Go the other way and buy 3mm tubing and see what it does :)

Fair enough bowden has a lot of issues like filament compression (which can’t be fixed with tubing) making it difficult to maintain a steady flow when conditions aren’t steady (e.g. acceleration and so on).

With larger diameter tubes the issue of filament compressing gets worse. In a nutshell, a larger tube diameter for “rigid” materials somewhat is similar to the effect of a softer filament/material.

With direct drive, I would say it doesn’t make a difference.

For bowden there is a benefit.

I also think that direct drive is superior to Bowden setups in nearly every situation. The current generation like the Orbiter v3 is a very light direct-drive extruder.

Don’t buy anything you can’t afford. Sounds like you already got that part right. :)

If so good 3D-printer are lower cost than ever. Sure you could pick up an used Ender 3 and the like for $50 and put in the elbow grease to learn like it was done in the old days. Alternative is spending $200 on an BambuLab A1 and skip most of the learning curve.

If you can handle frustration than a cheapo used printer could be a good option. Please ask before purchasing as not every $50 printer is a good deal/option and as a newbie you can’t distinguish the good from the bad.

Use ASA. PETG will creep.

ASA will creep significantly less. If you have fibre-filled materials they also excel in this regard but are probably overkill here.

PETG will creep even at temperatures below 80°C and the higher the temperature is the faster the process is: https://en.wikipedia.org/wiki/Creep_(deformation)

I have a small collection of parts that all experienced creep (as a showcase). Structurally they where all fine in simulation and practice. Over time they all failed due this deformation.

JLC3DP

For production parts look at other materials. Nylon might still be within the range of your 3D-printer.

For attaching: If you can’t clamp then look at gluing it in place. If you need it detachable maybe drill a hole in the shaft and use a R-clip.

You can also tap and die the shaft on top. Make sure to use a left or right-handed thread depending on the rotation/load (thread is so that the rotation/force isn’t unscrewing it).

What is the plastic of your choice?

PLA, PETG, … behave like a very slowly flowing liquid as such the print will deform/expand (creep).

Interesting 3D printer design.

Using 3 linear rails to mount the bed. I suppose each linear carriage is one “corner”/point and all three of them to define a plane without overconstraining the buildplate but you keep all of the rigidity of the linear rails?

The issue I see is that the three linear rails in one plane are overconstrained so it is still is a pain to setup (especially on an ender 3 like frame construction).

What is the reasoning for using two motors instead of one powerful in the middle?

Heat set inserts are an interesting topic. You don’t need them if the screw is only installed once. Sure enough you can drive a metric screw into a 3D-prind and form/cut the threads but better are self tapping plastic screws. For nut vs. insert: It depends. if you can get away with a square nut (e.g. DIN 557 | don’t use hex nuts like DIN934) they are quicker to install. Make the cutout slightly undersized so the nut is fixed/wont move when the screw is installed.

If you use inserts consider the flange type to get a very nice aesthetic: e.g. https://de.aliexpress.com/item/1005006135129074.html

Removable plastic rivts are also a great fastening option. Push pins with integrated spring are equally briliant. Parts with snap fits are awsome but very diffuclt to design as the tab needs to flexible to be pushed in place but still strong enough to not break (layer adhesion).

Warping is indeed a big selling point for these materials. Major selling point/driving force I belive is still the esthetics of them as they provide a surface finish which can be sold (without post processing) without looking like a FDM 3D-print. Less warping, improved properties is more of a nice to have for most applications.

don’t seem to be, at least at the hobbyist level

This can be sad about a dozens of constructions. ISO1234 /DIN91/GB91-2000 splints are everywhere in the industry but nobody in the hobby space knows about them.

It’s pretty common to see cheap bearings in 3d printed parts, actually mildly interesting to me that bushings don’t seem to be

For bushing vs. bearing: It always depends on the application and industry.

. Do wonder though about the wear of 3d printed bushings, surfaces won’t be smooth,

Talked a while ago to somebody that run studies on FDM printed bushings. The verdict was that print orientation/layer adhesion was a limiting factor with their setup.