the results might be quite good. If I could find a slightly thicker plastic film, it would mask the steps even more, so the hunt is on.

I assembled some acrylic plates to form a temporary box in which I can build the mold. I will be making a two piece mold with a part line around the rim of the typewriter knob. The first step is to form the inside portion which also must have the gate for entry of the resin and air holes at crucial points to avoid voids in the casting caused by trapped air.

I have picked places for the air holes and gate, which I will form by epoxying some screws and nails to the inside area of the model. This part of the model will be glued flat on the bottom of the box, with the screws and nails sticking straight up higher than the depth of my mold pouring.

In addition, I will hot glue a few "features" on the bottom plate which will form holes in the face of the mold. Those holes will create matching protrusions when I cast the other half of the mold with the other half of the model in the next step.

I have the glued up bottom ready to go, unfortunately the acrylic box I glued together to make the mold is not liquid-tight. I either have to get this sealed properly or dump the box and use something else in which to pour the mold.


My mounting method required a new improved version of the holder to keep the switch attached to the button holder plate. In my prior design, the toggle switch could twist so that one edge slipped into the button holder plate opening, which skewed the toggle handle out of position.

To fix this, I cut and drilled some plates to mount on the switch, over the threaded barrel that normally holds a switch onto a panel. These keep the switch in proper alignment. The other issue was the holders for the toggle switches didn't have enough material on the ends to hold when inserted into the button holder assembly.

I cut the new holders on the waterjet at Techshop on Tuesday. These worked well and the switches are now fitted into place. There is still a bit of play when the switch is thrown, something I need to deal with. I believe some kind of rubber stop or soft wedge is the right answer, given the complexity of the mounting hardware already in place.

The base plates under the toggle handles cover the rectangular opening of the black plate, leaving a thin slot for the switch handle to extend through. One of the base plates is red, but both were initially produced with white plastic on the 3D printer.

I printed new baseplates on the 3D printer on Tuesday, one in white and one in red. The new design sits flatter in the opening and has a better recess on the underside to fit against the dome shape of the switch itself. They also snapped right into position in the black plate, exceeding my expectations. The only downside is that they are not glossy smooth surfaces, but I don't think that matters with their position on the machine.

I drilled the slots on the backside of the toggle switch handles, then tested their fit on the switch assembly. The handles fit but not deep enough and they let the toggle handle rotate on the tang. I then had to deepen the slots and test again.

I epoxied the toggle handles onto the switches. Although that is permanent and will prohibit removal of the switches for any repairs, it was quickly becoming the only way I could make these handles stay on the very tiny tangs of the switches. I was also able to make and insert wedges that keep the switches locked into place in the desired alignment.


There are small logos that go on the front door of the 1130 and on the top cover of the console printer (typewriter), but neither of those sheet metal parts are fabricated yet, or even designed.

I have decals prepared for the smaller text, but haven't decided how to handle the larger IBM lettering or remainder of the logoplate images.


I have the buttons and some of the light cubes for the 2501 control panel (the IBM 2501 was the fast card reader that was sold with 1130 systems, the slower option was the 1442 models 5 or 7). What I need to make to complete the set are three light cubes - one white with the title ATTENTION and two yellow cubes labeled READ CHECK and FEED CHECK.

I have a white cube without titling sitting around already, thus only had two to fabricate. The first amber cube is done and the second is in the mold overnight. Once I print the titling, turn them into decals and apply to the cubes, the panel with its buttons and lights can sit waiting for me to build the Documation card reader's psuedo-2501 enclosure. 

No comments:

Post a Comment