From Mechanical Computing to a Voting Machine - My Residency at Pier 9

... how I navigated the Deadlines, Delays, Distractions, Do's, Don'ts and Divigations of an being an Autodesk Artist in Residence.

This Instructable (if I can call it that) is an outline of how I went from working on a Mechanical Digital Computer to making a Mechanical Questionnaire to assess our relationship with the machines we have become so dependent upon and in which we often see ourselves. I hope to add links to more specifics along the way.

It is also a record of some of the many roads not taken. Maybe I will return to some of the great ideas I abandoned along the way...

There are no overall plans and hardly any measurements for this instructable. This is just the way I work. I start building and then make changes and changes to the changes and keep going until I am happy with what I am doing. If something more interesting crops up I may jump to that. I will elaborate on how and why I am making choices and I do measure as I build but the basic sizes and proportions are just a "seat of the pants" guess. You do have to measure things to get them to work with each other but...

I am assuming a fair degree of woodworking and general fabrication skills on your part although I will throw in tips about specifics of how I did some particular operation. These are idiosyncratic solutions as I have little formal training.

What you will see here is a glimpse into how (and some why) of the artworks I made during my Pier 9 Autodesk Residency. It doesn't include all the CNC training which I didn't use directly but, everything about my residency was encapsulated in my projects.

The Artworks reflect my responses to the experience of being at the pier. Making things is just the way I go about thinking and at Pier 9 I have been thinking about making.

In the April,1988 I saw a Computer Recreations column by AK Dewdney proposing a rope and pulley stone age digital computer -- it was such a great idea that I built my own version for a show in 1992. It took 8 participants tugging on ropes to get the mechanism to operate. I kind of wanted to make a wall mounted indoor version of this .... but I got distracted by many other projects.

You can see more images of my Archaeology of Intention at: http://bernielubell.com/03archaeoinstall.htm

and get a pretty good idea of my low tech approach to contemporary technologies at my web site:

http://bernielubell.com/index.htm

When I arrived at Autodesk's Pier 9 workshop in the fall of 2015 I thought it would be perfect to make a low tech computer at a residency with a company at the forefront of digital design and manufacture! And as a visual pun on the idea of "Distributed Computing" I imagined a distributed network of logic gates on walls around the building or at least in various places in Studio 9. All connected by strings and pulleys.

The AND gate is the most complex logic gate for a mechanical digital computer. An obvious starting point. My first prototypes -- seen in the left 2 images have a great simplicity but they also have a transfer disc that needs some sort of retainer. The next version was borrowed from an image on this web site:

https://mechalogic.wordpress.com/2013/01/26/and-ga...

There seems to be an interest in reversible mechanical digital computing in the world of nano technology. My computer isn't reversible -- it relies on weights and gravity to reset. And this model I like requires springs but I am looking for a good metaphor -- not a potential drug delivery system.

I really liked the action of the linkage that transfers the push rods from input to output ... but ...I wanted to use string and pulleys to interconnect the logic gates so that I could separate them more. I took careful measurements off the linkage (see the last image) and determined to use slotted guides and weights for the components instead of the sticks in guide holes.

Since I am at Pier 9, home of digital fabrication, I thought I would design templates in Illustrator and laser cut them to serve as a guide for routing the slots in all my logic gates.

Coming soon-- an instructable on how I did the Laser cut template

I did discover that the template should have been made in plexiglas the plywood template got a little gummy and I had to keep the slots clean of sawdust to get accurate cuts.

It took many weeks to fabricate all the parts for 2 AND gates, one exclusive OR and one NOT (Inverter). The slotted board was made from Birch multiply and all the sliders string guides and pulley shafts are solid maple. Weights are center drilled from 1" steel rod .

At some point I will publish an instructable with more details on the construction of these components.

Here is a view of the assembled AND gate without the confusion of the slots.

I want the NOT to switch which AND gate will go to an output through the OR. Something about this contingency appeals to me -- it sort of resonates with the sense of not being able to do everything you want. And this sort of constraint is certainly operating at Pier 9 where so much is on offer but time and space have limits.

One interesting aspect of the residency is the obstacles that it throws up AND how they provide you with all the resources you may need to overcome those obstacles. This resonates with so much in my life -- I like to make obstacles for myself to solve -- and It reminds me of a Stravinsky quote that I misremember as -- if I didn't have a system to create against I would have had to invent one. Not in quotes because I found the original passage in "A History of Western Music" by Donald Grout and it is much more elaborate -- and profound.

Anyway, the first image is my quick sketch of how I should hook up my components but clearly I need more sets of pulleys to shift string direction and pulleys to move one string over or under anther as they cross. Basically I need rollers and guides and pulleys to change directions-- I made o bunch of these components as well as little ways to connect the strings to the logic gates and ….

Here is where things started to get interesting. I now have my basic components (In Image order) 2 ANDs (one shown), one NOT and one OR and a host of small interstitial components to connect them. To test this network connected together I need to mount them on a wall. But it turns out that I can't put any holes into any walls on the pier. There are environmental restrictions about how often patching and painting chemicals can be used right over the bay water.

Not only is the testing going to be a problem but my overall idea to make a distributed network is not going to happen.

OK, so I have to make a wall and it needs to be not too large and it needs to be mobile!

Eventually there will be an instructable about this moveable wall.

Anyway, I want my wall to be sturdy but light and it doesn't need to last too long. My "go to" material for this is selected pieces of Hem Fir. It is a fast growing Hemlock that can be incredibly light weight and yet quite strong.

The rails are all made of 2x4's with an 1-1/2" channel running their length. This is for the panel frames and the legs etc… to slip into. The panels are added later.

Anyway, as I am building this wall it is looking more and more like an iconic portable blackboard I remember from grade school. So....

I have to follow through and make the board black. I mask off everything but the panels and look into blackboard paint...

Because of environmental restrictions on the pier this is not so simple. I found a low VOC waterbased paint but our environmental safety person found the VOC's were too high. So I got a zero VOC water based paint from Benjamin Moore, figure that will be OK and use it. Then I discover that I should have had the data sheets checked out first. A week later, it turns out that the paint is OK. Phew!

I went for the Blackboard big time. I made chalk trays and an eraser out of the Hem Fir and a sponge.

You can see how the pieces of the frame fit into the uprights here. The chalkboard tray was cut on the table saw with the blade ~15 degrees and the 2x4 on edge then flipped. Test cuts first to get the cuts to meet. The last pic is one of the tray supports which slips into the side rail.

Meanwhile, I was not the only artist who needed a wall to test things. Chris opted to make pedestals rather than a wall and when I told him that my wall had become a blackboard he had a dream...

I made Chris's dream operational and wrote out the lesson I had learned that "I will not put a hole in the wall" morphing into "I will not put any hole in any wall" I haven't written in cursive in many decades.

But now I love my blackboard so much I don't want to put any holes into it either.

but I do need to test out the computer. I made an arrangement of the parts on a couple of work tables, made a white board (1/2" ply ~ 4' x 5') to lay over and protect the blackboard on one side. I used a chalkline to do the layout on my "white board" following my dimensions and then mounted the computer parts, ran the string and tied all the components together. And, to my amazement, it worked pretty well the fist time. A few hours of juggling weights to get the components in balance and...

Here is a video of the operation with a pretty inaudible soundtrack. Better video to come.

This seems pretty complete but lack the mystery of my original idea of a widely distributed network Since it won't spread out over a larger area but be confined to this one mobile board I thought I could use both sides of the board and… Mei & I were talking in the shop and the idea of using the space inside the board, between the two panels, came up -- maybe even having something pop out of the invisible space between. A little mystery or surprise. Since the board was already constructed I removed the top piece, cut some nice neat slots into it and then used it as a guide to brutally hack some holes into the top and bottom of the frame. I vacuumed from the bottom with compressed air pumping from the top and sprayed the inside of the board so it would be sealed.

Then I took a break for 2 weeks -- which is a whole story in itself.

When I got back from my break I wanted to do a survey with everyone at the Pier looking at how they felt about their relationship to machines. Although I love machines, I am a low tech guy. As one of our reviewers put it -- I am interested in the locus of control. Just how much control do we cede to our machinery.

I have no idea how to show the development of the questionnaire. My wife suggested a picture of us talking or of me thinking and there was a brainstorming session with Vanessa our AIR traffic Controller.

Ultimately I decided to open my questionnaire with a quote from Saul Bellow's "Adventures of Augie March" :

“…probably she was waiting to see what I would do. My real desire was to get out. But the car had
already gone a way over the cobbles and having just got it under way I couldn’t check it. That’s so often what it is with machinery: be somewhat in doubt and it carries the decision.”

and I asked:

Does this quotation correspond to your experience with machines?

Who has made the decision for the driver not to stop?

If you are some sort of machine, What sort of machine would you be? and other questions that included some demographics like age gender zip code and whether you went back to change any answers. And a whole other questionnaire that I haven't looked at the results from yet.

A preliminary analysis of some of the responses showed a few interesting things especially sorted by Age, Gender and whether you went back to change answers. More results will be forthcoming but I am not sure where to publish them.

The median age of the respondents was ~32 and the younger group were less inclined to think that the machine decides. Or perhaps they are just less in doubt. Also the younger group thought the unconscious or the car made the decision. The older group felt it was the conscious mind. Females thought the car was the most likely culprit anbd were also much less likely to think that the machine decides.

I'm not sure what to make of this -- or how to automate the data analysis. My sample was small but the counting by category was done by hand -- which was tedious.

But the most common answer to my first two questions was 3 -- which translates as "sometimes" and reveals nothing. It flattened all my results -- I really wanted to force people to choose but by leaving a midpoint I lost that option. Hmnn...

Meanwhile I decided I was being too precious with my BlackBoard -- I should mount my computer right onto the board except that the components were too big since they were made to spread around a room not fit crammed together on a 4' x 5' wall. So, I have remade everything with smaller with black slotted boards -- maybe I have improved the mechanisms a bit, maybe made it worse -- we will see.

One improvement was to put the strings through holes in the sliders so that they are much closer to the frame. This should reduce the moment arm when activated and might result in smoother action.

Probably need a separate instructable for all the details of making these as well.

This time, instead of slots in a solid ply board I built up a slotted frame out of modular maple components, glued together and painted black. I relieved the bottoms of these frames to reduce the apparent weight and to make them standoff from the black board. My intention is to h=ave the frames disappear and you would just see the wood components.

It was also time to think of how people might input a string pull for the computer. I

thought about the stools we have had to sit on through various meetings and classes. They are memorable because they are not very comfortable. Memory is a possible output for the computer -- but more on that later.

I thought that leaning left or right would either input the NOT causing a switching of ANDs or...

...but the stool was incredibly unstable -- left/right and front/back and every which way and it seemed to be breaking before my eyes. Lots of problems to solve here. Had to make a jig to enlarge the flat bottomed holes, recenter the threaded inserts and use bigger ones, widen the pivots and use threaded inserts to hold them to the base and to the seat (screws were just not strong enough -- one of the many problems with the original design.

I put the inserts in from the back of the boards so they worked more like nuts and bolts. Only problem was if the inserts weren't tight they might unscrew as I tried to torque the machine screws. Solved with a bit of superglue get on the inserts.

The resulting stool was still unstable. Just in one plane this time but hard to hold any position except all the way to one side or the other.

The pictures may not do it justice but it was impossible to control your position on the stool. The springs just weren't strong enough and being partially compressed they both pushed and pulled you off the mid point. Maybe it is a perfect abs workout because with enough muscle control over your center of gravity you just might manage to stay centered.

So I quick turned and painted some pull handles -- just in case the stool didn't work out. I use my metal lathe for wood turning.

1" maple blanks into my 3 jaw chuck tapping them with a rubber mallet until they have no wobble. Then I center drill and drill a 3/16" hole and repeat on the other side of the dowel. Then I can turn it between the chuck and a dead center, sand and finish.

The red paint wouldn't dry and I was away from my lathe so I rigged up a1/8"" arbor which matches my hole and put a 1/8"" hole in a plate clamped to the drill press table as a dead center and sanded off the offending paint. And I am glad I did because...

I was chatting with Andrew by the OMAX and telling him that my stool was too unpredictable -- really it was just like I had wanted my questionnaire to be because it had no middle/sometimes position. Then It struck me that I needed to pursue this idea and somehow make this a questionnaire machine. What was especially delicious here was that the stool would shift so quickly -- it manifested Bellow's line; "Be somewhat in doubt and it (the machine) decides."

The middle is missing because the seat swivels on it's base and the weak springs will center the seat without a load but with a load the center is not stable. Only the two extremes provide stability. There may be a political metaphor in here a well.

So I have a mechanical metaphor for the decision -- now I need a way to count the decisions -- I need memory. The ratchet building should be it's own instructable but I don't know when I will get to that.

The ratchet gear is laid out traditionally. A circle is inscribed with the dividers on the disc cut with a 4" hole saw from birch multiply. Then the divider is used to step off increments along that circle and adjusted until I came back exactly to the starting point. Repeat the process pressing the points firmly into the disc. Teeth are then cut on the band saw. You can see the layout in the second picture.

I only need two but wanted an extra. The ratchets have to turn a drum to wind a string up a scale.

The first picture shows redrilling the 1/4" hole to 3/4". I do this with a short 1/4" dowel that I center drill and then use a forester bit to enlarge the hole cleanly. The second pic shows pining the gear to the shaft with a 1/8" copper plated welding rod. The third pic shows nylon washers hot glued to the drum which is glued up square with a 3/4" dowel as a center spacer, cut to a hexagon profile on the table saw and then turned round on the metal lathe. 4th pic is component mounting boards. The bearing holes are reamed oversize with with an adjustable reamer. Holes and shaft are waxed. 5th pic - drum pinned. And then some shots of the assembly

the shape of the ratchet teeth requires an extreme tapered curve for the Pawls. shaped and tested and shaped and treated until it is right. Note the transfer punch to get the holes in the same spot on all 3.

The ratchets need to register their responses onto the scales which will be connected to the ratchet holders in the dados you can see in photos 1-3. The base will hold all these components and the stool. The gray color is to match the industrial stools. The 2x4 triangles are to stiffen the connection of the base to the uprights.

The no trip slope on the base is made with a 10 degree setting on the table saw running the ~30" square plywood piece through vertically.

I used 12' finger-painted primed pine to which I added marks every 5 cm as my scales. The indicators are painted maple pieces that are held to the strings with 10-24 nylon screws (threads in the maple pieces) The 1/8" braided dacron line is wrapped once around the cylinder and routed through a set of guides going over a HDPE pulley I made for the top. On the return leg there is a spring to maintain friction on the drum. There is an additional guide string -- very fine twisted nylon line running up the center of the pointers to keep them pointing towards the scale.

Here is the arbor I used. It is a 3/8" threaded rod with 2 nuts as a stop on one end and center drilled at the other to go into the live center.

Cut pulley blanks out of HDPE. I could fit 3 on the arbor. Turned it on the metal lathe with a standard 60 degree replaceable carbide bit.

I mounted the ratchets before I realized that they had to be mirror images of each other, duh. I had to rebuild them to make this happen. and then discovered that I had the drums turning the wrong way so I had to swap them and the scales and the alignment wasn't symmetrical so I had to enlarge the hole for the ratchet shaft to accommodate it. looks the same but now the drums turn the right way.

Indicators were made from maple blocks drilled lengthwise for the string and cross grain in 2 spots to tap for the

10-24 nylon machine screws. Then a dad is ripped to create a thinner area for the point edge. Then cut for the point on the bandsaw. Sanded painted and installed on the string.

There was also a fine string to running through a hole in the pointer

The first 2 images show the tracking guides for the indicator string. I angled the holes so that they are in line with the pulleys at the top of the scale at the top of this pine guide plate and the holes are offset at the bottom of the plate to keep the string from bunching up on the drum. This didn't work Next time I need to probably turn a crown on the drum. Still seems to work. the 3rd image shows the string on the drum I decided on one wrap -- makes resetting easier.

The next 2 pics show the attachment for the guides for the vine guide string to keep the indicators pointing at the scale.

Next pic is one of the 2 stops needed for the ratchet arm to get just a single click per movement.

Th last pic shows both stops and the stop for the activation string which goes between the transfer arms to the seat.

Some shots of laying out the guide lines for the computer parts to go onto the blackboard directly

The black slotted components are added using finish (small head) screws.

The pulley mounts are all installed separately to keep the spare look. Behind each cutting board pulley are nylon washers as spacers and on top is another washer and toothpick as retainer. The guides help keep the string from popping out of the pulleys.

Finally painted handles, chalk text and stops to prevent overdriving the components.

Instructions were laser etched into a glued up 17" square piece of pine. An outline guide at 16" dis was cut as well. the edges of that were measured from the board edges and transferred to the back to determine the circle center. A shallow hole was drilled at the center in the back for the router circle cutting jig.

Additional words were added with metal stamps colored with acrylic paint to match the laser etching.

The text on the seat are the instructions for use. Harking back to the questionnaire:

"Be in Doubt and the Machine Decides?

Lean to Answer

"Always" to your Left "Never" to your Right

Transfer bars are made from maple, notched for the pivot extensions and attached to the seat pivot extensions with screws. Holes were drilled and countersunk for the string which attaches to the ratchet bar which is similarly drilled. There is a small maple block with tapped for a 10-24 nylon screw that acts as the lifting stop.

I tried doublestick tape but people were leaning so far that it was still popping out the 7/8" hole in the base. I made a couple of notched maple hold owns that screw to the ply base.

Some images of me testing the installation and a short video. Opening is tomorrow night. I'll count the votes before each reset.