Ex. 07 “Yes, we’ve heard of Bruce Shapiro” (Analog Machine)

This page is dedicated to analog machine-building. To re-iterate, we are very familiar with Bruce Shapiro’s Sisyphus project. So no need to refer us to him, thank you.

Our research for the analog machine is focusing primarily on pantographs (translational drawing machines). Materially, we opted for a bed of sand, with a magnetized ball as a drawing utensil (i.e. Bruce Shapiro), and later a stylus from above. After finding a prototype off of which we could base our initial scheme, we began sketching/fabricating;


1wt-01 b1-wt c2-wt cc1-wt g1_wt
2wt-01 b2-wt c1-wt cc2_wt g2_wt
3wt-01 b3-wt c3-wt cc3-wt g3-wt
4wt-01 b4-wt c4-wt cc4-wt g4-wt
5wt-01 b5-wt c5-wt cc5-wt g5-wt
6wt b6-wt c6-wt cs1-wt sg1-wt
7wt-01 b7-wt c7-wt cs2-wt sg2-wt
8wt-01 b8-wt c8-wt cs3-wt sg3_wt
9wt b9-wt c9-wt cs4-wt sg4-wt
10wt-01 b10-wt c10-wt cs5-wt sg5-wt





In order to update from under the drawing surface to above, and to change the logic from cam tracing to tool-pathing, we needed to scale up. Our first attempt was made simply by scaling up the workspace by 4. img_8501
Although the machine worked (loosely), it was difficult to manage with only one person, and took up way too much space for the tiny aspect ratio we were achieving in the final drawing. The machine was huge, but it didn’t even fill the original black work space
Here you can see how very small the original drawings were. This is the maze drawings you see above, but with a super bizarre distortion and tiny scale. maze2
So we updated. Andrew made us a grasshopper definition that makes a perfect 2-1 Pantograph workspace, so that all dimensions can be grabbed before making.
From there, we shrunk our workspace while maximizing our aspect ratio. This is our final setup used to produce the above drawings. img_8551
We had major success by dragging a stylus through the sand from above. At first we attempted the maze on an even bed of sand
We decided to shift our pursuit after realizing the exciting potential of moving a mound around.
In elevation you can see how the arm originally interfered with the sand, a potential pursuit in the future. img_8553
Here is a simple elevation of one of the final mounds with the labyrinth drawn through it. It is interesting to think about how the machine might be able to carve pathways up or down the mountain. img_8559




This is the setup for a drawing with the casting box and modified raised bed.

the height of the bounding box is dependent on how much sand you place

After making the drawing, secure the sand particles with a PolyCrylic spray. Be sure to mist the sand, and not spray directly on, or else you may disturb the surface quality of the sand img_8606
You should see a thin white veil on the edges when you have sprayed enough. img_8611
It is important to not waste the alginate (it’s a little expensive) so to keep your mold tight, we can take up space with blocks of foam. img_8612-1
In order to seal the edges, use plastecene clay (or any other oil based equivalent) img_8616
Now on to the mold making! img_8620
Be sure to follow the instructions on the package, and that your alginate is liquid enough to capture the detail of the sand. img_0396
This is the right consistency and pour img_8623
After it hardened (3 minutes) We flipped it over to remove the sand. Here you can see the distinction between the soft sand, and the thin shell which was sprayed. All of this will have to be excavated.  img_8625
We started with a brush in order to keep the nice soft sand img_8629
Now the acrylic shell needs to be removed img_8628
By spraying it with a low pressure (50psi) airgun, you can remove the thin shell. img_8630
This is how the final mold came out. img_8631
And the final cast.


Minor success. In real life the hydrocal really managed to grab the detail of the sand. However, it seems to lose some of its magic by becoming solid.





Interested in moving sand around and input, we began playing, while also capturing our resultant toolpath. In this drawing, you can see the “Jog Path” in red, and the “Push Path” in black. costline2 image5
Here we tried to modify an island a bit, and have fun watching what happens islandsculpting image7
This was an attempt at tracing a “coded” toolpath, as opposed to making it on the fly. maze2 img_8506
We pushed a mound around conebase conetweez
Tried to make a bridge between two islands twodune bridge
Or a new coastline. tweez aftertweez_2
And here are some process videos on how these drawings are made.

We also tried pushing sand around with a little UFO shaped thing. It can be flipped so it also looks like a cymbal. The pushing idea is manifested above as well using an L bracket, but there it is more “pulling” the bracket behind the ball, as opposed to trying to actively displace sand. It was rather difficult, as can be seen in the video below.



Sketch of the machine layout

and laser-cut acrylic arms

Go Here for the exact dimensions we used to make the pentograph https://www.peter.com.au/articles/pantograph.html

         image2 img_8168
Milled drawing bed -buffed with an orbital sander to remove tool-pathing trace image1 img_8172
Arms feature variable attachment positions in order to test multiple scales of translation (we keep ours at a 1-4 ratio). Make sure you assemble using this hierarchy. img_8169 img_8170
Arms attached to the base board with M4 bolts, washers, and nuts. Additionally, you can see how the magnetic tip is simply two small magnet balls whose strength keeps them together. img_8176 img_8180
Tested the arm to determine an area of the base board to keep the cursor within, and the machine is ready for testing!! image3 img_8179
Adding a light dusting of sand to the bed. We added a tracing template for people to hone their skills!! image4 image5


There are some specific things to consider when bolting things together. The use of Washers for free movement, lock washers for tensioning bolts, and lock bolts when you have them are all essential to make sure a  constantly moving machine doesn’t rattle itself to death. See below

Here you see two bolts pinching a lock washer, allowing for just enough space for the acrylic pieces to move without pushing the bolt around.
And here you can see how the bolt is secured to the mdf, lifts the acryllic high enough to move with the stacks, and also secures the bolts in their place so the motion of the pieces doesn't cause their motion
And here you can see how the bolt is secured to the mdf, lifts the acryllic high enough to move with the stacks, and also secures the bolts in their place so the motion of the pieces doesn’t cause their motion.



Our joint flex test, notice the bolt not moving due to proper washer, lock washer, and bolt use.
Our hello world
An example of the muscle groups necessary to move the cursor with accuracy. IT’S A FULL BODY WORKOUT



Talking with Che Wei and Duks on thursday, we’ve made some changes. For instance, our handle was previously a sliced up Poland Springs water bottle fastened to the hinge with an M3 screw. Now, we have a 3d-printed piece that assumes a similar form and does the exact same thing. Secondly, and this is the important one, we’ve upgraded from freehand drawing to laser-cut guide rails that we can now use to create much more specific linework in the sand:


First we need to make sure the cam frame is centered on the cam gear. img_8241
Using the alignment cam (it will become scrap afterwards so no need to use fancy material), center the frame on the gear. Use a bolt to tighten the cam and gear together. img_8242
Dab a little bit of acrylic glue or super glue onto an applicator stick img_8244
Apply lightly, to the edges. We don’t want the cam to get stuck in there as well! img_8245
Do the same for the space underneath. img_8248
And then the same for the transitional gears img_8251
And voila! The holes are anywhere too specific in the base board, simply drilled one at a time, making sure to bolt the gear in place before setting up the next drill so nothing moves. img_8257


Our updated handle image2-1
Translational gears. The rotation of the gear at the end of the chain 360° rotates the gear upon which the guide-rail sits 20° (UPDATE: ONE OF THE TEETH BROKE. STEER CLEAR OF PLASTIC GEARS)

By doing this, the same toolpath can be executed around variable origin points on the bed (You’ll see what I mean in a second)

The first pass using the centered square cam shown in place above  image1
 Some of the cams we made after a successful first pass with the square. image3-1
This is a drawing made with the offset square shown above. img_8263

This is a video of the broken plastic gear, in case you were wondering.