Ideas for my next project
I've mulling over ideas for my next project. I'm currently leaning toward integrating mobile devices into Filament DB. You'll be able to use your phone to scan QR code labels or NFC tags on spools or dry boxes and have that immediately sync into PrusaSlicer. But I'm open to other ideas as well.
So far here are the projects I've presented in under this section
- Filament DB - a Next.js + Electron desktop and web app for managing 3D-printing filament profiles, spools, and inventory — with NFC tag read/write (OpenPrintTag + Bambu Lab) and PrusaSlicer/OrcaSlicer/Bambu Studio integration.
- PrusaSlicer Filament Edition - Adds built-in calibration tools (temp tower, flow rate, pressure advance, retraction, fan speed, extrusion multiplier, dimensional accuracy, max flow rate), live bed mesh visualization with USB probing, XY skew correction, virtual filaments for multi-color mixing, and FilamentDB integration for syncing filament presets across machines.
- FDM Polymers - A Technical Reference is a ~126-page engineering handbook covering the chemistry, mechanical/thermal envelopes, calibration guidance, post-processing limits, and brand surveys for the polymer families (PLA, polyesters, styrenics, polyolefins, polyamides, polycarbonates, and high-performance specialty filaments) used in engineering-grade FFF/FDM 3D printing.
I've been continuously improving the first two and the third I will add a section on solvents and adhesion (thanks to a suggestion for @conrad). I will most likely also make the FDM Polymers document a wiki so that others can better contribute.
As for the PrusaSlicer integration, I have no clue what if any of what I've implemented will show up in PrusaSlicer 3.0 except for the virtual filament color mixing which I'll most likely remove in favor of Prusa's implementation.
RE:
Just for one (AFAIK) unsolved problem, layer height in multiple-nozzle prints. E.g. 0.6 mm vs 0.2 mm gives a by-three speedup since it can do three narrow lines at the same time. But it could be the square of that - printing three layers at the same time.
I've been thinking about slicing a project twice for each nozzle then combining layers in G-code by postprocessing. I don't think there is a fundamental obstacle, but so far I don't have any print project interesting enough to make it worth the trouble. Once INDX is online, I'd expect there is an audience for this approach.
Recently, I've been toying around with a 0.1 mm nozzle - lots of "but"s e.g. its sensitive to abuse and dust - and only buying cheapest filament doesn't exactly help. But in principle, it works reliably (0.2 mm is my "go-to" size for detail, ordered 0.15). And for that, "N²" vs "N" makes a massive difference.
RE:
@mnentwig I thought about this a lot as well which is why I purchased a 0.2, 0.8 and 1.0mm nozzles for the INDX in anticipation of experimenting with this.
With 2026 being the year of the tool-changers. It's only a matter of time before someone implements this in Orca. There is the blog post Prusa did a year ago (?) on how to hack PrusaSlicer to get something like this to work on the XL? I need to go back and reread it. I sure though it probably kept the layer height constant.
RE:
My approach would be as follows, a file with multiple objects (annotations in G-code in the "cancel object" framework):
- slice with (e.g. 0.6 mm) nozzle into file a
- slice with (e.g. 0.15 mm nozzle) into file b
- python script to collect layer G-code into a hash per file by object name, then another hash by (top) Z
- pick which layers I want to print with which nozzle = which file
- print in ascending Z, combining adjacent-tool layers for speed.
That actually comes fairly close to a REQUIREMENTS.md I'd throw at Claude if I could use the corporate subscription for fun stuff... not yet complete but it'll ask questions, fill in the gaps and write the agreed requirements first and then the code. And this is already meta-metaprogramming when people are trying to come to grips with the insanity of the former, what, you just tell the computer to write its own code ?... no, we're beyond that stage...
RE: Ideas for my next project
I code for a living and have been doing it for many, many years. So I like to think I'm good (at least they pay me enough to think I am). Anyway, before LLM's I almost never finished a home project because it was too much of a time sink and I was burned out from coding for 8-10 hrs at work. Without Claude I just wouldn't have had the patience to finish the projects I started this topic with.
That said, because I'm a professional developer I can look at the code it produces and know when it's hallucinated garbage. I occasionally then have to step in and correct it.
@mnentwig I have a personal Claude Max subscription. If you want to put a spec together I can start something.
RE: Ideas for my next project
Wanted to add that I would prefer to do this in C++ as an extension to PrusaSlicer. Python post processing is good for some tasks but it slows things down otherwise.
RE: Ideas for my next project
Long-term, absolutely... I'm still thinking along the lines of proof-of-concept validation, lots of assumptions in there. I'd first go for the most straightforward tools - and of course, slicing once per nozzle is a horrible hack. But no point in polishing the implementation if it fails conceptually on the printer.