Project Timespan
Late August and early September 2020
Inspiration
There is a backlog of a few dozen projects over several years that I want to share through these clogs, and I am torn on whether I should continue the roughly chronologic order that I have followed thus far or simply write on whatever suits my fancy. As a compromise, moving forwards, I will incorporate a categorized-chronologic system, in that clogs within each craft will be written in chronological order, but I will freely jump back in forth in time between projects of different crafts.
Project Overview
This project can be subdivided into a few sections:
1. Designing the CAD file (Autodesk Fusion 360)
2. Tool pathing the file (Autodesk Fusion 360)
3. Machining the design (GRBL Control Candle)
4. Finishing touches
In this post, I will explain the design process I used to make a Weill Cornell doorbell cover, which happens to be one of the first projects I made with the a 3018 hobby-level CNC machine: SainSmart Genmitsu website. I’m not jumping back too far in time, as it has been but nine months’ time that I have dabbled in CNC machining.
1. Designing the CAD file
I am a beginner in CAD, and I did not parameterize any of the design; I still don’t really do this, but I didn’t then, either. As a student, I have a free version of Autodesk Fusion 360. I’m not going to go into too much detail in the actual Fusion360 work, as there are plenty of excellent how-to guides on designing CAD files. I will, however, elaborate on my design process.
I went through a few iterations on design of the doorbell cover. While I was thinking through possible designs, I decided to add a nameplate that covered the original plate of our apartment number: 13B. The design of this piece was very straightforward, and I probably spent the most time on font selection and size! I did mess up on the machining of the backside, but this is not readily visible!
In the end, I settled on a design of the Weill Cornell shield. I figured that it would be a safe choice given that I lived in school-sponsored housing, and anyone I invited over for a get-together would immediately recognize the logo.
The original doorbell is a rectangular shape that abuts from the door itself to a depth of approximately 1/3″, and the doorbell button itself further extends another 1/3″ or so. One of the images below presents this layout (albeit it is an image of a random apartment 6J). There were two main possibilities that I could envision with the doorbell button. I could either cut out a near-exact fit of the button, essentially a circle, from the Weill Cornell shield, or follow the divisions of the Cornell shield to create a non-circular button. I opted for the latter and decided to incorporate the book of the Cornell shield as a pressable button – I figured that it would not look out of place if the book jutted out from the rest of the doorbell cover, while it would look odd if a small circle popped out from the rest of the Cornell shield.
I had some hassle with sizing of the Cornell logo – I wanted to have the actual button fall in the center of the book, as this would make the balance of the book piece much easier. However, I given the shape of the shield and that of the rectangular doorbell I had to cover up, I wasn’t able to accomplish this goal.
2. Creating the Tool Path
This was a fun learning process, and fortunately it wasn’t very difficult. The initial step is to create a tool library; essentially, inputting the dimensions of one’s tools. This involves taking digital calipers and measuring specific geometries of the bit, including the bit diameter, shaft diameter, overall length, shoulder length, and flute length.
There were three pieces total for the doorbell cover: the bottom layer, the top layer, and the button. For each of these pieces, I used a 2D contour to cut out the border, and the 2D pocket to clear inside areas. Then, I ran the simulated tool path. When I was satisfied that the created tool path was reasonable, I exported the gcodes from Fusion 360.
3. Machining the Design
The SainSmart 3018 Pro comes with a software package called GRBL Control (Candle) that reads in the G-code and sends the cutting instructions to the CNC router. The most important details I kept in mind with respect to this process was checking the level of the base, carefully zeroing the z-axis, and ensuring that the geometry to cut was fully within the confines of the machine’s cutting bed. As with many 3018 models, the bed itself moves in the x-axis while the z-axis and y-axis motions are dictated by the spindle mounted on the gantry, and I overestimated the cutting capacity along the x-axis.
4. Finishing Touches
After cutting out the design on the CNC router, I sanded down the parts by hand (as I didn’t have a random orbital sander with me at that point in time). While I did like the look of the natural wood color and grain, I wanted to pain the doorbell cover in school colors (red and white) so that it was more recognizable.
I had a bit of difficulty in the installation of the doorbell button, as I went ahead and cut of the pieces without thinking through the design. The actual button on the door is rather wobbly and made it difficult to simply glue the doorbell button on. This was complicated by the awkward position of the actual button – it did not sit in the center of the doorbell button, but rather contacted off-center towards the top. To solve this issue, I decided to just fill in some of the space with sponge and used a not-insignificant amount of tape to hold everything together. From the outside, everything is nice and clean, though the inner workings are admittedly far from aesthetically pleasing.
Lessons Learned & Improvements
As a first CNC project, this was certainly a big success. The main issue I had was that I broke a couple of bits with settings that involved too fast or too deep a cutting pass. I expected this, though, as few projects will go completely without hitch, especially one that marks a foray into a new craft. All in all, I was quite pleased with the results, though I had a few gripes with my final design. I had to rely on a somewhat inelegant solution for the button that involve tape and sponge, and if I wanted to spend more time on this project, I would have opted instead for a hinge-like knocker contraption, which I think would work much better. In addition, I was not able to think of a good way to incorporate the viewport into the design.
Reflection
I highly recommend making a custom apartment doorbell cover as a project – it’s just so much fun!