Week 5, 3D Printing and Scanning

Objectives

Group assignment:: Test the design rules for your 3D printer(s); Document your work on the group work page and reflect on your individual page what you learned about characteristics of your printer(s) Individual assignment:
Individual assignment:: Design and 3D print an object (small, few cm3, limited by printer time) that could not be easily made subtractively; 3D scan an object (and optionally print it)

Group Assigment

The group assignment had some interesting results. It was insightful to see the capabilities of the FDM printers and the slicing software. It is good to know that the minimum wall thickness is about 0.5mm and that they will print well at 70 degree wall thickness. This will be helpful as I prototype my final project and other things.

Link to group assignment page

3D Print an Object

For the 3D printing part of the assignment. I wanted to create another tool for my camera, so I created a flexible flare hood to create custom shape vignettes in-frame. This part is not easy to make subtractively because it has undercut areas that cannot be accessed by cutting tools, and it has long thin leaves that would be hard to make accurately.

Inspiration

While shooting a training session in the Dominican Republic last year, I was looking for ways to make more interesting images of the session. Since we had a bunch of Post-It notes hanging around, I attached them to my flare reducers and used my left hand to pinch them into the frame. The result was that I could make custom out of focus borders to frame the people in a pleasing way. Since then, I have been wanting to create a more robust way to create this effect without having to use Post-Its.

Photo shot using green Post-Its to frame the subject.

First Iteration

I did a quick sketch of how I wanted the device to work. I wanted to have some flexible fingers to be able to press into the frame with my left hand, and I wanted them to be all the way around the lens to customize each frame.

Then I took the concept into CAD and used my existing lens hood CAD as a base and added 16 fingers along the circumference. I then wrapped the leaves back on themselves so my fingers would have a place to push but not be in the frame themselves.

I test printed the CAD on a paper printer to gut check the dimensions before sending it to the 3D printer. Since the part had overhangs I used grid support material to make sure it would print properly. The part was imported into Prusa Slicer to setup the job. I oriented it with the side connecting on the camera on the floor, and the leaves vertially up in the air. I did it this way because I wanted the geometry that connects to the camera lens to be printed accurately without supports, and it was a good orientation for the support material to go up under the wings. If I did it sideways I would have had support material everywhere and it would have been a nitemare. I know in this orientation the leaves will not be as strong as the bending force is tring to break the layers apart at their weakest point, but it was the compromise I had to make.

I left the default 15% infill but selected to add supports. Then I sliced the file to see what the material path looked like.

Looking deeper into the model, while I chose 15% infill, the model was effectively 100% infill as the nominal wall thickness of the part is only 2mm. Thus there was not enough surface area in any of the layers for the infill pattern to kick in.

Output from Prusa Slicer inside the layers. Note that wings of the design are completely full of filament and thus effectively 100% infill.

Then I sent the file to the Prusa Mk4S in our Fab Lab and ran the print. As I used PLA for this print with a clean print bed, and there is so much surface area on the printe bed for the print to stick to, I did not use any glue or extra adhesion for the job and it ran just fine.

The print of the first iteration. I used yellow PLA on a Prusa MK4S

I installed it on my lens and gave it a test. Generally it worked pretty good. However, there were a few issues.

While the lens is circular, the image sensor is rectangular. So the upper and lower fingers of the device were too hard to push into the frame of the image
The ring where it locks onto the lens was a little too thick and was hard to twist onto the lens where the stock lens hood mounts.

However, I was able to get some shots that had the effect I was looking for.

Second Iteration

I updated the design to pinch in at the top and the bottom to get the fingers closer to the field of the image and I thinned out the engagement ring and re-printed.

Iteration 2 CAD. Note the top and bottom of the ring of leaves is squared off.

I used a Bambu A1 printer and gold PLA filament for this iteration. I liked the brightness of the yellow color of the first print and thought the gold would make some good effects. However, I dropped the part while I was trying to mount it and it shattered on the floor.

However, it was not a total loss, as I learned a couple of things. 1. I found that the pinched in areas needed to be 90 degrees from where they were on the model so that the would align with the locking nubs on the lens hood. 2. Clearly there was a massive difference in print strength between the two printers and materials. I suspect that whatever makes the gold shiny may also make the material brittle.

Third Iteration

For the 3rd iteration I rotated the fingers 90 degrees and I added support ribs near the root of the fingers to provide some strength in that area.

Then I printed it with Bambu grey basic PLA and tested it out.

This one did perform well and was more robust than the second iteration, but still had some problems.

The fingers were still a little bit fragile and I broke one in testing
The top and bottom sections are still too far away from the frame and need to be pushed down too much.
Grey is a boring color for the framing.

Image made with iteration 3. Note the grey hazy shape in the lower right.

Fourth Iteration

Undeterred, I continued on for a 4th iteration. I narrowed the opening of the flare hood to try to make it easier for the upper and lower leaves to get into frame. I also completely changed the leaf structure to have a curve to them. I was trying to increase the effective length of the leaf beam to help soften them up a little and keep them from breaking.

For the material, I selected a semi-transparent PETG. I thought this might be a bit more robust and have more interesting optical properties.With a different material this required some differnt settings, but overall the same strategy. I used the "Generic PETG" settings inside the Bambu software to setup the print and used support material. Again since the part is constant small wall thicknes, even though I set the infill to 15%, the part was 100% in actual printing.

The 4th iteration showing the support material.

The drawback was that the weeding of the support material was harder as the bond strength of PETG to itself is stronger than PLA. I used side cutters to trim away the material. Usually it was easy as the support only contacts the part in small areas. However, depending on how the adhesion of the support material to the part was, there was the change to break the leaves, and I had part of one leaf break while trying to trim it out.

After weeding the print. Note the broken part of the leaf.

Despite the extra work weeding, this was the best working version of the bunch. The leaves were still a bit too far apart, however they were much stronger and I was able to bend them into the frame in a nice way. They had good spring back and I didn't feel like I was going to break it. The semi-transparent color also made a nice effect in the images that I would definitely use on a real shoot.

Test shot pinching the leaves into the shot.

Edrawings part file of iteration 4 Note that the native SolidWorks files are too large to upload.

STL file of iteration 4 Note that the native SolidWorks files are too large to upload.

3D Scan an Object

For the scanning part of the activity I attempted to scan two things, my camera lens and my kids foot. For all of the scanning I used the Creality Scan Ferret and the associated scanning software.

Camera Lens

Since I have been doing a lot of CAD work and my final project around my camera lens I decided to do that first.

For the setup, I used the Creality Scanferret scanner. Since I already had it installed it was pretty easy to plug in and get up and running.

For the setup, I put the lens at the center of a cake decorating spinner. Then I held the scanner by hand and slowly spun the platform 360 degrees to capture the data. I then moved the sensor around the top of the lens to capture front side.

It obviously did not capture the lens, but did capture the aperture visible through the glass.

Scan data from my 50mm Sony lens. The blue surface inide the main body is the aperture.

Then I processed the scan inside of the Creality software to make a CAD mesh. Unfortunately it deleted the aperture plane and wallpapered over the lens area with a garbage surface.

Then I saved this as an .stl file so I can view it in other CAD packages.

Then I exported as a .ply file to have a slightly smaller file to post here.

PLY file of the lens

Feet

My youngest daughter has been begging me for about a year to make 3D printed shoes and I haven't really been able to come through on that request. But I used this opportunity to take a step towards that goal.

Initially I just had her sit on the table and try to hold her foot still. This worked sortof ok, but the scanner kept losing its orientation and it was double scanning.

Then we had a better idea of trying to have her make a footprint and scanning that instead. At first we tried to use Playdoh, but we did not have enough that was still squishy.

Then my daughter found some Crazy Aarons silly putty. This was pretty hard, but workable. We rolled out onto a piece of parchment paper with a soda can. Then she stood on it and made the print.

I then took the putty foot print and put it on the same platform used for the lens. It took some fidling to get the setup right to get it to scan the footprint. It felt like it was thinking the putty was part of the background, so I had to tell it to not ignore the scan surface. Then it scanned ok. I trimmed alot of the data around the print and processed the surface.

I took this a step further and tried to process the scan with nTop software. I brough the .obj file into nTop and turned it into an implicit body. Then I created a rectangular cube that overlapped it and trimmed the cube with the scanned surface. I tried to get it to create and exportable CAD body from the mesh, but the model has some non-manifold areas that I have not been able to figure out. However, this has potential to be something I can bring back into SolidWorks to create a sandle out of.

nTop workflow processing the scan data and trying to make a real CAD body.

STL file of the foot scan