Show older

If this seems to you a bit out of the blue then you're a bit correct, but I have a history of reproducing real and theoretical intellectual infrastructure (e.g. my two Memex builds) and I used to work as a prototype engineer in the PARC lab where the Altos (and laser printing, and Ethernet) originated. That all happened well before I worked there, but that's how I'm connected to that community.

Now that I have permission to share this work I'll start posting WIP photos, videos, and snippets of the design, build, and installation of the replicas. I'm hecking excited to share!

In the meantime, if you're into the Alto then definitely swing by this link to see a metric boatload of photos and document scans like this one from the hardware manual.
bitsavers.org/pdf/xerox/alto/

While looking into texturing techniques to replicate the original cast parts of the Alto displays I ran across this video about wall texturing and painting techniques. Now I may never be satisfied with flat, boring walls.
youtu.be/2XX5QRMKG54

I have a few rollers ordered but if those don't work out then I might need to print TPU rollers or CNC cut foam blocks for a custom texture. The last time I tried to replicate a surface texture was before I had additive printers or CNC cutters so this is a new kind of fun for me.

It's fun looking for less than a dozen steel hemispheres because their listed uses are so random. It's a handrail topper. No, it's a chocolate mold. No, it's a salad bowl. No, it's a fire pit!

So far I'm pretty happy with OSH Cut. Their instant quote system is pretty good at telling me how to tweak my designs to get the best cut and bend. The one hiccup I had was quickly and effectively responded to by a support person, Emily. I ordered a couple of samples to see the quality of their cuts and finishes and they look good, so my next step is to polish the design a bit and then order parts for a prototype!

The display swivel mechanism design is coming together. It's basically a cylinder and a hemisphere that squeeze a hemisphere that's welded to the display's base.

Here's one a reference photo of Kay's Alto that shows the display resting on the base's hemisphere. He told me that in retrospect they didn't need a free swivel like that because they only ever tilted the display up and down like nodding your head. It could have been a single-axis.

There's also a sliding potentiometer peeking out from the bottom of the display. I can only assume that it allows the user to change the intensity of their 70s vibe.

A Creality CR-10 S5 is coming my way so I need to rearrange my shop to make room. It has a print volume of 500mm x 500mm x 500mm so it'll handle Alto case parts without adding unnecessary seams. 😸
creality3dofficial.com/product

I plan on reselling it after the Alto replicas ship and I've printed helmets and busts for a few friends. I try not to keep any of my work (except my custom tools) and I generally make smol mechatronics so this big printer isn't a good use of space in my decifactory.

I sort of like the inebriated look of this sweep before I add parametric constraints. In the future it might be fun to make a desktop display with this sort of semi-organic and irregular shape.

There are roughly eleven bazillion shades of old computer beige. I want these replicas to look like they are old new stock that we recently found buried in PARC storage.

This morning's task: Model the case that will slip over the metal frame of the base. The metal frame will be laser cut and bent so I had to design in some material on the front arms that will be ground away to fit inside the case.

Here are the labels attached to the back panel of the Alto model II display that I'm using as reference. It was manufactured for PARC in St. Paul, Minnesota by Ball Electronic Display Division in 1979.

TIL about the Alto "nose boot" in which eleven keys needed to be simultaneously pressed in order to set specific bits during boot of an early build of the OS.
history-computer.com/xerox-alt

It's not in scope for this project but once I've delivered the Alto display replicas I'll be tempted to replicate the custom Honeywell keyboard that slips into the front cutout of the base. Along with the chording keys and mouse, it would complete the user-facing aspects of the Alto. The Dorado CPU could also drive an Alto display when necessary.

I drew up the bezel in which is the last major external element of the Alto. There are still a number of internal features like fasteners and the LCD fixture as well as a few tweaks to the curves to more closely match the curvaceous cases.

In testing the CR-10s5 to prepare for printing the Alto's large case parts, I've come to the conclusion that I need to design custom rafts that can minimize edge lifting and handle it when it does occur. Bed adhesion will be a struggle, especially since the outer edges of the bed aren't heated.

The biggest printed part is ~450mm in its largest dimension which takes a bit of trickery to print without failure.

Cura estimates that a standard quality print with a 0.4mm nozzle of just the back case of the display will take around six days and 1.4kg of filament. So... I really don't want failures. Once I've nailed a CR-10s5 profile for the stock nozzle I plan to switch to a 0.6mm nozzle and run a few tests to balance speed and quality.

Ok, after much futzing I have a working profile for the stock CR-10s5. Leveling a bed this big is a different skill than leveling Ender 3 sized beds. The longer lever arms at the corner springs and the greater flexibility of the larger surface gave me fits at first. The next step is to swap in a 0.6mm nozzle and the engineering PLA to tune their heat, flow, and adhesion for days-long prints.

Test parts from OSH Cut arrived. This is the inner frame of the base. I will take an edge grinder to the front legs to take off the horizontal tabs that were required for the bends and then I'll stick weld a steel hemisphere over the hole. The display will have a cylinder that swivels on the sphere.

Printing test coupons for finishing tests: surface roughness, color, texture, etc.

I moved the absolute unit of a printer and needed to re-level it. I have learned that big beds needs to be at printing temperature when leveling because the heat causes shifting. Leveling cold can take care of large scale warp but for precision it's level hot or go home. Also, I'm using the Miranda method of bed adhesion (clean with soap and water, then never ever ever ever ever touch it) and so far it's a success.

Based on the metal parts I received from OSH Cut I've made a few changes. The base frame will now peek just a little bit below the level of the surrounding case to provide more durability during rough moves when the weight might fall on the base instead of the rubber feet. The frame won't be visible during regular use. I also lowered the wings on the upper display frame so that it can be bent by OSH Cut's beefy hydraulic press. I may make support parts to re-strongify the display.

Also, I purchased the cutest little stick welding setup so when that's all here I'll attempt to connect the swivel hemisphere to the base frame. I've used higher end MIG and TIG before but these will be my first stick welds so I'm glad that I have a test frame and a few of the steel spheres. Also, an angle grinder. 😸

Today I printed 1:4 scale versions of some of the Alto display case and frame. The floppy part on the upper right is floppy because the full size version is 0.1" thick steel so when scaled down that's not much material. The first coat of primer is drying and eventually I'll give them a nice coat of beige.

Once the six full sized replicas are complete and delivered, maybe I'll make and sell a few miniatures. It's pretty dang cute at quarter scale. One might make a neat collectable or for electronics folks they're probably big enough for some neat projects. A tiny Alto display might make a neat homelab interface, maybe running Mycroft.
mycroft.ai/

The quarter scale Alto display has a working swivel and and has jumped up another cuteness shell. Once it's further primed and painted I may die of squeeeeeeeeee.

My laser cut and bent parts from OSH Cut arrived and I've started printing slices of models from CAD to measure fitment and to test a chonky 1mm nozzle. Also, I rediscovered my fuzzy wrist and head bands which are great for sweat in a shop heated by the CR-10s5.

I received a shipment of LCD panels and driver boards which were thankfully not damaged by whatever poked a hole through three layers of cardboard and a thick layer of foam. I ordered a few extra for spare parts and maybe to extend my ultrawide display to the left and right. (I ❤️ pixels)

I've settled into this layout for on an ultrawide display. I usually open two files: one for parts and one for an assembly. So far it's working well.

Follow

Here's a rough* print of an Alto display base, set on my desk to give you a sense of scale. Tomorrow I plan to grind the bend tabs off of the inner metal frame and weld on the swivel hemisphere so that I can fit the print to the metal.
* no chamfers, sanding, paint, etc

· · Web · 1 · 0 · 2

This morning I started the first full size printed part for an Alto's monitor case. Even with a 1mm nozzle and 0.7mm layer height it's estimated to take ~12 hours. I could probably squeeze a bit more speed out of the CR10s5 but the cycle time is slow for testing the cascading effect of speed changes on large (for FDM) parts. If I had more space then I'd throw money at the problem by bringing in two more of these printers, but here we are.

The correct saw blades and an angle grinder made quick work of the Alto base frame's metal tabs that were left over from the bending operation. This is the first time I've assembled full sized printed parts with their metal frame and friends, it's solid af! The last pic shows the bottom of one of the arms. When secured, the tip of the metal arm will rest against the inside front of the plastic arm.

It hasn't caused a print failure or much delay but twice now the CR-10s5 has incorrectly decided that it was out of filament.

I replaced the filament sensor and it's worked without trouble for ~10 hours. *whew* That eases my anxiety a bit.

This morning I spent an hour in VR with Alan Kay for a design review of my CAD model of the Alto display. We ended up talking about his experiences with HCI and VR. He was a grad student in Ivan Sutherland's group when they worked on the Sword of Damocles so it was interesting to hear Kay's take on the current state of VR from the perspective of someone who was there at the beginning. All in all, a good morning!
en.wikipedia.org/wiki/The_Swor

Sign in to participate in the conversation
Mastodon

Machines.social is a Mastodon instance for people who Trevor Flowers knows and are of the maker, machinist, and/or manufacturer persuasion.