3D Printer? - Githubissues

ezorita commented 3 years ago

Hi Scott, this is a fantastic project. Thanks for your contribution.

Not sure if this has been asked before. Is there an easy way to export the laser cut model to 3d printed parts?

lastcoolnameleft commented 3 years ago

@ezorita I have been working on exactly this. It is possible; however, there's some gotchas that I haven't quite solved yet. My spool doesn't seem to spin as loose as the laser cut ones. I'm sure I could change some of the dimensions, but it's quite time consuming. I've modified his splitflap.scad so that I can 3d print the flaps, spool and frame. Let me know if you're interested and I can share with you my findings so far.

And yes, this is an exciting project and I think some others have tried 3d printing it, but I haven't seen any final product of it yet. I feel like I'm close, but the tolerances are really low.

ezorita commented 3 years ago

hey @lastcoolnameleft it's great to hear you already made some progress on this! I did start to adapt the different frame parts to 3d printable objects.

It would be great if we share our progress, I'll try to improve upon the printing tolerances. Would you like to fork the repo so that we got a place to contribute and eventually open a PR back to @scottbez1?

dmadison commented 3 years ago

For what it's worth the flat laser-cut design should work (nearly) out of the box for 3D printing even though it's not designed for it. Within the splitflap.scad file you need to modify:

render_3d = false;
kerf_width = 0.5;  // spacing between parts

And you'll get the entire 3D design packed flat which you can export as an STL for printing.

The one exception is that one of the spool struts is placed on top of the right side panel so it's cut out of it on the laser. You'd have to reposition the strut somewhere else.

scottbez1 commented 3 years ago

Yep, thanks for the summary, Dave!

By the way, a lot of the infrastructure needed to formally support 3d-printable exports should be there already; if someone wanted to create a script for it and send a PR that'd be really sweet!

I think it would just require a small tweak to projection_renderer.py to pass in the file extension to use when rendering each individual component (.stl instead of .svg), and then would need a new wrapper script akin to generate_2d.py (it would be simpler though, since you don't need to do any of the post-processing or merging that's done for the SVG files). Since that exports each component individually already, the spacing/layout becomes irrelevant and you can lay out the individual STL files yourself using your plater/slicer of choice.

The only thing that I think would be kind of weird is the engravings, which are exported separately from the body they're engraved onto. That approach wouldn't play nice with an STL export, but the current engravings are also probably too small to be worthwhile on a 3d print, so it may be best to just skip engravings initially to keep things simple.

lastcoolnameleft commented 3 years ago

Instead of having 3d printed parts like laser-cut parts, I wanted to build an integrated solution that let you toggle flags to render the specific piece you want. (e.g. instead of 4 walls for the frame, render the full frame)

I hope to have it better documented, but here's my attempt at a 3d printed model: https://github.com/lastcoolnameleft/splitflap/blob/3d-print/3d/splitflap.scad

The changes I made are to add a number of "render_xxx" flags to make it easier to export combined pieces instead of individual parts. For example setting render_spool and render_spool_motor_wheel to true will print one side of the spool as one piece and render_spool_struts as the other.

I also added render_frame and render_captive_nuts flags to render the frame as 1 assembled piece.

The issue I hit was that it just added to the code complexity and assembling everything became a lot trickier. One idea was to have individual pieces split up into separate files similar to how there's a pcb.scad which would make the code more composable. Then there could be a "laser_cut.scad", "3d_print_frame.scad", etc.

I'm definitely open to ideas.

dmadison commented 3 years ago

How divergent are you imagining the 3D printable design to be?

If it only differs mildly from the laser-cut design (e.g. hiding captive nuts) then a few flags might be fine, coupled with its own module for all of the parts in 3D-printable orientation. Anything more than that and it might be worth splitting the 3D printed version into its own design/file just to keep things from getting too disorganized. It would probably make more sense to rework the splitflap.scad file so that the 2D render is conditional; that way it can be include'd in other designs and you can reference off of all variables without having to expose them as functions.

I do think the separate flags for each part is overkill. A script could be used to automate exporting the individual parts, probably by iterating through child objects like the projection renderer.

lastcoolnameleft commented 3 years ago

Great question. I initially thought it wouldn't be very divergent. However, I'm hitting concerns that I don't have an answer to (e.g. If the frame is all 1 piece, how do you fit the spool + gear into the frame. Currently, I brute force it.)

I'm still learning the repo and poked around the /scripts and it appears that openscad.run() lets you override existing variables to generate STL's. Is that accurate?

I looked into creating a render_for_3d_printing() function which sets all of the appropriate variables and overrides to generate the appropriate pieces, but OpenSCAD variable scoping didn't seem to allow for that. This is my first OpenSCAD project, so let me know if there's an option I've missed.

scottbez1 commented 3 years ago

Yeah, the python scripts are able to override variables using the -D command line option (https://en.wikibooks.org/wiki/OpenSCAD_User_Manual/Using_OpenSCAD_in_a_command_line_environment ), whereas that kind of behavior won't be possible from within another scad file without first refactoring splitflap.scad to wrap more things in high level modules. The global variable situation is not great (but in my defense this was also my first OpenSCAD project :stuck_out_tongue: so I didn't have a great sense of OpenSCAD architectural patterns). It's definitely in need of some refactoring. Unfortunately I can't commit to that myself at the moment, but if someone does take on that heroic task I'd very much appreciate if it's done in small increments and separate from any functionality changes for the sake of reviewability.

That said, like @dmadison mentioned, it's probably worth considering the end goal before getting too deep in the splitflap.scad weeds. The design as it stands today was very much built for laser-cutting, with that assumption baked into nearly every design decision. At a certain point it may make sense to just factor out some of the core dimension calculations (spool dimensions, window size and position relative to the motor, etc) to a library and build the actual pieces up from scratch using just that core library.

The best example of "built for laser-cutting" is probably the spool, which is currently an assembly of 7 pieces plus a bolt and a nut, when it could easily be just 2 pieces total (or 1, depending how well your printer can handle supports) if 3d printed. If you're interested in making a variant that's optimized for 3d printing, rather than just 3d printing the existing design, consider this a gentle nudge towards starting there :wink: , as I think it's the biggest opportunity to really leverage the benefits of 3d printing over laser-cutting. For instance, it should enable much smaller diameter spools, which would solve the most common question of "how do I build a number-only display" without needing to increase the gap between flaps substantially (https://github.com/scottbez1/splitflap/issues/16#issuecomment-553883245 )

sohowaty commented 3 years ago

Just to add my two cents to 3d printing - I wouldn't want to print the whole frame as-is for laser cutting - it'll require a lot of time to print when compared to simplified version. NilSkilz made a fork with models designed for 3d printers - two-part spool and one-part frame. This fork is old and I'm not sure if all dimensions are still valid when copared to 'master' version, but I printed one module and going to test it ASAP.

lastcoolnameleft commented 3 years ago

I just took a first-stab at refactoring splitflap.scad so that it would be more easily 3d-printable. I'm interested in your feedback.

https://github.com/scottbez1/splitflap/pull/139