Week 6: Computer-Controlled Machining

Project Plan

I was sitting in my room and couldn't seem to find a project for the week that could be used in relation to my final project. My thoughts started to wander and I was looking around my room. I saw my feet resting on our coffee table and then considered the coffee table itself. I soon convinced myself that I could design and make a better one. The project for the week: a new coffee table for my bedroom.

The Training

I went to the training for the Onsrud 3-axis router in the Architecture department's wood shop in building N51, hosted by Chris Dewart. During the training, we discussed shop policies and safety procedures. We then began working with Mastercam and the Onsrud. Chris didn't elaborate much on how to work within Mastercam. He told us that once we sent him the DXF files of our designs, he would handle the CAM setup and then we could cut out our parts. During the training, we cut out the test parts to see the fit of the plywood that we would be using for our project. The Onsrud has a built-in tool changer, and so Chris added the tool changes to the CAM program before he sent them to the machine. The standard order would be to drill out the corners of any dogbones with a 1/8-inch drill bit. Then, use a 1/4-inch downcut end mill to cut out all of the bodies, leaving an onion skin to hold the parts in place. Finally, we'd use a 1/8-inch downcut end mill to clean up the fillets left by the larger end mill in the corners of the dogbones, where the relief holes were drilled.

The Onsrud has a powerful vacuum table that sucks the stock down to the cut bed, keeping the parts flat. It is important to use screws to mount the stock, as during the cutting process, the vacuum decreases as more holes are added to the stock.

The bodies have been cut out, but they aren't free from the stock because of the onion skin that was left behind in order to hold the parts down during the rest of machining.

Chris also then showed us how to use the router table to clean up the edges of our parts, which improved the appearance and quality drastically.

During the training, Chris talked about the importance of constantly measuring the stock, as it can change based on the weather. I.e., if it rains, the stock will absorb water and be wider than it otherwise would be. When we tested fits with the components we had just cut, we determined that the best fit was the 0.42" gap. We measured the stock to be 0.433".

Exploration

I wasn't quite sure what I wanted the new coffee table to look like. In the past, I have been interested in tensegrity tables, even making a very simple one out of popsicle sticks in high school, and so I figured I'd start there. I looked online at a bunch of different examples, and they were certainly interesting. I found an interesting model of a 3D-printable dual tensegrity table online that caught my attention. I printed it and then assembled it.

While cool, there was an underlying problem: I live in a fraternity house, and I think that a tensegrity table would be too fragile to survive any stupidity that might occur. I wouldn't put it past someone to be curious and then quickly test the limits of the table and accidentally break it. With that being said, I moved on and sought another idea.

At some point, I remembered seeing what I now know are called contour slats. This idea quickly caught my attention, and I began modeling different ideas using Fusion 360. Below, I will walk through the general process that I used for each of the designs and then show a bunch of concepts that I developed in the process.

The General Process

This step was the most difficult of the process. Depending on the initial sketch, Fusion would fail to compute the loft properly as the model almost folded over itself. I couldn't find a proper solution or a way to determine if a design would be able to work, and so I would just loop, editing the sketch and testing the loft.

This is the general process that I used to design the different tables. Eventually, I began experimenting with cutting additional holes or shapes through the lofts. I found interest after one of the lofts accidentally yielded a hole.

Different Designs

Design 1

Design 2

Design 3

Design 4

Design 5

Execution

The final design that I ended up choosing was interesting, as I tried lofting through the middle of the sketch as opposed to simply cutting ovals through them, as you can see in designs 4 and 5.

Here is a short video showing how I got the rough shape of the final design. There was some post-processing that I did beyond the point in the video. This was to clean up some of the sharp corners left after the loft, but that was it. I then went and modeled the upper and lower supports so that the different slats could actually be mounted together.

With the supports modeled, everything has been designed, and here is the final model of the table that I want to make:

With the design completed, I had to move on and begin preparing the DXFs to send to Chris. To do this, I used Fusion's Arrange function to lay out all of the components on two sheets the same size as the plywood we'd be given. Once all laid out, I created two sketches to project all of the geometry onto. With the geometry projected, I then extruded the geometry through the sheet to get the negative of all of the components.

With these sheets, I could then create sketches on top of them and all of the geometry would carry over—perfect for creating the points and bodies DXFs that I'd need for Chris.

I then went through and selected all of the points in the dogbones that would need to be drilled for clearance.

With all of that done, it was time to cut! I met with Chris. To start, we measured the new material that had arrived, and it was around 0.45"; my originally modeled gaps wouldn't work! Fortunately, I used some parameters in my original design to account for this and was able to scale all of my dogbones to fit the new stock. Once I altered them, I sent him the new files, which we then prepped to be cut. Due to the size of some of my components, we had to double the thickness of the onion skin he typically uses to ensure that the parts wouldn't fly off.

Once the files were prepared for the Onsrud, the cutting process was actually quite simple. Due to the quality of the machine, there was little concern for failure, and so I simply waited for the cuts to be done.

Post Processing

The image on the left shows an example component just after it had been removed from the stock. As you can see, the edges are super rough and not in any condition to be used for a table. To fix this, I used the router table to clean up all of the edges and create a much nicer finish.

The other form of post-processing that I had to do was to remove the onion skin from the dogbones in each component. To do so, I simply used a box cutter and cut the material out.

With all of that completed, it was time for the best part—assembly!

With that, the project is complete!

Thoughts and Reflection

This week was a lot. I started modeling different tables on Thursday, and yet somehow I found myself in the wood shop from 12–7:30 on Tuesday, cramming to try and get it all in. I was gone over the weekend, which is why I had to cram, but nonetheless, it was stressful.

I'm super happy with the result and how the project turned out in the end. I swung through a range of emotions throughout the process. Originally, once I modeled the slats, I was super excited with the design. Once I added the supports to actually attach them all, I thought they ruined the design and I was quite disappointed. I told myself that it didn't matter at that point and I just needed to move on and get the project done. Then I went to cut it all out and assemble it, and I swung all the way back and was super happy with the design. I found myself really enjoying using the router table; I found it quite satisfying to take such a rough part and drastically improve its appearance and feel.

As for the design itself, I am glad that I decided to go with the contour slats. For one, I think it turned out quite nice—I really like the organic shapes that came out of the modeling process. Importantly, the table is stable, and I think it can withstand living in a fraternity house, unlike a tensegrity table. I was even able to sit on it (it looks an awful lot like a bench).

It would be cool, once I have more time, to go back through the process, refine the design, and make it out of nicer wood to serve as a proper coffee table.