INtroduction
To begin to print an impossible object, I initially experimented with a few different objects that I found on Thingiverse. I initially attempted to print a couple of different hinged butterflies by mathgrrl and jijimath, but I encountered some issues with print resolution and fine features on the butterflies with FDM printing. I ultimately selected this rolling knot object by kitwallace to print as my selected impossible object.
Scaling
I started by attempting to print the object on an FDM printer, the Prusa i3 MK3, since I have used an FDM printer before. My first attempt at scaling the object to fit within the bubble gum capsule did not go as planned. Upon downloading the file, the dimensions were approximately 60mm x 60mm x 50mm. I thought that I scaled the file down to fit in the approximately measured to be 50mm x 50mm x 50mm capsule, but it appeared that the scaling did not save properly in the file and the object printed full size & did not fit in the capsule. I re-adjusted the settings so that the maximum dimension was 30mm and the print then fit nicely into the capsule.
PrintING with FDM
Printer and Settings
I printed the impossible object on the Prusa i3 MK3 printer with PLA material. I began with the default OEDK settings for infill and layer height. After a couple of failed prints, I adjusted the layer height from 0.2mm to 0.1mm, thinking that this may increase the strength of the part and prevent clogging, and this did seem to help with the prints. I also encountered failed prints when I tried to print more than one object within the same print file. Therefore, I stuck to only printing one object at time. I also at one point attempted to print with the NinjaFlex material on the Prusa i3 MK3, but it seemed that this material did not work well for the design of this object, as the print did not look clean around the angled portions.
SUPPORT MATERIAL
I did end up using support materials for all of my FDM prints. A couple of times I attempted to use to original orientation provided by the object designer without support materials; however, the nozzle clogged each time on the Prusa. I found most consistent success with the prints when I used supports generated by 3DPrinterOS. I modified the orientation of the object when using support material to optimize print time & material used with supports.
POST-PRocessing
Post-processing involved removing the support material from the prints. This was not particularly easy to remove (I used tweezers and scissors), and I can see why it would be helpful to print these impossible object knots without support material. It would also save print time & resources. I would be interested to know if the nozzle clogging without support material was due to my own printing abilities, or if printing without support material was intended with the design. From reading about others who printed this file, it sounds like minimal support material may have been used when using FDM printers.
PrintING with SLA
Printer and Settings
I used the same scale of the part to print on the SLA printer, since the previous scaling fit nicely in the capsule. The printer used was the Prusa SL1S SPEED printer; this was my first time working with liquid printing. The material used was the Prusa UV Sensitive For 3D Printing Tough Resin. I selected the “0.05 Normal” setting in the PrusaSlicer, since I did not believe that the print required “0.025 UltraDetail.” In the future it could be optimized to use the “0.1 Fast” setting, but at this point I wanted to ensure the quality of the print.
SUPPORT MATERIAL
After some basic research, I found that SLA prints do best with some support material and when positioned at an angle. Since the object was already angled, I decided not to further angle it. I also kept the default support materials in-place rather than attempting to see if I could print without supports as I did with the FDM prints.
POST-PRocessing
Post-processing of this part involved washing, drying, and curing. This was done with the Prusa CW1S washing and curing machine. First, I placed the finished part into an isopropyl alcohol bath in the washing machine (while still attached to the print bed), and started the washing cycle which lasts 10 minutes. Afterward, I removed the bath from the washing and curing device, removed the print from the bath, placed the print (still on the bed) back in the machine, proceeded to dry for 3 minutes, and finally cured with UV light for 10 minutes. Following, I removed the support material.
COST ANALYSIS
FDM (Prusa i3 MKR)
The final file I used to print on the Prusa i3 MKR used 4.63g of filament, cost $0.09, and took 1 hour and 45 minutes to print. Post-processing time was minimal. Labor time once the process was established was approximately 30 min. To calculate machine use cost, I used the metric of expected printer cost/printer lifespan to establish a printer use rate of $0.38/hr ($300/800 hrs) + a baseline fee of $5.00. I then multiplied this by the time it took to print to achieve an estimated machine use cost of $5.67 ($0.38/hr x 1.75hrs + $5.00).
The following is a cost break-down for the printing of one object:
- Materials: $0.09 (from 3DPrinterOS)
- Labor $5.00 (for 30 min, $10/hr)
- Machine Use: $5.67
- Total: $10.76
With scaling, I would anticipate that labor cost would decrease but materials and machine use cost would remain somewhat constant.
SLA (Prusa SL1S SPEED)
The final file I used to print on the Prusa SL1S used 3.4 ml of resin (2.58 ml for object, 0.83 ml for supports) and took approximately 48 minutes of printing time, plus the additional 23 minutes required for post-processing procedures. I found that the cost of the resin used is $80 for 1kg, which contains approximately 1000 ml of resin. Therefore, the materials cost for 3.4 ml of resin calculates to be $0.27 (3.4ml/1000 ml x $79.99 = $0.27). For post-processing with the SLA print, I found that 1000 ml of isopropyl alcohol costs ~ $10. I estimate that around 1000 ml of isopropyl alcohol was used in the bath bin. This can be reused; so this cost would be reduced at scale. Labor time once the process was established was approximately 1 hr. For this printer, I again used the metric of expected printer cost/printer lifespan to establish a printer use rate. From this, I established a printer machine use rate of $0.27/hr ($2000/7500 hrs) + a baseline fee of $5.00. I assumed washing and curing machine use to be negligible. I again multiplied the machine use rate by the time it took to print to achieve an estimated machine use cost of $5.23 ($0.27/hr x 0.83hr + $5.00).
The following is a cost break-down for the printing of one object:
- Materials: $0.27 (See above calculations)
- Post-processing materials: $10.00 (For isopropyl alcohol, see above estimations)
- Labor $10.00 (for 1 hr, $10/hr)
- Machine Use: $5.23
- Total: $25.50
At scale, I would expect the cost to decrease significantly, as isopropyl alcohol could be reused many times per part. Additionally, I would expect labor cost to decrease at scale.
CONCLUSIONS
Both prints appeared to be fairly high-quality. However, the SLA material was smoother and produced what I would consider to be a higher-quality print in comparison to the FDM. Due to this difference in quality, it is understandable that the SLA overall cost was more expensive. SLA also required more labor time due to the post-processing procedures. While I appreciate the smooth texture of the SLA print, I think that this print could be consistently printed at high-quality on an FDM printer, since there are not incredibly fine details that require extra resolution or attention. Through this process, I learned how common it is for prints to fail and the precision required in executing a successful print. I also learned the basics of SLA printing, liquid printing, and relevant post-processing procedures.
