Corey M. Rice : Fab Academy 2017

3D Scanning and Printing

• Test the design rules for your printer(s) (group project)
• Design and 3D print an object that could not be made subtractively
• 3D scan an object (and optionally print it)

Learning outcomes:

• Identify the advantages and limitations of 3D printing and scanning technology
• Apply design methods and production processes to show your understanding

Evidence to show mastery:

• Described what was learned by testing the 3D printers
• Shown how you designed and made your object and explained why it could not be made subtractively
• Scanned an object
• Outlined problems and how you fixed them
• Included your design files and ‘hero shot’ photos of the scan and the final object

3D printers are great. There is no denying it is incredible to watch an idea materialize, layer by layer, until it forms into a complete thought. This seems to be the closest thing we have to a Star Trek Replicator. I cannot wait to see where this technology will develop. I look forward to a day where ideas can quickly turn into complex, multi-material objects. However, that day is not today.

In the lab were I do my work, we have a few MakerGear M2 Printers and typically slice 3D models for printing with Simplify3D software. This is the case because of where we are located. MakerGear is a 3D printer manufacturer that is located in another nearby suburb of Cleveland, Ohio. Their printers are an excellent quality product and they recommended Simplify3D to our lab/ school. A student of mine is an intern at MakerGear who has been able to learn a lot with the company. He even designed the MakerGear holiday ornament. One of the teachers at the school personally owns the sole MakerBot that is in our lab, and he has had terrible luck with the thing. Right now it has a broken print head.

A Test Print

In the little that I have worked with the 3D printers, I have learned this: they aren’t perfect. A test print is a great way to see what you can expect from a printer, and how you can design for the combination of printer and slicer software that you have. So, this week, I set to work on the group project. Again, I do almost all of my work in isolation from the rest of my local Fab Academy classmates, since it is a long way to the official Fab Lab and I have resources closer to home. So I designed this test print and used it to gauge how my designed for the week might be restricted by the realities of the machine. It only takes a little experience with 3D printing to see that a 3-D model is often an idealized version of what will be printed. A slicer and then printer will each reduce the perfection of a model into the fallibility of real-world objects.

I knew I wanted to make sure that I hit all of the marks set forth by Gershenfeld in the lecture and assessment book: minimum printing and overhang angles, wall and gap thicknesses and minimum text sizes. This meant that my test print model would need to include all of these features. First, I set out to design separate incline angles of decreasing degrees. The idea with this feature would be that I could identify what angle was too sharp to be resolved by the printer accurately. All of the model wedges would terminate at the same endpoint; but, if the print was shorter, I knew it would be too sharp to design for a real-world.

The wall thickness would be tested by modeling walls of decreasing sizes – each half the width of the previous. The test print will show at what thickness the slicer and printer will not faithfully produce the model.

The gap thickness would similarly be tested by comparing the 3D model to the printed part. If a gap was too small to be ‘practically useful’ then it can be considered outside the limitations of the printer.

Again, comparing the model to print is useful for seeing what overhung angles are practically possible. In these images, it is possible to evaluate the discrete angled sections and the rounded cutaway.

The takeaway from a test print like this is the limitations of the printer. Knowing these limitations can inform design for items that will be made on this printer. Also, it is worth considering that the Z-axis is often the weakest of a fused deposition modeling (FDM) style printer. So, if this test print were in another orientation, different results may arise. If I were to revise this test print model further, I would place one or more vertical cones protruding out of the top of the model to see how small of a print width could be produced.

A Thing Made Additively

There are many ways to make things. Woodworking or sculpting marble are ways to make things, and these are similar in one way: they remove material from the stock. 3D printing is different in that it is fundamentally additive: material is added to the thing being made, rather than taken away. This distinction is more than a curious difference of production. Some things can only be made additively. Internally complicated materials, objects with interior features, or interlocking objects are a few examples of items that can only easily be made additively.

This week’s assignment is to design and print an object that could not be made subtractively. I believe that some of the things that I designed during week 2 may meet this qualification. The hopper that I designed has a channel cut out (of the blue part in the video) that would be difficult, if not impossible, to cut out on a mill. While this may not be able to built subtractively, it is not clear to me that this is so.

Consequently, I decided to produce a new model that unambiguously meets the week’s criteria. I wanted to design a car that is propelled by air pushed through a windpipe that a mill could not cut, and is printed with the wheels in place. To achieve this I had to do some calculations. I designed a pair of wheels and axle that is 1.5 inches long. The car body is also 1.5 inches wide and the first iteration had wheel wells that were only 50 thousands of an inch larger than the wheel themselves.

I based this design on my test print. It appeared that the 50 thousandths gap should have printed properly. Also, the wind pipe was designed to have an interior diamond shape since this would mean that the interior overhangs within the windpipe would never exceed 45 degrees. This also meant that the slicer would not automatically place supports within the windpipe (which may be impossible for me to remove after the print completes). In fact, I did not allow Simplify3D to automatically place supports for my first iteration. I manually placed then in Simplify 3D under the chimney of the model only. This was the first iteration. Lots of the model printed beautifully, but it certainly had some flaws. The 50 thousandths of tolerance turned out to be enough to let me break the wheels free, but not to have anything move smoothly. It takes serious effort to get the wheels to move.

So, a second iteration was made. This version had much more space around both the wheel and axle, but the wheels were still about as thick as the wheel wells. Not a terrible plan but, without supports between the wheels and car bodies, the inside of wheels that printed on top were ‘stringy’ and thus jammed up the car. So, on to revision three.

The third revision had a serious expansion of the wheel wells in all dimensions. This time the axle would be able to move with (perhaps too much) slop, but everything should at least work. The chimney and windpipe had been working on the previous versions of the car so I can take comfort in that…

So I sliced the model again on another computer’s version of Simplify3D and loaded the gcode onto an SD card. Normally Simplify3D will let you pass the gcode to the printer of USB, but I needed to use the same printer that was connected to the computer with a problematic version of simplify 3D. (All the other printers were in use.) In any case, I could load the SD card into the printer and print the gcode from the SD card directly. I left the lab tonight with this print still going on the machine. I look forward to the outcome in the morning.

[UPDATE: The last version worked beautifully, and was green.]

3D Scanning

3D scanning is delightfully fun. It lets you take a real object and, using a series of 2D photos, stiches together a 3D model of that object. There are, of course, other technologies to get this done. Certainly several of those technologies must be better or worse at 3D scanning. I specifically used a borrowed a 3D systems Sense2 scanner. The process was delightful. The software is beautifully configured to work easily. I had one of my students (the one I mentioned earlier who works for MakerGear) sit in a chair and spin as I pointed the scanner at him. While it didn’t come through perfectly on the first try, the second went much better. The software even has a few touch-up features built-in that make the process convenient and fun. Honestly, this process plays more like a toy than a tool. It was just that easy. So I had the same student scan me. How could I not? Once the scanning was complete it was easy to drop the exported stl files into Simplify3D and prepare the print. It took 10+ hours, but I would say it was worthwhile.

The next day I had a shockingly accurate self-portrait bust, and so did my student. I believe that I could learn to use an STL mesh editor software to touch-up a parts of the scanned models, but the scan quality was so good that it didn’t seem to matter. The following day I just enjoyed having a bust of myself sitting on my desk. But I think I will have to put it somewhere else long term – it feels to narcissistic to be sustainable in the long-term. Anyway, the bust spurred enough curiosity in another of my students that I agreed to scan her as well. It turned out that long hair was much more difficult for the scanner to deal with, and it took a few attempts to get everything to work properly. However, I was able to scan her successfully as well. Now her model is printing and she will have her own bust in the morning.

There are certainly plenty of limits to 3D scanning, even moreso with the device that I used. Since the scanner is really just stiching together images, it doesn't have the resolution to resolve fine details of things like hair or small wrinkles in the skin (perhaps thankfully). Other, more expensive, structured light scanners sould do a better job, but there would still be limitations. This seems like a useful technique, but at the level of resolution that I can access at my school, it is little more than a novelty toy.

You can see all images (used and unused) from this week's work HERE