A lot of making today happens with digital fabrication using machines like 3D printers and laser cutters. What if we could use digital tools to capture our design process in a way that is visibly represented on our fabricated objects? I think this would be a powerful way to tell the story of how a design came about and what iterations led up to a final product.
I wrote a paper about this and conceptualized three types of what I call Process Products:
- Process Heatmaps: a full-color, 3D printed part that indicates aspects of the 3D printed part that went through the most changes in the digital model.
- Process Stacks: a physical stack of laser cut components that were created in the prototyping of a final part
- Process Textures: a representation of machine settings that creates textures along the surface of the part. The textures show the number of iterations and the nature of the changes from one iteration to the next.
The paper I wrote reused some components I had created for a previous project. But I thought it would be fun to see how these ideas might play out in a real, designed product. So my friend and I are working together to create a design that incorporates these different examples of Process Products for a conference in January.
I worked on brainstorming some potential themes to explore for the Process Product demo, including:
- Can we incorporate something that might be potentially useful / interesting to people at a conference?
- Can it be something that is familiar (so we don't spend very long explaining the actual product), but the critical dimensions are non-obvious?
- Can the design incorporate all three process products?
I came up with a few ideas, including a toy train set with laser cut tracks and 3D printed cars and a business card holder.
After meeting with my friend, we came up with the goofy idea of having a business card launcher / shooter. We set off to prototype some designs on our own before settling on a design.
I used LEGO and rubber bands to prototype a roller launcher. It worked decently well when it did work, but it seemed to jam quite a bit. I think there are a few key things that led to issues with the model, including that the wheels I used for the rollers would jam against the frame and the rubber band would get jammed between the rollers. Still, you can see a video of it working above. ( :
This is the model my friend designed - it uses a taught rubber band to "flick" to launch a card from its side. It turns out that business cards go pretty far when you hit it off-center. (:
Nonetheless, we liked how this design had less parts that were likely to fail. Next is planning out what different parts will be laser cut and which will be 3D printed. The plan right now is that the grip and connectors which will attach to the laser cut components that are at 90 degrees from one another will be 3D printed - the rest will be laser cut.
I'll be working on the model in the next few days, and he'll be working on seeing how we can map a color gradient (heat map) onto a 3D model.
The plan for TEI is to 3D print the handle for the business card launcher and any necessary brackets for connecting the 3D printed parts with laser cut components. The base of the launcher (where the card is placed) as well as the rubber band holder will be laser cut. Since the base is large, I thought it would make the most sense to use that for the process texture and use the rubber-band holder for the process stacks.
With this plan, I began by modeling the foamcore prototype in Solidworks and then started experimenting with some surfacing to create a 3D model for the handle. But after thinking more about the layout for the demo session, I started to shy away from creating a handheld launcher. I imagined a chaotic demo session with business cards flying all over the space, which, at the very least, could lead to a length cleanup, and at the worst, could interfere with other people's demos. Also, while this is less important, I found it uncomfortable to model what looked like a gun when I was at home for winter break!
So I decided to think about what a desktop model might look like that can launch a business card to a person at the other side of the table (~3 feet in depth).
My original plan was to have an angled C-clamp to attach the launcher to a desk. I also wanted to incorporate a handle that you would rotate to release the rubber band, which I thought might add some consistency to the launcher.
Eventually, though, I headed in a different direction because I thought clamping the model to the table would be kind of clunky, and the handle was large enough that it would likely be expensive to 3D print, even if hollow.
In this version, I created a standalone base for the launcher with rubber feet to help prevent it from sliding. A bracket connects the base to the launching surface at an angle of 10 degrees.
I designed the base so that you could rotate the base to aim the launcher, but to simplify the design, we decided to remove the rotational element.
What's left to do is to design a grip on the base so that the model can be secured while the rubber band is extended - I'm imagining finger imprints right now for holding the base down. Will work on this in the coming week!
Before refining the CAD model for the base (and 3D printing it), I wanted to get a sense of how someone might grip the base on a table. The easiest (and cheapest) way to do this that I know of is to make a foam model.
I used some scrap blue foam lying around in the shop and sketched out a general form in my notebook. After dimensioning the sketch, I drew the outline onto a piece of paper, including the top, side, and bottom views of the design.
I then attached these images to a piece of blue foam using push pins and used a hot wire cutter to cut through it. This gave me a pretty rough foam that I spent some time sanding to get to the form I wanted.
I then carved out a thumb rest using the sand paper, which I thought gave the form a pretty sturdy feel. I have a medium-sized hand, so I'm going to enlarge the dimensions slightly for the final model.
Note: sanding down the blue foam made me remember why you shouldn't wear black pants in a shop - I was covered in dust afterward!
I updated the base of the model to resemble the foam model I created. To reduce the amount of material for the part, I shelled the part. I also added space for rubber feet to attach to the 3D printed part and for nuts to attach the launching plate to the base.
By angling the top of the base, I was also able to get rid of an additional bracket.
Since I had the foam prototype, I was able to print out a 1:1 scaled drawing of my part to ensure it conformed with the physical model, a trick I use all the time.
Time to 3D print!
To be honest, I never liked using Maker Bots as I was pretty spoiled using more professional 3D printing machines while working at various design companies. But I definitely see its usefulness for lower-resolution prototyping.
That being said, I ran into a ton of problems trying to use the Makerbot Replicator 2 today, which we have at the Media Lab. The extruder would thin out as the job printed, and the plastic wouldn't stick to the bed (see images). I cleaned out the printing head 3 times because it got jammed with melted plastic.
After these frustrations, I ended up looking for other options. Luckily, another research group in my lab had a newer Makerbot I could use and I was able to get the job set up on that. Fingers crossed that it works - it's a 10 hour job!
Came into lab today to find that the print job ended up taking almost 13 hours. It took a while to remove the support material, and I ended up not bothering for a lot of it since it was internal. When I started peeling off the support from the bottom of the piece, the base started coming apart, so I didn't remove the base and cut around it instead. This also meant that I didn't attach rubber feet I bought from a hardware store, but luckily it didn't really need it.
I also used an assortment of 10-32 nuts and bolts from the Media Lab shop and laser cut pieces I cut yesterday in the shop.
As you can see from the video above, the model works relatively well, but there are a few issues I'd like to fix:
- If you look at the second video, you'll see that the assembly is unbalanced and falls over on its own. I think a large part of this is that I used metal nuts and bolts since they were on hand in the shop, but if I used plastic fasteners, it would be lighter.
- I realized that the launcher plate and bow were larger than they needed to be, so I wanted to resize them, both to make the entire assembly smaller but also to help with the balancing issue.
- I had added a small guide that helps you extend the rubber band but think it should be a little bit longer.
I adjusted the dimensions in the model to make the laser cut parts smaller. You can see how the outline of the parts changed in the images above (the process stacks will be useful for this!)
I then went ahead and laser cut these new pieces, which I'll describe in the next step.
All the business cards I've accumulated over the years are finally being put to use!
You can see a video of the new assembly above. It is still slightly unbalanced but much better than the previous version (it stands on its own). I think that using plastic fasteners will solve this issue, and I plan to pick some up tomorrow!
Otherwise, I think what's left to do is just planning what the actual Process Products will look like. I used the Pack & Go feature in Solidworks to create new versions with most major changes, and I'm now on Version 9 of the CAD model. More soon on how these models will be integrated into making process products!
After finalizing the CAD model, we could get working on the actual process products. There are three:
- heatmap - the 3D printed base, which will be color 3D printed to reflect changes to the CAD model
- stacks - laser cut components representing changes to the outline of the launching plate (where the card sits)
- textures - a laser cut component (the rubber band holder) that shows how the speed and power settings were adjusted over time.
I started off by modifying the launching plate part file from each version of the assembly, adding a small hole to the center of the part.
I also purchased some ball-chain from the hardware store to hold the stack together. I ended up using a ball chain instead of a screw because I thought making the stacks more of a swatch could make it easier to compare individual version of the design.
Then I laser cut each piece and rastered version numbers on each part. Since I was mostly using scrap material from the shop, the design files I shared here are arranged to fit on those pieces.
You can see the final laser cut pieces above!
In Illustrator, I imported the outline of the "V" piece (that holds the rubber band). I then used the move tool to create a striped pattern representing adjustments to speed settings over time. (I had copied down changes I made to the laser cutter speed settings the first time I had used it to create prototypes).
Finally, I added the exact speed settings to the file to raster them directly on the part.
After assembling the first version of the process stacks, I realized it was difficult to see the changes between the iterations for a few reasons. First, when they were stacked directly on one another, it was difficult to distinguish between which iteration you were looking at. Second, the swatch design meant that each piece could rotate relative to one another, which compounded the problem.
I decided to return to the original idea of using a screw but physically separating the layers using spacers. I wanted to use standoffs, but since there wasn't time to purchase them, I used a combination of set screws and spacers that I found in the Media Lab. We have an amazing selection of hardware.
Generally I feel better about this version - it is easier to compare the versions when they are spaced apart. But it will be more likely to break since each piece is only supported at the center. We'll see how it goes...
We had to design a poster for the conference, which I've shared at the following link:
You can check it out if you'd like to read more about the project. (:
Assembly was a fairly quick process after we got all the parts together - we assembled it for the first time the night before the conference, and luckily everything fit!
Presenting at TEI was fun and very low-stress. I was kind of amazed that nothing broke as I expected the part might get knocked over (I had brought extra laser-cut pieces just in case). Overall, people seemed to have fun learning about the concept of Process Products (and also shooting business cards around the room).
One of the most helpful feedback was to connect the physical representations of process with additional information. For example, if someone wanted to learn more about why certain changes were made, the physical parts could somehow link to digital content that would explain more about the rationale behind design changes. Perhaps there is something that could connect the physical parts with Build in Progress?
I'm not sure what exactly will happen with the project now, but if anyone is interested in collaborating on it, just reach out to me. (:
My friend gtsai worked on the heatmap element of the project, which involved incorporating two iterations of the base CAD model. It ended up being an involved process of exporting projected images of the colored model from various perspectives as well as the STL. I'll leave space here for him to write about it.