Great article. This is all above the skill level of the average part on thingiverse or printables, but the good parts on there are going to follow similar ideas. Love the mouse ears, press-fit holes and step-by-step alignment of layers to build impossible bridges.
Notably, in fusion 360 this would all be designed in "plastics" mode, and yet that mode is oblivious to whether the part is printed or moulded. I wonder if any CAD engine can do "production-aware design" that constrains design to the capabilities of standardized machines, e.g. keeping a metal part 3-d millable. I've seen strict design rule enforcement with PCBs, and I have seen sheet metal macros, but nothing for general mechanical CAD.
I've investigated this space, and I'm not entirely sure its even a desired goal from the perspective of a mechanical designer. The benefit tends to be for smaller aspects (ensuring hole sizes are appropriate for the desired thread, or that holes aren't too close to a bend line on a sheet metal part, etc) but the final design of a 3d part is so non-deterministic, and the variety of manufacturing methods are so varied and unique, it might just cause more issues than benefits.
I've been playing with 3D printers for 7 years, and I even assembled mine at home during the pandemic. Some topics described here I already found out by practice and I think most people with experience in 3D printing also do that.
But having everything studied, compiled and explained in that level is just, again, amazing! Not only that, but there are so many other topics covered here that I still have to learn.
I know they get a lot of hate in the HN community but my Bambu Labs P1S is mind blowing. It’s so easy to use I print 100x more than with my old Ender. It’s motivated me to learn Fusion360 … i’m actually printing droids for my kids to color this very minute.
Enders were... not a great experience. I understand they were in a good price spot at the time, but from my experience and from what I gather online, very finicky. People who liked tinkering with the printer itself loved and recommended them because 3D printing became a skill of its own (Not for the design considerations in the article, but to make the equipment work consistently).
I've heard that Bambus are much better. I have a Raise3D E2 from the Ender era, and it's rock solid. A step up in price, but no finicking. Just works, when new, and now.
I know two people with that exact model of 3D printer. Both printers are routinely out of commission for weeks on end due to some failure that the owners lack either the technical expertise to diagnose and fix or the desire to pay exorbitant prices for proprietary replacement parts to fix (or both). Meanwhile my Ender 5 is always chugging along, and is never out of commission for more than a day or two while awaiting replacement parts from Amazon that cost between a few cents and up to maybe $20 each.
I don’t actually think Bambu makes unreliable printers; to the contrary, they are excellent machines that, if anything, are much more reliable on the whole than Creality. But they’re kind of like sports cars, in that their target market is either people who want something fast and flashy and are willing to throw money at any problems to make them go away, or for technical types who want something they can take out on the track and don’t mind wrenching their own machines. The problem is that Bambu printers are marketed and touted as being great for beginners, and while they certainly make it easy to get into 3D printing for nontechnical people, I think most of them will end up ultimately being disappointed at either the lack of customization they allow or amount of time, effort, and money required to diagnose and fix them when something goes wrong.
I think that conclusion is wrong, they are absolutely for beginners. No bed leveling. Lidar scan of first layer. Filament sensors. Good software. Enders are sold to beginners but you actually need to be an expert to get good results and keep them running.
My Bambu A1 just works. I had an Ender 3 before and it almost killed my interest in 3d printing because my prints constantly failed. I don’t see a path where the A1 could disappoint me.
GPL issues and concerns about the SaaS-y aspect. Folks on HN and often techy folks in general don't like it when hardware requires an internet connection vs local control. These concerns are somewhat warranted based on recent moves Bambu has made
My favorite continues to be hardware from Prusa. They're rock solid and respect user freedoms (serviceability/upgradability/hackability). Being made in the EU is also a big upside for me.
I've had an MK3S+ for years and even though it's a primitive machine in comparison to the current Bambu hardware I see no reason to upgrade to something else. It just keeps printing whatever I throw at it and the results continue to be very good. In fact, I seem to have better luck with it than the Bambus I sometimes use at various hacker/makerspaces.
If you just look at the numbers (speed, volume, ...) against Bambu hardware they're not as good, but the reliability and simplicity make up for it IMO. The main missing feature is multi-material support, but that's something I'm not really interested in due to how wasteful the current technology is.
> There is no excuse to not add text to a printed part.
Super off-topic, but I've always kind of been let down by the appearance of 3d printed text. As noted, engraved seems to be better than embossed, but it still just looks kind of weird. I envy the clean, crisp labels that seem to be commonplace on commercial injection-molded plastic parts.
The toner transfer technique seems kind of promising. I think I've also seen people spray painting 3d-printed parts, and then lasering away the paint to draw text, which is interesting (if somewhat more materials- and equipment-intensive).
I've heard people have had pretty good luck laser engraving text onto 3d prints with fiber lasers, though it is pretty steep price bump just to get some text on a 3d print
Has there been any interest in leveraging LLM's for 3d modelling? Sort of an AI assistant with CAD software, to help beginners get going and also more rapidly design simple objects.
Yes, there has been. Unfortunately, there are a few core issues blocking this from becoming a big thing:
1. The majority of 3D modeling is not done parametrically, meaning there is not a lot of data. The little data there is is generally in OpenSCAD, which isn't very powerful or extensible for useful CAD.
2. Generally, when you want to do CAD, you need to come up with a way to define everything precisely. Like I want this hole 2 millimeters from the bottom, and this exact edge next to the hole to be beveled, etc. Saying all that to an LLM is slower than just making the whole.
That said, these still can be useful for beginners, and there are things like Adam AI that are starting to catch on for simple stuff.
6 months into 3D printing and I couldnt have asked for a better article to stumble upon. What a massive field this is and I love some of the take aways. Paricularly circles into hexagons, and making things adjustable.
I’m not making my own designs yet. It is too difficult. Modifiying a little here using Blender is where Im at
It's super easy to design using OnShape. Hit me up with private message and I will show you everything you need to model 3D printable parts in under 5 minutes.
* Sketch a 2D design on a surface
* Make the elements in that design depend on each other (this is parallel to that, this is equal to the other, X is at an angle to Y) as much as possible
* Pull the 2D shape up into 3D space
Now you know how to design your own things! The rest is just learning the buttons, but there's usually one called "sketch", one called "constrain", and one called "extrude".
Incredible article, learned quite a lot. To me, a very good supplementary reading would be Structures by J. E. Gordon [1]. Helped me grasp a lot of the mechanical design notions necessary for that sort of work.
My friend and I have been getting into forge molding carbon fibre using 3d printed molds like this: https://www.youtube.com/watch?v=25PmqM24HEk. It is a great technique for making small batches of really strong parts and I'm surprised it isn't more common.
I always thought 3D printing would make multi widget machine[0] manufacturing possible
While it’s done a lot of cool stuff and enabled rapid prototyping etc it never scaled the way I really thought it would
[0]: there may be a better turn for this however this is what I mean: that is one machine that can output a wide variety of different things using the same common material, IE maybe one day it produces ball bearings and the next it could produce a bunch of car pistons, with only having to make minimal changes to the machine itself if not changing anything at all
"Flexible" or "Quick Turn" manufacturing are terms used for this kind of thing. Quick-turn comes from being able to change from one kind of part to another, quickly, with no added setup cost.
In theory, it seemed perfect for flexible manufacturing: same machine, same material, endless outputs. But in practice, it hit limits in speed, material properties, and post-processing. You still can’t print a high-tolerance metal part at scale and cost-effectively replace traditional machining. It’s amazing for prototyping or niche parts
There are companies with big print farms that offer this service. But of course it’s limited to materials that can be 3D printed, and if the product reaches a certain scale, it’s likely best to invest in injection molding or some other process.
That said, for smaller scale products, news businesses, or things where 3D printing is the only way the thing can exist, these services exist.
One technique which bears mentioning is printing in 100% infill using a filament which will allow re-heating/cooling and then putting it in a tray of powder salt (very finely ground table salt) and then backing and cooling it.
Nice article, though what I'd personally love to see is a resource where I can go from zero to actually making (basic) designs using open source tools, which can then be taken to a 3D printer and printed.
The learning curve was steep, but FreeCAD has allowed me to start playing with 3d printing gears and other things on my Bambu Lab P1S. I'm largely self taught with electronics and programming, so just starting and making small experiments got me going. For inspiration, there are lots of sites that share 3d print designs.
This article reminds me of another I read first here, 'Reality Has A Surprising Amount of Detail' by John Salvatier. At first blush 3D printing seems easy, but especially with smaller parts that might go through many duty cycles it's anything but. I'm going to have to do more than skim this, I think this one is worth multiple reads over many days to really absorb the densely packed information.
Thanks to the author for being willing to put so much of their hard-earned experience into a resource for the rest of us.
This is fantastic-- while I'm aware of most of the techniques in it, it would have saved me a ton of time and trouble if I had it a few years ago.
Each of the points could basically be expanded to an article on their own. E.g. they don't mention for vase mode that you can get much better results using a big nozzle with it.
- Solvespace --- small and lightweight, the UI may be a bit off-putting
- FreeCAD --- hugely improved in the recent 1.0 release, this is a large and impressive system
- Dune 3D --- the new kid on the block, it has the advantage of a modern appearance and UI standards, and the consistency of being a one-man project
If one moves away from traditonal/contemporary CAD there are a few other options:
- BRL-CAD --- intensely old-school, this is one of the oldest opensource codebases
- OpenSCAD --- programmatic CAD, this has inspired more successors than I would care to count (esp. look up libfive and Matt Keeter's Master's Thesis if you are academically mathematically oriented)
For that last, one of the more successful hybrids is "OpenPythonSCAD" which is just what it says on the tin --- Python in OpenSCAD:
OpenSCAD is an underrated but powerful modeling tool, especially for developers and engineers who appreciate precision and code-driven design. It has a low barrier to entry — the syntax is simple, yet expressive — and with just a bit of practice, you can build tight, parametric models that are incredibly robust.
One of its standout features is the `hull()` function, which computes the convex hull of multiple shapes. When used skillfully, `hull()` becomes more than a geometric operation — it’s a design primitive that lets you smoothly bridge components, create enclosures, and generate complex organic forms without manual sculpting. It's like having a smart “connective tissue” for your model.
If you're comfortable with code and want exact control over your 3D prints or CAD designs, OpenSCAD delivers precision with minimal overhead. It rewards clean thinking and composability — making it ideal for rapid prototyping, parametric part libraries, and even mechanical design.
Learn FreeCAD. Getting trapped in commercial software and having to abandon years and years worth of project files isn't a mistake I'm making twice. Fusion seems attractive, but look at how they treat their shit tier users.
Onshape is amazing. The learning curve is much more forgiving than other software while still being a feature-rich, optionally constraint-based and parametrizable CAD application. It works on any OS, even on a laptop with an iGPU, a Chromebook, and for basic stuff like exporting a part for printing, a phone.
Consider signing up via your favorite YouTuber's sponsorship link to support them.
Downsides are that the CAM plugin is paid-only (irrelevant for 3D printing) and you're obviously trapping yourself in a commercial, proprietary walled garden that might start charging subscription fees or otherwise rug-pull you once it gets popular enough. I've decided that the ease of use benefit is high enough to warrant the risk - I'd rather risk not being able to edit my models in the future than not creating them in the first place because the alternative software is too painful to use.
It's helpful to understand how the software works, because it's different from what you might have experienced from other software: It essentially stores operations, like "start with this sketch, then extrude this part of it to a height of 10 mm, then add a fillet". You can go back and edit previous steps and the following steps will be directly re-applied.
In sketch mode, you can just draw, but you can also add arbitrary constraints, e.g. "these points have to be exactly 3 cm away" and it will adjust your sketch to match the (new) constraints. This makes it really easy to change some aspect of the part later. This is common in CAD software, although OnShape's implementation seems more intuitive to me than e.g. Fusion 360.
If you want to do actual 3D CAM (for CNC machining), Fusion360 seems to be the only free option (not available for Linux).
In general, with all CAD software, the common "just poke at it until you figure out how it works" approach doesn't work well, although once you've understood the basic concepts that I've explained above and know some CAD terms/concepts like creating 3D parts by extruding or rotating 2d drawings, Onshape will mostly let you get away with that approach. You probably should still watch tutorials before you start.
Is there any realistic free alternative for 3D (not 2.5D) parts?
You certainly won't want to use it for mass production, but for hobbyist use where getting the model and CAM config right, setting up the machine etc. are the biggest time sink and most parts are made in quantity 1, I found it acceptable.
FreeCAD has a built-in CAM. It's not very powerful, but it's only going to get better with time (while the proprietary alternatives will only continue to get worse as companies try to squeeze money out of their users).
> does anyone has any good tips for a complete Newbie where to begin?
Start with Tinkercad: https://www.tinkercad.com. It runs on the browser, it has some limitations, but it is really simple to use, just open and model whatever you want joining and extracting shapes and importing SVGs for extrusion.
After that, if you know any programming language you'll find OpenSCAD easy to learn. I gave a course last year about it, the slides are available here: https://lucasoshiro.github.io/posts-en/2024-03-24-openscad/. They are in Portuguese, if someone shows interest I can translate them to English, but I think they are easy to follow even by non-speakers.
I've been a newbie too and tried to use FreeCAD as others mentioned but I found myself enjoying build123d (basically a python library that uses an long-existing technology called OpenCascade and a viewer called OCPViewer generally used within visual studio code).
The learning curve is still there, but I felt more empowered to adjust/share 3d printing designs made in it over dealing with quirks of GUI-based CAD applications. The discord community on there is rather helpful too.
I'll still use FreeCAD on occasion as a secondary viewer for stl files, though my hope is to use build123d entirely including for describing joints as well.
BTW there is an open source project on GitHub named 'Mayo' which is a pretty incredible viewer for 3d files including most CAD formats. 'F3d' is another great viewer. Both are cross platform.
I just got started recently with OpenSCAD - it's a different beast, but very useful for simple parametric designs. You write code to describe the form of your object - no clicking and dragging things at all.
I use Fusion 360. Free for hobbyists. Yeah it's quirky and they constantly screw the free plan out of features (e.g. less saved editable designs, having to use the cloud to export STL) but it is also a highly capable tool that aligned best with the stuff I already knew.
Not entirely sure if it's available for Linux.
I probably shouldn't use autodesk but I'm not trying to make the world a better place. Just to unleash my creativity.
Ah I see. I've been looking at FOSS options like FreeCAD and Blender but both didn't feel right (especially blender as it's more a tool for animators).
And I rather spend my limited free time creating stuff than to learn a new tool. Unless it is actually a more powerful one for the purpose that enables me to do things I can't now. But this doesn't seem to be the case.
It's the same reason I use BambuLab printers. My hobby is making stuff, not tinkering with printers. They're just tools, a means to an end.
Ps forgive me my defensive attitude but I often get people at the makerspace that take my choice of tools as a political statement. But I don't care. I just want to use what does the job for me.
I can't vouch for this, but maybe you could get SolidWorks working in Wine? (e.g. https://github.com/cryinkfly/SOLIDWORKS-for-Linux). Of note, SolidWorks is cheap if you're a student or veteran, for a non-commercial license. It is a dramatic improvement over FreeCAD. (I wish CAS were in a state like EDA and artistic model makers where the free/OSS software was on par with commercial, but we are not.)
I use FreeCAD, but it definitely leaves some UX refinement to be desired. There are a couple of web based options like OnShape that seem to work well, too.
OnShape is great (we have been using it exclusively for a project over the past four months, the collaboration tools are phenomenal), but FreeCAD has made some fantastic progress over the past year. Some of the underlying technology problems have solved, and the UX has improved a lot with 1.0. The customization and scripting opportunities are also wonderful with FreeCAD. That said, if you’re coming over from Solidworks/NX/Inventor, as much as there are buggy parts of those, FreeCAD still has extremely frustrating workflows and buggy parts that you have to work around. It feels like it’s moving closer to Blender-like quality, but it still has a long road ahead of it.
All of Solidworks, Onshape, and Freecad have a very similar operating philosophy (I believe they're all based on the same backend engine). I used onshape for a while because I found freecad unusable but recent improvements solved most of those issues and now I prefer freecad.
Great article. This is all above the skill level of the average part on thingiverse or printables, but the good parts on there are going to follow similar ideas. Love the mouse ears, press-fit holes and step-by-step alignment of layers to build impossible bridges.
Notably, in fusion 360 this would all be designed in "plastics" mode, and yet that mode is oblivious to whether the part is printed or moulded. I wonder if any CAD engine can do "production-aware design" that constrains design to the capabilities of standardized machines, e.g. keeping a metal part 3-d millable. I've seen strict design rule enforcement with PCBs, and I have seen sheet metal macros, but nothing for general mechanical CAD.
I've investigated this space, and I'm not entirely sure its even a desired goal from the perspective of a mechanical designer. The benefit tends to be for smaller aspects (ensuring hole sizes are appropriate for the desired thread, or that holes aren't too close to a bend line on a sheet metal part, etc) but the final design of a 3d part is so non-deterministic, and the variety of manufacturing methods are so varied and unique, it might just cause more issues than benefits.
Amazing. Again: amazing!
I've been playing with 3D printers for 7 years, and I even assembled mine at home during the pandemic. Some topics described here I already found out by practice and I think most people with experience in 3D printing also do that.
But having everything studied, compiled and explained in that level is just, again, amazing! Not only that, but there are so many other topics covered here that I still have to learn.
Great work, thank you!
I know they get a lot of hate in the HN community but my Bambu Labs P1S is mind blowing. It’s so easy to use I print 100x more than with my old Ender. It’s motivated me to learn Fusion360 … i’m actually printing droids for my kids to color this very minute.
Enders were... not a great experience. I understand they were in a good price spot at the time, but from my experience and from what I gather online, very finicky. People who liked tinkering with the printer itself loved and recommended them because 3D printing became a skill of its own (Not for the design considerations in the article, but to make the equipment work consistently).
I've heard that Bambus are much better. I have a Raise3D E2 from the Ender era, and it's rock solid. A step up in price, but no finicking. Just works, when new, and now.
I know two people with that exact model of 3D printer. Both printers are routinely out of commission for weeks on end due to some failure that the owners lack either the technical expertise to diagnose and fix or the desire to pay exorbitant prices for proprietary replacement parts to fix (or both). Meanwhile my Ender 5 is always chugging along, and is never out of commission for more than a day or two while awaiting replacement parts from Amazon that cost between a few cents and up to maybe $20 each.
I don’t actually think Bambu makes unreliable printers; to the contrary, they are excellent machines that, if anything, are much more reliable on the whole than Creality. But they’re kind of like sports cars, in that their target market is either people who want something fast and flashy and are willing to throw money at any problems to make them go away, or for technical types who want something they can take out on the track and don’t mind wrenching their own machines. The problem is that Bambu printers are marketed and touted as being great for beginners, and while they certainly make it easy to get into 3D printing for nontechnical people, I think most of them will end up ultimately being disappointed at either the lack of customization they allow or amount of time, effort, and money required to diagnose and fix them when something goes wrong.
I think that conclusion is wrong, they are absolutely for beginners. No bed leveling. Lidar scan of first layer. Filament sensors. Good software. Enders are sold to beginners but you actually need to be an expert to get good results and keep them running.
My Bambu A1 just works. I had an Ender 3 before and it almost killed my interest in 3d printing because my prints constantly failed. I don’t see a path where the A1 could disappoint me.
as I said, as a Bambu owner, i’m really impressed with mine and highly recommend them to others.
Hate? I missed this. Why hate?
GPL issues and concerns about the SaaS-y aspect. Folks on HN and often techy folks in general don't like it when hardware requires an internet connection vs local control. These concerns are somewhat warranted based on recent moves Bambu has made
More than that. They tried to gaslight people after people found out the changes Bambu Lab was making.
Non-compliance with GPL and other opensource licensing.
Predatory licensing agreements and cloud software which presumably allows the company to access/steal designs.
What are some alternatives? Ty in advance for any hint!
My favorite continues to be hardware from Prusa. They're rock solid and respect user freedoms (serviceability/upgradability/hackability). Being made in the EU is also a big upside for me.
I've had an MK3S+ for years and even though it's a primitive machine in comparison to the current Bambu hardware I see no reason to upgrade to something else. It just keeps printing whatever I throw at it and the results continue to be very good. In fact, I seem to have better luck with it than the Bambus I sometimes use at various hacker/makerspaces.
If you just look at the numbers (speed, volume, ...) against Bambu hardware they're not as good, but the reliability and simplicity make up for it IMO. The main missing feature is multi-material support, but that's something I'm not really interested in due to how wasteful the current technology is.
Thanks for the throughout reply!!
> There is no excuse to not add text to a printed part.
Super off-topic, but I've always kind of been let down by the appearance of 3d printed text. As noted, engraved seems to be better than embossed, but it still just looks kind of weird. I envy the clean, crisp labels that seem to be commonplace on commercial injection-molded plastic parts.
The toner transfer technique seems kind of promising. I think I've also seen people spray painting 3d-printed parts, and then lasering away the paint to draw text, which is interesting (if somewhat more materials- and equipment-intensive).
Really cool article though.
Another option is water slide decal. It can give a really seamless look, but is time consuming and expensive.
I've heard people have had pretty good luck laser engraving text onto 3d prints with fiber lasers, though it is pretty steep price bump just to get some text on a 3d print
Also useful to turn spheres into two parts you can screw one with the other, like in this design of mine: https://makerworld.com/it/models/99223-death-star-christmas-...
Has there been any interest in leveraging LLM's for 3d modelling? Sort of an AI assistant with CAD software, to help beginners get going and also more rapidly design simple objects.
Yes, there has been. Unfortunately, there are a few core issues blocking this from becoming a big thing:
1. The majority of 3D modeling is not done parametrically, meaning there is not a lot of data. The little data there is is generally in OpenSCAD, which isn't very powerful or extensible for useful CAD. 2. Generally, when you want to do CAD, you need to come up with a way to define everything precisely. Like I want this hole 2 millimeters from the bottom, and this exact edge next to the hole to be beveled, etc. Saying all that to an LLM is slower than just making the whole.
That said, these still can be useful for beginners, and there are things like Adam AI that are starting to catch on for simple stuff.
There are AI models that can generate 3D models, e.g. Hunyuan3D. Not quite CAD models, but maybe this could eventually be adapted to that use case.
Then there's the possibility of an agent automating an actual CAD program. This has already been done with game dev, e.g. Unity MCP.
Just last week: https://willpatrick.xyz/technology/2025/04/23/teaching-llms-...
Yeah, tons, there are already products like this in use
Can you name them?
6 months into 3D printing and I couldnt have asked for a better article to stumble upon. What a massive field this is and I love some of the take aways. Paricularly circles into hexagons, and making things adjustable.
I’m not making my own designs yet. It is too difficult. Modifiying a little here using Blender is where Im at
It's super easy to design using OnShape. Hit me up with private message and I will show you everything you need to model 3D printable parts in under 5 minutes.
Agreed, you need to know three things:
* Sketch a 2D design on a surface * Make the elements in that design depend on each other (this is parallel to that, this is equal to the other, X is at an angle to Y) as much as possible * Pull the 2D shape up into 3D space
Now you know how to design your own things! The rest is just learning the buttons, but there's usually one called "sketch", one called "constrain", and one called "extrude".
Incredible article, learned quite a lot. To me, a very good supplementary reading would be Structures by J. E. Gordon [1]. Helped me grasp a lot of the mechanical design notions necessary for that sort of work.
[0]: https://archive.org/details/StructuresOrWhyThingsDontFallDow...
My friend and I have been getting into forge molding carbon fibre using 3d printed molds like this: https://www.youtube.com/watch?v=25PmqM24HEk. It is a great technique for making small batches of really strong parts and I'm surprised it isn't more common.
I always thought 3D printing would make multi widget machine[0] manufacturing possible
While it’s done a lot of cool stuff and enabled rapid prototyping etc it never scaled the way I really thought it would
[0]: there may be a better turn for this however this is what I mean: that is one machine that can output a wide variety of different things using the same common material, IE maybe one day it produces ball bearings and the next it could produce a bunch of car pistons, with only having to make minimal changes to the machine itself if not changing anything at all
"Flexible" or "Quick Turn" manufacturing are terms used for this kind of thing. Quick-turn comes from being able to change from one kind of part to another, quickly, with no added setup cost.
In theory, it seemed perfect for flexible manufacturing: same machine, same material, endless outputs. But in practice, it hit limits in speed, material properties, and post-processing. You still can’t print a high-tolerance metal part at scale and cost-effectively replace traditional machining. It’s amazing for prototyping or niche parts
"You still can't print a high-tolerance metal part at scale and cost-effectively..."
Dan Gelbart has a response (with caveats)
https://www.youtube.com/watch?v=kLgPW2672s4
oh wow - that's cool! - Thanks so much for sharing!
There are companies with big print farms that offer this service. But of course it’s limited to materials that can be 3D printed, and if the product reaches a certain scale, it’s likely best to invest in injection molding or some other process.
That said, for smaller scale products, news businesses, or things where 3D printing is the only way the thing can exist, these services exist.
One technique which bears mentioning is printing in 100% infill using a filament which will allow re-heating/cooling and then putting it in a tray of powder salt (very finely ground table salt) and then backing and cooling it.
What is the purpose of this?
You get a solid plastic part without layer lines. This makes it about as strong as injection molded plastic.
Nice! Want to try.
Nice article, though what I'd personally love to see is a resource where I can go from zero to actually making (basic) designs using open source tools, which can then be taken to a 3D printer and printed.
The learning curve was steep, but FreeCAD has allowed me to start playing with 3d printing gears and other things on my Bambu Lab P1S. I'm largely self taught with electronics and programming, so just starting and making small experiments got me going. For inspiration, there are lots of sites that share 3d print designs.
These are some great tips. The teardrop shaped holes are a neat idea.
Those were a staple of early reprap designs.
In fact, it was such an iconic piece of early 3d printing design language that it became _the_ RepRap logo!
https://reprap.org/wiki/RepRapLogo
Then overhangs got good enough that people just started doing normal holes again. :)
This article reminds me of another I read first here, 'Reality Has A Surprising Amount of Detail' by John Salvatier. At first blush 3D printing seems easy, but especially with smaller parts that might go through many duty cycles it's anything but. I'm going to have to do more than skim this, I think this one is worth multiple reads over many days to really absorb the densely packed information.
Thanks to the author for being willing to put so much of their hard-earned experience into a resource for the rest of us.
This is fantastic-- while I'm aware of most of the techniques in it, it would have saved me a ton of time and trouble if I had it a few years ago.
Each of the points could basically be expanded to an article on their own. E.g. they don't mention for vase mode that you can get much better results using a big nozzle with it.
3D printing is fun because there's always something new to learn
What an impressive looking article (I've only skimmed it so far).
I've been meaning to try my hand at CAD and designing models to print but I haven't quite made the jump.
One thing that has given me pause is a good CAD program for Linux, does anyone has any good tips for a complete Newbie where to begin?
For traditional CAD the notable candidates are:
- Solvespace --- small and lightweight, the UI may be a bit off-putting
- FreeCAD --- hugely improved in the recent 1.0 release, this is a large and impressive system
- Dune 3D --- the new kid on the block, it has the advantage of a modern appearance and UI standards, and the consistency of being a one-man project
If one moves away from traditonal/contemporary CAD there are a few other options:
- BRL-CAD --- intensely old-school, this is one of the oldest opensource codebases
- OpenSCAD --- programmatic CAD, this has inspired more successors than I would care to count (esp. look up libfive and Matt Keeter's Master's Thesis if you are academically mathematically oriented)
For that last, one of the more successful hybrids is "OpenPythonSCAD" which is just what it says on the tin --- Python in OpenSCAD:
https://pythonscad.org/
which I have been using for a project on the other side of the fence --- making DXF and G-code for CNC mills and routers:
https://github.com/WillAdams/gcodepreview
EDIT: One additional tool to note is Fullcontrolgcode Designer, which to bring things full-circle, is the 3D-printing version of the above:
https://fullcontrolgcode.com/
OpenSCAD is an underrated but powerful modeling tool, especially for developers and engineers who appreciate precision and code-driven design. It has a low barrier to entry — the syntax is simple, yet expressive — and with just a bit of practice, you can build tight, parametric models that are incredibly robust.
One of its standout features is the `hull()` function, which computes the convex hull of multiple shapes. When used skillfully, `hull()` becomes more than a geometric operation — it’s a design primitive that lets you smoothly bridge components, create enclosures, and generate complex organic forms without manual sculpting. It's like having a smart “connective tissue” for your model.
If you're comfortable with code and want exact control over your 3D prints or CAD designs, OpenSCAD delivers precision with minimal overhead. It rewards clean thinking and composability — making it ideal for rapid prototyping, parametric part libraries, and even mechanical design.
Learn FreeCAD. Getting trapped in commercial software and having to abandon years and years worth of project files isn't a mistake I'm making twice. Fusion seems attractive, but look at how they treat their shit tier users.
As a fellow linux users and 3D printing newbie:
- Tinkercad (browser) fun and great for very simple projects. Like the MS Paint of 3D.
- OnShape (browser) seemingly pretty powerful, but not the easiest to learn in my experience, and has some annoying bugs.
- Plasticity (desktop) I played around with the free trial and liked it a lot, found it more intuitive than OnShape.
- Womp (browser) not CAD software, but easy to use and great for making free-form/organic looking designs.
- Blender (desktop) not CAD software and haven't used it myself, but I've seen others use it to design 3D prints.
Onshape is amazing. The learning curve is much more forgiving than other software while still being a feature-rich, optionally constraint-based and parametrizable CAD application. It works on any OS, even on a laptop with an iGPU, a Chromebook, and for basic stuff like exporting a part for printing, a phone.
Consider signing up via your favorite YouTuber's sponsorship link to support them.
Downsides are that the CAM plugin is paid-only (irrelevant for 3D printing) and you're obviously trapping yourself in a commercial, proprietary walled garden that might start charging subscription fees or otherwise rug-pull you once it gets popular enough. I've decided that the ease of use benefit is high enough to warrant the risk - I'd rather risk not being able to edit my models in the future than not creating them in the first place because the alternative software is too painful to use.
It's helpful to understand how the software works, because it's different from what you might have experienced from other software: It essentially stores operations, like "start with this sketch, then extrude this part of it to a height of 10 mm, then add a fillet". You can go back and edit previous steps and the following steps will be directly re-applied.
In sketch mode, you can just draw, but you can also add arbitrary constraints, e.g. "these points have to be exactly 3 cm away" and it will adjust your sketch to match the (new) constraints. This makes it really easy to change some aspect of the part later. This is common in CAD software, although OnShape's implementation seems more intuitive to me than e.g. Fusion 360.
If you want to do actual 3D CAM (for CNC machining), Fusion360 seems to be the only free option (not available for Linux).
In general, with all CAD software, the common "just poke at it until you figure out how it works" approach doesn't work well, although once you've understood the basic concepts that I've explained above and know some CAD terms/concepts like creating 3D parts by extruding or rotating 2d drawings, Onshape will mostly let you get away with that approach. You probably should still watch tutorials before you start.
> If you want to do actual 3D CAM (for CNC machining), Fusion360 seems to be the only free option (not available for Linux).
The free CAM available in F360 has been artificially limited to only allow extremely slow travel speed. It's almost useless.
Is there any realistic free alternative for 3D (not 2.5D) parts?
You certainly won't want to use it for mass production, but for hobbyist use where getting the model and CAM config right, setting up the machine etc. are the biggest time sink and most parts are made in quantity 1, I found it acceptable.
FreeCAD has a built-in CAM. It's not very powerful, but it's only going to get better with time (while the proprietary alternatives will only continue to get worse as companies try to squeeze money out of their users).
> does anyone has any good tips for a complete Newbie where to begin?
Start with Tinkercad: https://www.tinkercad.com. It runs on the browser, it has some limitations, but it is really simple to use, just open and model whatever you want joining and extracting shapes and importing SVGs for extrusion.
After that, if you know any programming language you'll find OpenSCAD easy to learn. I gave a course last year about it, the slides are available here: https://lucasoshiro.github.io/posts-en/2024-03-24-openscad/. They are in Portuguese, if someone shows interest I can translate them to English, but I think they are easy to follow even by non-speakers.
I've been a newbie too and tried to use FreeCAD as others mentioned but I found myself enjoying build123d (basically a python library that uses an long-existing technology called OpenCascade and a viewer called OCPViewer generally used within visual studio code).
The learning curve is still there, but I felt more empowered to adjust/share 3d printing designs made in it over dealing with quirks of GUI-based CAD applications. The discord community on there is rather helpful too.
https://build123d.readthedocs.io/
https://github.com/bernhard-42/vscode-ocp-cad-viewer
I'll still use FreeCAD on occasion as a secondary viewer for stl files, though my hope is to use build123d entirely including for describing joints as well.
BTW there is an open source project on GitHub named 'Mayo' which is a pretty incredible viewer for 3d files including most CAD formats. 'F3d' is another great viewer. Both are cross platform.
I just got started recently with OpenSCAD - it's a different beast, but very useful for simple parametric designs. You write code to describe the form of your object - no clicking and dragging things at all.
FreeCAD is fine (the author also uses it). Make sure to follow the official documentation (eg. PartDesign tutorial) to not get immediately frustrated.
I highly recommend MangoJelly Solutions's tutorials.
Here's a playlist for FreeCAD 1.0: https://www.youtube.com/watch?v=t_yh_S31R9g&list=PLWuyJLVUNt...
But he has a bunch of other videos.
The parametric workflow can be confounding to some people, but most pick up the newer FreeCAD interface fairly quickly:
https://www.youtube.com/@4axisprinting/videos
Best of luck =3
I use Fusion 360. Free for hobbyists. Yeah it's quirky and they constantly screw the free plan out of features (e.g. less saved editable designs, having to use the cloud to export STL) but it is also a highly capable tool that aligned best with the stuff I already knew.
Not entirely sure if it's available for Linux.
I probably shouldn't use autodesk but I'm not trying to make the world a better place. Just to unleash my creativity.
Not sure if they changed this, but you used to be able to local export an STL without cloud by going to Utilities -> Make -> 3d print
It's not. There is a flat pack version but it says it's not supported
Ah I see. I've been looking at FOSS options like FreeCAD and Blender but both didn't feel right (especially blender as it's more a tool for animators).
And I rather spend my limited free time creating stuff than to learn a new tool. Unless it is actually a more powerful one for the purpose that enables me to do things I can't now. But this doesn't seem to be the case.
It's the same reason I use BambuLab printers. My hobby is making stuff, not tinkering with printers. They're just tools, a means to an end.
Ps forgive me my defensive attitude but I often get people at the makerspace that take my choice of tools as a political statement. But I don't care. I just want to use what does the job for me.
For Blender, try adding:
https://www.cadsketcher.com/
and
https://blendercam.com/
I can't vouch for this, but maybe you could get SolidWorks working in Wine? (e.g. https://github.com/cryinkfly/SOLIDWORKS-for-Linux). Of note, SolidWorks is cheap if you're a student or veteran, for a non-commercial license. It is a dramatic improvement over FreeCAD. (I wish CAS were in a state like EDA and artistic model makers where the free/OSS software was on par with commercial, but we are not.)
I’ve had a lot of success with https://onshape.com, which just needs a browser.
I use FreeCAD, but it definitely leaves some UX refinement to be desired. There are a couple of web based options like OnShape that seem to work well, too.
OnShape is great (we have been using it exclusively for a project over the past four months, the collaboration tools are phenomenal), but FreeCAD has made some fantastic progress over the past year. Some of the underlying technology problems have solved, and the UX has improved a lot with 1.0. The customization and scripting opportunities are also wonderful with FreeCAD. That said, if you’re coming over from Solidworks/NX/Inventor, as much as there are buggy parts of those, FreeCAD still has extremely frustrating workflows and buggy parts that you have to work around. It feels like it’s moving closer to Blender-like quality, but it still has a long road ahead of it.
All of Solidworks, Onshape, and Freecad have a very similar operating philosophy (I believe they're all based on the same backend engine). I used onshape for a while because I found freecad unusable but recent improvements solved most of those issues and now I prefer freecad.