Final Project: Assembly and touchups

It took a lot of time to cut all the aluminum pieces (16 pieces X around 2 hours for each piece). Fortunately the wood was simpler. I used the bitbreaker mode with Mahogany wood settings from otherplan website (0.05″ depth instead of 0.003″) which let me speed it up at least 15 times faster. The rpm was down from 15k to 10k, it made the operation kind of less noisy.


The only problem was sticking the wood to the spoilboard- I had to be extremely careful with wood and spoilboard surfaces before mounting it on with a two-sided tape.

I engraved 0.145″ outline rectangle on the tiles, it allowed me to have precise and identical shapes. Although they were all rectangles, it is practically very difficult to device a system that cuts exact rectangles. Chop saw, thread saw, band saw etc. tend to develop some error over time, and it’s difficult to get 90 degrees cuts. The outline engraving was a nice trick ; I sanded the tiles till those outline marks. The final tiles arrangement looked like this:


The aluminum pieces snapped perfectly in the corresponding places, there was 0.006″ tolerance:


When a few other students used the othermill, I took some brass and aluminum and turned it into this:

IMG_7249 2

Basically I made a hole in brass, measured it with calipers, turned aluminum to a diameter little larger than the hole, and pressed the two together in the arbor press. Then I turned the composite cylinder on lathe again, faced it till it reached the point where the two metals had no gap between them, and cut it using a parting tool. Next I quickly made an acrylic stencil on lasercutter, fixed the metal piece in it and put it on othermill bed. The stencil provided a great reference in aligning the drawing.



Ultimately it became a compass for the map-


Next I aligned all the tiles, tested the alignment with a few aluminum pieces, and glued all tiles together to a flat 12″ X 15″ ply:


I left it to dry overnight. Next morning I fitted all the metal pieces into engraved cavities and masked them with tape. I applied minwax dark walnut finish to the oak tiles, and then made a frame around it using the same oak. Later I applied a coat of polyurethane on everything (frame + tiles), with the masking still on. I still had to use some turpentine to remove a few walnut finish stains from some metal pieces, where the masking didn’t do a great job.


And here’s the final map!


Finals: Exploiting the Othermill!

I’m going to make a 12″x15″, *non-interactive* map of NYC using wood and metals. I spent a day visualizing the map and creating SVG from it.

Screen Shot 2016-04-28 at 10.36.10 AM


I’m planning to make wooden tiles with embedded metal pieces. Wood is land and metal is water. Following factors had to be taken care of to avoid unpleasant surprises at next levels:

  1. Size of milling bit: The bit should be able to trace outlines of both the positive and negative shapes. The outlines had to be rounded a little.
  2. Size of the Othermill bed: This determined the size of individual tiles.
  3. Materials: I wanted to use aluminum and walnut, so I tested a simple friction fit:

IMG_7204 2

I’m really happy with the results: Both materials are sanded down to same surface and it feels elegant, smooth and rich. Unfortunately I couldn’t find walnut after hunting a few places across Manhattan and Brooklyn, and online delivery would have taken at least 5 days. So I’m going ahead with oak. I’ll stain it to add some contrast with metal.

Next I carefully made two sets of SVG files: One for oak and other for aluminum, taking into account the tolerance values and tile alignment. This project is taking more time than what I had thought earlier, but it’s extremely enjoyable and I’m learning new facets of the Othermill.


I also learned about a lot of new uses of dremel: there’s a cutting wheel that I use to cut off the pieces from metal slabs, when othermill leaves a thin aluminum film on the contours.

4 Axis Mill: Job #2

I tried 2-sided cutting on 4 axis mill this time, with a block of mahogany wood.





I didn’t have a round bit at time, so a spherical concave surface is cut into steps by the flat bit. I had to sand it manually later.

Once the roughing pass finished cutting the surface and notches, I stopped the operation as I realised that it’s just cutting tabs. Tabs were simple to take care of manually on the sander belt.



4 axis mill


I am enjoying 3D CAD/CAM as I’m learning more and more tools from this class. I used Vectorworks to create a helix. SRP Player and VPanel were then used to enter dimensions, material details and bit details. I got a piece of delrin from Ben, I used metal  lathe to clean it up by facing the two ends. Facing also marked the center of the surface, and I used a little nail to make a dimple so that it fits on the tail of the mill.


The machine has many possible operations to offer, so the interface is somewhat complex. I had to be careful while setting x,y and z origins.


220 minutes! I opened the fall 2016 registration page and took a good look at the next semester courses 🙂 Vacuuming was necessary at a few steps as there was a lot of material being removed. I used the small vacuum cleaner from 3D printers space.


The profile looked really beautiful as it was being machined. Because of the small size of job and bit (1/8 inch), the changes were not immediately visible after each pass. The final piece looks like this:

thumb_IMG_7088_1024 thumb_IMG_7089_1024

I canceled the finishing pass before starting the job, and it still took over 3 hours for roughing. Roughing produced facets on the surface which look nice for this piece, but it will take more time to finish the surfaces for utilitarian parts that require precision.

Metal Lathe

While ITP schedule gets crazier each week, subtraction class & assignments never fail to soothe! Using metal lathe was an enjoyable, satisfying and rewarding experience. Never before in the shop I was so much focused, geared up, and happier.

For the first assignment with basic lathe operations, I used a piece of 1″ aluminum rod and also tried delrin rod. I started with a piece of aluminum Immanuel was using before he finished his assignment. It was partially turned and faced and I learned a few tips from him before I started.


I faced and turned the job, which left a sharp edge around the faced circle. I used slant edge of the facing tool to make it a little blunt. Next I used the parting tool, mainly to see how it feels and how different it is from the facing tool. I made a little notch around the turned surface using parting tool:


Using a lot of oil and precaution, I tried using the turning tool to mark a rather deeper notch on the surface. Although it felt really smooth while operating, the edge has radial marks on it, like that on a knob/ wristwatch crown. This was an unexpected result and I like the effect.


Compared to aluminum, facing and turning delrin was less fun and I didn’t spend much time on it. Here are the two pieces:


Next, I used the aluminum piece to make the doo-dad. I mounted it on the chuck, used the #3 center bit first:


I struggled for a little while with the drill chuck as I wasn’t initially able to install it on the tailstock. I saw a few youtube videos and made sure it is done right before actually drilling the hole. I used 3/32, 3/16 and finally 7/32 bits to gradually drill the hole larger in diameter. Next I stopped the machine completely by pressing the red STOP button and used the tap set. Initially I tried mounting tap on the tailstock and used the wheel on right hand end of tailstock to drive the tap into the job; but somehow the tap always refused to come out. It went in when pushed into the job, but stayed inside when I turned the tailstock wheel back.


Finally I used the key and did it without using the tailstock. I thought of tapping as an operation that would require a good amount of efforts; but it was really smooth. Like Ben had rightly mentioned, this particular aluminum ‘wants to be machined’.


Next I installed the piece with the finished face inside the chuck. Parting needed more attention as there was a lot of material being removed continuously. I used lot of oil:


I used vernier calipers earlier to measure the depth of hole drilled into the piece, and therefore the hole was visible from the other side:


Next I did a little facing and tapping from this face as well. The final machined piece looks like this:



Wood joinery was an interesting assignment that made me think of the CAD and CAM process to the last detail, and I decided to make an object that has good amount of joinery. I made a crate (14″ x 10″ x 8″) from plywood.

The drawing was simple, but I had to be careful and make sure that the notches will align. The most important part was the corners where three surfaces meet together- calculations involved thickness of the material, tolerance, and length of the notches. I designed the joins so that the opposite walls are identical pieces, like the ones on left and right with handles, as shown in the drawing below. I used mirror images so that top plywood surface texture and color would show on inside of the box, and bottom surface of the plywood becomes the outside.


IMG_6927 IMG_6928

The tolerance worked perfect: the notch length is 1.98″ and corresponding indentation is 2.02″. That is NOT good enough gap for two pieces to join together without glue; but with 5 pieces and 8 edges it just perfectly worked. Next I cleaned the cutouts with sandpaper, applied glue and clamped the pieces together for a day.

IMG_6931 IMG_6930

With glue, it got a little more tighter and fitting in the last piece was a little difficult, but I didn’t have to file or sand down any surface in order to join it together. A minor error was the top edge engraving on 4 walls that creates a little 0.25″ step for a lid: On longer walls I forgot to engrave the extra 0.25″ on each end to compensate for shorter wall engraving, and it looked like this:


I had to use chisel and hammer to subtract a little 0.25 x 0.25″ square.

Next I used sandpapers to clean the edges, and handheld sander tool to clean the surface. This tool was effective in almost completely removing the stains where the glue had popped out of he joints, forming a bulge after drying:


Following is the before and after image with the stains removed.


Wood Joinery with Techno CNC

Last time I was lucky enough, and/or the CAD/CAM job was really neatly done. Outcome: the wooden container came out to be a beautiful piece with finished details.

This time the techno CNC sorcery won, I could not make a seamless wooden joint after a several attempts. As always I worked on the assignment on weekend when the shop is usually empty. I did some sketching first with paper and pen, looked at the reference examples on Subtraction class page, and then made vectorworks drawings:


I tried Oval-shoulder halving and Tenon halving with ellipse and rounded square shaped window. For the oval shoulder I tried two different drawings: One with zero tolerance and other with a 0.01″. The pieces with zero tolerance needed a lot of sanding, and the ones with 0.01″ tolerance were somewhat simpler to assemble.

IMG_6910 2 IMG_6911 2 IMG_6915 2 IMG_6917 2

It required considerable efforts on sanding the surface areas and hammering the pieces to fit them together, but this could have been a lot better and easier with more tolerance.

I ran into and learned a lot from two issues-
1. Stupid mistake I did in CAM: I made a significant mistake in the parts that had oval and rounded square windows. In setting contour and engraving orders, I had put the engraving first, then outside contour for the whole piece, and then the contour for window inside the piece. As a result the piece was not stable when the oval window was being cut. I had to pause the router and fix the piece into the spoilboard with two screws so that it wouldn’t wobble. The oval window got completely spoiled though, but I was able to manage it in time for the square window.
2. Bad quality of the material: The ply had an uneven, kind of burnt layer underneath the oval window piece. The layer got chipped off as engraving reached a certain depth.

IMG_6914 2

Techno Router: Making a wooden container on CNC

Of all the machines we’ve learned so far in Subtraction class, techno router is the one I enjoyed the most. It felt like using a scaled up Othermill router where bed size has been enlarged considerably from 5.5 X 4.5 inches to 8 X 4 feet. Also this assignment involved the most hacks and impromptu workarounds than other machines.

I thought of complex ideas initially, such as making a simple chair or stool with a few joints, but soon dropped that idea as making drawings in vectorworks for the first time was really tricky. I then decided to make a cylindrical wooden container with a lid, and made a few sketches:

IMG_6881 IMG_6882

As shown in the drawings, I wanted to make stackable ring layers that would form the walls of container, with interlocking protrusions and cavities. However I dropped that idea as the size and shape of the mill bit was another complex parameter in designing the joints. Finally I made digital drawings of flat rings that could be stacked and glued on top of each other to form a cylinder.

For some reason the MasterCAM always threw an error when I tried to apply contour operation to more than one objects, “The selected sub-chain does not touch the branch point.” I googled the error and most of the blogs mentioned some fixes in vectorworks which I did not understand. As a solution I tried making a similar drawing with .dwg extension in Adobe Illustrator, and it fixed the error. I was able to apply all the contour and pocket operations. (Using illustrator triggered another problem later!)


Three circles from the column on the left are the bottom (solid circle) and two lids (I planned to make two, just in case), with some engraving on the rim so that it would snap on the opening of cylinder. I set three operations:

  1. Cut the inside circles from the rings (Contour, counter-clockwise)
  2. Engrave rims of the lids (Pockets)
  3. Cut all circles and rings from outside (Contour, clockwise).

Next I prepared the techno router with a plank fixed to the bed. I required a 2’6″ x 1’6″, but found a slightly bigger sheet of plywood.


Next I set the origin and preprocessed the GCode file,

IMG_6865 IMG_6862

…And then something weird happened on hitting fire. The router head further advanced in -Y direction and started cutting a circle:


I immediately hit the pause button, thanks to Ben’s instructions. The reason was I was working in bottom-right quadrant in the MasterCAM software, and I had completely forgotten to take care of it when I shifted from Vectorworks to Illustrator! This is evident from the blue MasterCAM image shown above. I thought of possible solutions andcould think of two:

  1. Go back to MasterCAM, move the drawing to top-right quadrant, export GCode and start all over again.
  2. Put the sheet in bottom-right quadrant. That is, set the Zero a couple feet away in +Y direction, and start the operation from beginning. The drawing would still be cut in bottom-right quadrant.

I decided to go ahead with the second option as it was quicker. I stopped the current file, moved the head approx 3 feet in +Y direction, and selected Zero All. This worked perfectly well, with drawing set in +X and -Y directions. Next part involved patience, vacuuming, and monitoring the router operation:

IMG_6871 IMG_6876

The ply had a few underneath cavities and therefore a few rings were not perfect. Good I cut a few extra:


Next I applied glue and stacked the rings one by one on top of the bottom piece, and applied clamps to hold everything together for about an hour:



The cylinder looked beautiful and I sanded it evenly on the belt sander tool in the shop. This part involved careful handling as the cylinder was a bit slippery on the sander, because it was rolling like a wheel as soon as it touched the sander belt. The next trick was sanding it from inside. I used one of the circular pieces cut out from within the rings, glued sandpaper to its rim, and mounted it on drill machine. This custom-sander was really effective in finishing the inside surface of cylinder:



I used the same tool to sand the lid. The sanding made the lid a little smaller in dimension, and allowed it to snap perfectly on top of the cylinder:


Here’s how the final product looks:



The Othermill

Operating Othermill was a satisfying experience on many levels: simplicity of input file (just a flat SVG works!), user friendly design of the software, aesthetics of the machine, and finally the quality of the finished job. The 1/8″ flat mill bit ordered from Amazon seemed to be a quality product. I spent some time observing my classmate when they worked on their projects and that helped a lot while setting up my job.

The setup was much simpler and quicker than what I had anticipated. Steps were easy to follow and the only thing that went wrong was setting the Z axis. The drill bit descends gradually till it ‘touches’ (makes electric contact with) the metal bed, but a small piece of double-sided tape acted as an insulator and the bit kept descending further, making a harsh noise. Thankfully I managed to stop it as I was all set with the cursor hovering on the Cancel button.

Ben handed us all Derlin plastic sheets. The material has an appealing sculptural quality. The white material is translucent and I decided to make something that looks beautiful with lighting. I found this piece of pipe in the junk, and made drawings for Derlin caps that would fit on the opening of pipe.


It is great to see the software and the actual machine in action, sitting next to each other. The software displays the progress and simulation in real time.


Othermill is pretty precise when it comes to milling small and intricate parts:


I selected engraving for half the thickness of material. The final engraving and cutout looks like this:


The circles snap on the piece of pipe, and it looks wonderful with LED illuminating the pipe from inside.



Assignment 1: Router

I worked on the assignment over the weekend, and it was fun to operate the router. I am comfortable with the workshop tools by now, still I was really excited about the router as I had never used it before.

The tools I used were a ruler, set squares, pencils and pens, tape, the router (with router bit, default round jig, circle cutting jig, jig pin), drill machine (with drill and screw driver bit), sander belt, and of course the vacuum cleaner. I started with fixing the wooden plank on the table and drawing on it.


Next I put on the circle cutting jig on the router, and figured out the settings that allow manipulating the depth of drill bit. Using drill machine with a screw driver bit was a quick way to remove the default round jig and replace it with the circle cutting jig.


Initially I had only a couple of millimeters of router bit popping out of the jig level, and I realized after the first run that the router is indeed much powerful. I then changed the depth of router bit to around a centimeter.

The jig pin didn’t fit in snugly in the center of the piece, so I applied some blue tape to make it thicker.


After carefully doing around four passes, the job looked like this:


The next part was a bit tricky. I removed the jig pin, mounted on the smaller jig, and used the edge of circle cutting jig as a guide to cut the straight line in the piece.


Lesson learned: After affixing the guide on the wooden plank, I should double check if it is still parallel to the line that I want to cut. The guide was not exactly parallel after taking in all the errors that arose when I drilled the screws to affix it to the plank below. The result was a slightly deviated (and disappointing) cut:


After fixing the error, I was able to get the line cut along exact course. Next I sanded it on the machine, and it improved appearance of the piece a lot.

60 70

Here’s the final piece:

80 90

Lessons learned:

1. The corners were tricky. To generalize it better, the ends are tricky while using a machine- the start and stop points need to be carefully thought of. I started cutting the arc from exact intersection of lines in the line drawing, but I stopped where the router bit touched the drawing line on the other end. So the error introduced is equal to the radius of the router bit.

2. Spending more time and efforts on the initial setup doesn’t cost, it indeed pays off.

3. Think of the number of hours it will take and multiply it by 3! The error handling, rectification, and most importantly post-completion cleanup takes considerable time. In summary it was a great experience to start the new semester with learning new tools.