Design on Thingiverse: Geared Heart by UrbanAtWork

I saw this keychain on Thingiverse and was surprised by how the 3D printer would be able to print 3 gears contained in a heart outline in a single print. No supports needed, and the gears would be able to rotate each other. 

At first, I was really skeptical that it didn’t need supports since I saw a part of the keychain that was, essentially, a floating bridge connecting two axles. However, after talking with Fernando at the Maker Bar, I learned that 3D printing can bridge quite far without them.

 

First print: Scale was too small for toolpath

In the first print, I noticed that some corner edges of the gears were barely still attached to the rest of the gear, and eventually by touch they would fall out. The gears were able to move, but not very smoothly due to the impaired grooves of the gears. I had scaled down the file to 75%, and I believe it was too small for the toolpath of the edges.

Next iterations: Scale to 85% had better gear edges

In the next couple of prints, I scaled it up to 85%, which made the keychain just big enough to barely fit in the capsule. I noticed the edges of the gears improved and didn’t fall out as easily. The rotation was smoother, too. Some 85% scale prints, while they are an improvement, still had gear edge issues.

 

Print failure: Filament layers didn’t stick together

On one of the prints, the gears weren’t able to move, and this was because the layers weren’t adhering to each other towards the end of the print. Apparently, the printer I was using had similar issues on other prints, and it was taken for repair. Another failure was when the first layer wasn’t sticking to the bed, and so Fernando recommended that it is good to wash the print bed to get rid of any oils/grease when this happens.

Increasing number of layers didn’t improve print significantly

For one print, I decided to decrease the layer height to 0.1mm. This doubled the print time. I felt that the layers were more defined, however, it didn’t fix the problem I had with the gear edges. This illustrated to me that the main problem with the gear edges was the scale factor, and unfortunately, 85% was the max size that would fit in the capsule. Another factor I noticed was the filament type. The blue filament had a worse gear edge outcome than the orange filament. I’m not quite sure why that’s the case, perhaps different brands affect the filament, but that would be something to look into.

 

SLA print: Immobile gears, but better detail in logo

Using the SLA printing method, I saw how the gears were fused with the axles they were supposed to be rotating on, so obviously for the purposes of gear rotation in this keychain, the FFF method would be better. However, I noticed that the “Urban” logo on the back of the keychain was much clearer and more defined with SLA, showing how designs that require more detail could benefit from SLA.

 

Video clip showing lack of gear rotation in SLA and logo

Best print: Full gears with intact edges and smooth rotation

Video clip of good gear rotation and logo

 

Cost Analysis

FFF: The print is 5.20 g PLA, took about an hour, and costs $0.11. The print that had a smaller layer height took 1 hour and 57 minutes but had the same cost. Five PLA prints using FFF totals to $0.55.

Machine time: Prusa k3 MK3S = $750. Lifetime = 800 hours. Therefore, cost = $750/800 = $0.94/hr. About 7 hours of actual printing would be (7 hrs x 0.94/hr) = $6.56.

SLA: Material = 10.08 ml x (1 L / 1000 ml) x ($145/L) = $1.46

SLA Machine time: based on this website, taking into account average machine cost ($3500), lifetime hours (8000), maintenance cost ($500/yr), depreciation/hr of machine ($0.44), the total machine time cost/hr is about $0.70. My print took about 45 minutes, so (0.75 hr x $0.70/hr) = $0.53

Labor: Approximately 5 hours, from looking up designs to sending files to the printer and to monitoring the printers initially to make sure the print was sticking to the bed. Texas minimum wage= $7.25, so (5 x $7.25) = $36.25. The labor cost would decrease as the file improves through modification and the person (me) gets more practice 3D printing.

 

Total cost: $0.55 + $6.56 + $1.46 + $0.53 + $36.25 = $45.35

Labor= 80% of cost

For five objects, each one should sell for at least $9.07. So buying each capsule for a quarter is not the most profitable.

 

Conclusion and Insights: 

I found that many factors could influence the outcome of a 3D print, such as scale factor, filament type, 3D printing method (FFF vs SLA), and the state of the printing bed. While 3D printing is technically a hands off process, I felt that the beginning of each print required a lot of monitoring. Further, even though the file didn’t change, random variables can make one print less precise than another print. However, this process to make an “impossible” object taught me that you could make seemingly complex things “simple” depending on the method of creation. In this case, 3D printing.