As you know, I started designing custom 3D-printed shoes for myself. This is a follow-up.

So, I spent some time modeling a basic, thin-walled shoe. Nothing fancy, just a quick something to test the fit and do my first TPU print - although I'll probably start with test pieces before printing this doozie.

One shoe fits on the bed of our Prusa Mk4 - barely.

It's a 17-hour print. Yikes! I think I'll probably do this over the next week-end, otherwise my colleagues will kill me.

The support is mostly inside the toe box and promises to be a real b*tch to remove:

The shoe is very close-fitting, but I've left 2mm all around inside: if I mess up, or if the TPU shrinks like I think it will, it has a chance to remain wearable. If it doesn't, I'll simply put a sock on.

Stay tuned 🙂

Our Prusa Mk4 only has one extruder and no MMU. But I was playing around with filament change in the middle of a print to have several colors in a single part.

So I drew a small part with 3 characters connected by a common "bridge". Printing the characters face down, I tried inserting a color change at the first layer of the bridge in PrusaSlicer:

This works fine: the printer parks the head in the lower right corner of the bed, ejects the filament and wait for a new filament to be fed. Then it purges for a while and somehow manages to cut the purged filament clean and continue printing. A bit sketchy but it works.

Then I tried a 5-color print with multiple colors in X, Y and Z. Obviously the filament needs to be changed mid-layer for most of the layers. I defined 5 virtual extruders in PrusaSlicer, added a custom tool change G-Code to force a filament change with M600. After a bit of messing around with layer thicknesses, I managed to limit the print to "only" 16 manual filament changes. A bit annoying but it's okay for an experiment:

When I launched that print, it started out well, printing the first layer of the green "5" digit.

Then it hit the first M600 instruction, proceeded to eject the filament, and horrible noises from the extruder ensued:

https://toobnix.org/w/5M5n1ZxxPjHgp9uPWUKu8s

I could NOT free up the filament for the life of me, and the printer wouldn't come out of the menu to resolve the problem. Eventually I had to reset the printer and apply force to finally clear the problem.

Fortunately, no damage occurred and I could reload new filament as if nothing had happened. I thought it had been a bad fluke, so I restarted the print and it did it again.

Before I file a bug, anybody knows if I'm doing something wrong here?

The printer is running the latest firmware as of today - v6.1.3, released in September 2024.

I've started my next project, which is to design and print myself custom shoes. I'm gonna start with something simple: pool shoes.

Pool shoes are typically designed to be close-fitting and put as little material around the foot as possible. So they're ideal to iterate through the design of my shoes to find the perfect fit without wasting too much filament. And I need a new pair anyway...

I ordered a bunch of TPU with different Shore hardness and the rolls have arrived. So now I'm designing the shoes.

I have no experience with TPU, but a colleague at work does: he told me he tried to print a remote-control-like flexible sheet of rubber buttons that was 2mm thick, and the sheet readily delaminated when he pulled on it.

I want my pool shoes to be as thin as possible - apart at the front where I need protection. So bearing in mind what my colleague told me, I opted for a wall thickness of 2.5mm. Do you think this is enough?

As for supports, the printer I use at the moment only has one extruder head (it's not mine) so I'll have to print them out of TPU too. Is there a good strategy to limit the amount of material used and maintain structural integrity on such a large print? I'm thinking of drawing the supports myself inside the shoes to control exactly where they will be and limit wastage.

Well, not really sunglasses, but rather clip-on shades for my 3D-printed glasses

But here's the thing: they're FULLY 3D-printed. The "lenses" are in fact the finest and thinnest mesh I could print with our printer - basically one 0.1mm layer of 0.4mm lines spaced 0.4mm running horizontally, and an identical layer of lines running vertically right on top of it.

Is it perfect? No. The image through it is kind of "pixelated" But it's surprisingly acceptable. It looks like this when looking through them:

In real-life, it's quite a bit darker than this. But the photo shows fairly accurately how it looks like seeing through them.

It works because the mesh is very close to the eyes and totally out of focus. And although it's not optical-grade, the price is unbeatable 🙂

If you want to try printing it yourself, the model is here. It's meant to be printed with a 0.4mm nozzle and a 0.1mm layer height - including the first one.

Double-check how the slicer slices the first two layers, where the mesh lives, because it easily tends to "simplify" the lines by not printing them, which is obviously not what you want.

EDIT: as others have pointed out in this thread, don't use these shades as actual sunglasses without sticking some UV filter over the mesh on the inside. They're not eye-safe as-is. I made them more for the challenge of making them than anything else.

I designed a part that has compartments for small neodymium magnets. The compartments open up on the side of the part, but ideally the entire part should look smooth and featureless, and the epoxy I use is not the same color as the PLA. Also, I'd like to be able to fish out the magnets later, and epoxy is a bit too final for my taste.

So I'm thinking of dropping a small dollop of melted PLA into the openings to seal them, then file / polish them smooth. It would be sticky enough to hold the magnets in place yet easy to pop off with something pointy or sharp if need be.

And to do that cleanly, I figured I'd get me one of those cheap freehand 3D pens as a kind of precision "glue gun" for PLA. And it occurs to me that I might also be able to use it to "weld" small parts together, and hand-write things on parts with a different color filament.

I'm not much of an artist so I have no use for a 3D pen as an artsy tool. But it seems like a useful thing to have alongside a 3D printer, and they're not that expensive - even the more expensive Mynt3D 3D Pen Pro, which is the one I'm eyeing.

Does anybody know if those 3D pens can be used for small manual reworks / assembly of PLA parts?

This is 7 years old, but what a fabulous project!

You might recall that I tried printing this fake lens last week and found it quite challenging.

So today for S&G, I figured I'd try to print it in two halves like so:

Then instead of gluing or epoxying them, I joined them with acetone: it's quite a thin profile, so it's a good test of the strength of the bond.

Well, I let it cure for half an hour just to make sure all the acetone inside the bond had evaporated, and it certainly is plenty strong! Strong enough to take a vigorous chemical polishing - with acetone also - while applying a fair amount of pressure and stay in one piece.

I polished it by hand for 5 minutes and it's even a bit lens-y now 🙂

I'll let it cure for another 24 hours for good measure, then tomorrow I'll mount it in the lathe and I'll polish it some more with acetone, to see how clear I can make it.

It's never going to be any good of course, but it's interesting to see how much it can be improved, if only to find out what I can do with PLA.

I found a small length of filament in the parts bin. I don't know what it is, and nobody else here does, or remembers ordering it. I'm pretty sure it's a sample that was sent by Prusa when we ordered the printer, and it's probably not a special material.

It's feels "gummy" and a lot softer than PLA, but not really rubbery either. And I tried printing something with it at 230C as if it was PLA and it's clearly not hot enough: it's able to flow out of the nozzle but it barely sticks to the bed.

Any idea what it could be?

In the process of refining the design for my 3D printed glasses (yes, I'm still at it 🙂) and trying to streamline the lens ordering process a bit, because some folks have told me their optician, or their optician's lens supplier, didn't really want to mess with "unusual" things like this, I decided to draw and print a fake ophthalmic lens. You know, not optical-quality - or even see-through - but something that looks and feels like a lens, that can be mounted in my frames, to show an optician hands-on how it works and that it's not weird or delicate to work on.

I figured it would be a quick print in clear PLA, that would require only a bit of cleanup and mount right into the frames, complete with the bevel and the notch and everything. How wrong I was...

This part is almost impossible to print right:

• It's modeled after a real bispherical lens with an offset optical axis. I mean it's optically incorrect, but it has everything a real lens would have: uneven edge thickness, one convex and one concave side.

  You just can't set it flat on the bed in the slicer: none of it sits flat. The slicer has trouble generating support around the edge on the concave side that it interprets half of as overhang, and even if the beginning of the support doesn't get ripped off by the head and the print completes, the surfaces will be absolutely awful.

• If you print it vertical - which frankly is the least bad option - then the bottom of the lens, under the support, will be a complete mess. The bevel won't even be visible. It takes quite some time to create supports that won't mess up the bottom of the lens too much

• If you print it in clear PLA with only perimeters, it'll be transparent enough to see the spots were the perimeters are started at the next layer. And depending on your wall generation strategy, the spots where the printer tries to fill the voids will show up as round "halos" inside the lens.

  The lens' thickness varies everywhere, so the slicer tries its best to fill each layer, but it's slightly different at every layer. At any rate, it reveals the slicer's idiosyncracies in tool path stragegy right away.

• The bevel all around the lens is only 0.5mm high. If the printer is dialed in, the bevel will show as an actual 120-degree bevel near the front face of the lens, particular where the bevel is normal to the layer, about half-way up the lens if you print it vertically. If not, it will look like a barely-raised blob.

I've tried this print on 3 different printer and I just can't get a decent print. All I can do is play with the settings in the slicer until it comes out not too terrible.

So, that part is a bit frustrating for my original purpose, but it turns out to be a great part to test a printer and/or the slicer software! If you want to try it too, it's here:

https://github.com/Giraut/3D-printed_eyeglasses/raw/refs/heads/main/lens_example.3mf