31
votes
The deadliest of all dead ends in the 3D printing industry
Link information
This data is scraped automatically and may be incorrect.
- Title
- The Deadliest of all Dead Ends in the 3D Printing Industry - 3DPrint.com | The Voice of 3D Printing / Additive Manufacturing
- Authors
- Paweł Ślusarczyk
- Published
- Jul 22 2024
- Word count
- 1714 words
From the article:
I was saddened to hear that Shapeways, one of the early successes of bespoke consumer and small designer 3-D printing, filed for bankruptcy earlier this month. Every time I've thought about getting a desktop printer, I wind up asking myself, "But what will I really use it for?". Wax printers precise enough to prototype jewelry prints for casting are hellishly expensive (as were Shapeways' prices), and it's not like I could direct-print enough interesting items at a low enough material and time cost to sustain a business. The article agrees with my impressions on a larger scale - 3-D printing is a protoyping and bespoke design technology, not ready for manufacturing.
"Never say never" and all that, but the nature and cost of the materials, and slowness required for precision, mean that our fully automated on-demand luxury communist future needs further development.
My buddy has a 3d printer, and after a few years of learning to use the thing well, it's great for "Oh I need a thing that'll cost me $2, but I don't wanna drive to the store."
Latest he shared was a precise measuring cup for his powdered laundry detergent.
But the primary use is 3D printing plastic penises and hiding them in his friend's houses.
If he's printing all of these plastic phalluses, what does he store them all in? A bag of some kind?
A 3D printed individual enclosed holster.
I kid, though I gather the primary use case for 3D printers is printing organizational systems for all the equipment.
The problem with 3D printing as a tool is that it's wholly contingent on the user having CAD experience and a willingness to go through an iterative design process.
Just in my bedroom alone, I have
All of these designs were made in more than one iteration (industrial design is hard!), they are all made for a market of one, and with judicious application of texture embossing, fillets, and chamfers, they look reasonably professional to where no one notices them. That's a massive victory for a weird home plastic fab device!
Lurking through the massive 3D printing subreddit, I get the feeling that the tool's real usage will come when kids that cut their teeth on tinkercad as part of school STEM programs grow up. CAD is a weird beast for the uninitiated and a tall hurdle to clear just to make some one-off bracket but if you already know how to use it, then 3D printers become something much more powerful than a Yoda figurine generator. Especially considering that CAD used to only unlock the ability to put in an order to make some fab shop make things for you to great expense. Now you can own the entire process.
I've been considering buying a 3D printer on-and-off for years for exactly this sort of thing.
The thing about CAD isn't that I'm averse to learning it. It's weird for sure, but nothing worse than anything of there other things that I tinkered my way to usability with (which includes early-2000s Blender). The patience is there. What rubs me the wrong way is how the prevailing programs are either Windows-only or run like garbage even on high-spec machines (hello Fusion 360). If it were like image editors or IDEs or just about anything else where there were options that work well on macOS I'd be much more inclined to spend time on it.
Yeah, the CAD suite landscape is dire. When the grand enshittification started post-2020, the cheap/free options for non-profit hobby-use vanished or were spirited away behind expensive paywalls with monthly-payment locks.
Right now if you just want to mess around, OnShape (cross-platform, browser based) is the main option. It's pretty new to me but there are things about it that I really like. It's treatment of parameterization is much more integrated than Fusion. Like the little sun clips I mentioned, since the fit had to be so particular, I made every active dimension a variable so I could print-test-update-print till I got it nice and tight.
Fusion can do the same thing but it just feels like the parameter feature was tacked on. It's just this laggy pop up spreadsheet menu that's a pain to edit, drops mouse clicks, and lags the entire interface as it regenerates the model when you change something.
I guess I should also mention FreeCAD. It's about the only cross-platform FOSS option out there but it's pure rage fuel. A lone vigilante developer has created a fork to overhaul the entire UI and fix some big issues but getting his work pulled into the main repo has been met with some aggression from upstream. Who knows if it will ever get merged.
I'm glad to hear this isn't just me. I've put more time into FreeCAD than anything else and I still absolutely hate every bloody second of using it. I'm still not sure if I suck at CAD or FreeCAD is just making me feel that way.
I used SolidWorks professionally for three years, and also didn’t care much for FreeCAD, so I don’t think it was just you.
OpenSCAD is quirky but less infuriating than FreeCAD. I've found quite a number of great examples on thingiverse, like customizable text lego bricks.
There's no drag and drop functionality, entirely programatic, but it's not a bad thing, just a learning curve thing.
I've got a half-finished raspberry pi enclosure design which I would class as 'not bad for a 101 self-taught student'
I'm new to 3d printing and went for OpenSCAD as my first CAD tool, the "for programmers" description sold it to me as thats what I've done for most of the last 20 years.
Definitely an interesting learning curve, but so far it seems to work exactly as described and I'm just starting to explore the vast amount of plugin libraries.
Getting my head around composing 3d shapes is taking time, but I'm liking the "programming" aspect as I can encode physical dimensions into the design as variables that I can update in moments when test prints prove I need to tweak some aspect.
OpenSCAD does have drag-and-drop functionality. It's just that it only works for moving blocks of text. :)
In all seriousness, I really like OpenSCAD. I actually find working with it a lot more intuitive than graphical CAD programs, though it's nowhere near as powerful without a degree of knowledge in 3D geometry. The only thing that bugs me is that it doesn't like having negative space boundries that intersect positive ones; it gives you an infinitely think border. In theory they're too small to print and won't make a big deal, but it makes graphical bugs on the screen that drive me up the wall, so I always offset the negative space boundary to overflow by 0.001mm.
BTW, RealThunder has a patreon.
I'm still waiting for FreeCAD to get out of its Blender 2.5x phase. There's so much good work already done in the application, but it needs a drastic reshuffling around, probably both on the inside and in the interface. If we could get to the Blender 2.7 phase of "non-standard, kind of awkward, but stable, usable and reasonably featured", perhaps a small portion of the industry would start considering it potentially useful and start donating money for development, which could slowly snowball just like Blender did.
But so far I'm not very hopeful.
As a mechanical engineer with my own Solidworks license ($~5k, business expense), I cackled here. Solidworks uses an embedded instance of Excel for this, and it is terrible.
OnShape is what I use too. It’s free if you don’t mind making all your designs public.
I'd like to share my absolute hate towards Autodesk that grew inside me after using Fusion 360 for a while.
You see, Fusion 360 was actually a pretty good application relative to the extremely expensive or not so great competition, in my experience it ran pretty well too and it had a generous pricing model for both hobbyists (free with a lot of features) and startups under a certain revenue (cheap). So naturally we used it with a friend on a couple projects that we aimed to make commercial. I hate cloud-based apps, but there weren't really many viable alternatives.
We were naive and didn't realize that the whole time Autodesk aimed to do a bait and switch. I knew many people hate them, but I thought it was like Adobe hate, where people dislike subscription models and cancellation fees, but the overall service is still pretty good.
They gradually removed features from the free version. Most of them we didn't need, but it started to be concerning after a while. Then they removed cheap startup pricing (and made the full price slightly cheaper, but not nearly as cheap - and it's a subscription, so it adds up). Finally, and this is where I swore to avoid Autodesk at all costs, they removed collaborative editing.
This was one of the main features of Fusion 360 and one of the main reasons why we used it. We could take turns working on a model and it would seamlessly synchronize. The number of shared projects in the free version was always limited, but that's not a problem, we only needed to work on a couple.
They completely removed this feature and revoked access to all models that were not created by you. I needed to edit some stuff a year after we finished the models and realized I cannot access anything that I worked hours and hours on, because the models were started by my friend, even though I had full admin access before. And since Fusion 360 is a shitty cloud app, where you have no real control over anything, you do not have any offline backups that you could reasonably get to. The files can be exported as .step for example, but the native versions are cloud only.
I don't think I've ever been more disgusted with a piece of software than when I found this. Do not use cloud apps you have no control over for anything important, especially not by companies like Autodesk.
Ugh, yeah. I'd switch from FreeCAD in an instant if one of the commercial options satisfied my very short list of requirements: one-time payment of an amount reasonable for an individual hobbyist, doesn't require using their cloud service, runs on Linux. But FreeCAD, wonky as it is, is the best thing I've found that meets those, so that's where I invested my learning effort.
(It doesn't have to be open source. I did just buy LightBurn for some recent lasering experiments, because it met all three requirements and is nicer to use than the open source options.)
(edit, lest someone take the above as a recommendation: LightBurn emailed me today to say they were dropping Linux support. Sigh.)
While it's not really "CAD" in the traditional sense, 90% of people that have/want a 3D printer can design just about everything they could need with TinkerCAD.
Is it a more limited tool? Yes.
Is it still more than capable without the huge learning curve of more traditional CAD? Yes.
Will I just be limited to lego bricks? No, but it does have a lego brick creator build in.
Are there tutorials and gurus that have turned out incredible things despite the seemingly limited capabilities? Absolutely.
At work I have access to professional CAD software. For small home drawings I tend to use tinkercad as it is so quick and easy to use. It can import STL files, cut holes, draw any shape so even for editing existy models off thingyverse it is the best home user option.
Setting aside the learning curve for CAD as applied to 3-D printing, I've seen many marvelous (and marvelously useful/artistic) items come out of desktop scale devices. The article really drills in on the current lack of mass scale and material problems. There are thousands of manufactured products that could take advantage of the infinite adjustability and versatility of additive forms, but the printing technology isn't there yet.
If I wanted to make the one tiny Teflon gear that's busted in my no-longer available food processor, there's no desktop (or production-scale) printer which can handle that feedstock. Not to mention the nasty fumes even if PTFE was an easily extrudable material. If I wanted to make a wildly ornamented 12" flower pot, the material cost, time, and limited scale for home printing would be very hard to work with.
I've tinkered with precious metal clay for jewelry design. Extrusion and CAD would make much more precise and versatile forms from the clay than the usual pottery-like hand-formed results. However, the cheapest bronze clay is still expensive for a wasteful iterative design process (and yes, there are now printers that can handle a DIY version, which is what led me down the rabbit hole again).
For objects that can be made from common feedstocks, the sheer length of time and unreliability of output makes other processes (injection molding, casting, milling, robotic assembly, etc.) much more economical at scale. That's why the article focuses on the overcrowding of the desktop scale printer ecosystem, and the impending collapse of the nascent commercial printing marketplace.
This is tangential, but are you sure it's Teflon? That would be strange for a gear, maybe it's Acetal/Delrin/POM? If so, there is a POM filament from GizmoDorks. I have a spool but haven't tried it yet.
BUT - modeling gears is a challenge anyway, so maybe a moot point
POM is more common for sure, but I have seen PTFE gears before. Supposedly they're nice to have because they are self-lubricating, but the one I saw was broken so I couldn't tell you about how reliable they really are.
Ooh! May I ask if you make jewellery? Totally agreed that FDM printing would result in tonnes of work at the grinder were it deployed for direct use in lost wax casting, but maybe a desktop SLA machine for a two part plaster mold might work?
I’d note that my printer is only used for ugly functional parts, for which it’s already paid itself off, so I’m likely working in a different space as you. Still cool to hear about other peoples’ work, though, if you’re up to sharing :)
I'm mainly a hobby jeweler and metalsmith due to the costs and learning curve for producing anything that meets my (admittedly very perfectionistic) standards. I've done some hand-sculpting in wax, but don't have the spoons, space, or resources to build my own casting setup, which is what then led me to PMC work. Again, I wasn't satisfied with what I could make, and that led to toying with the idea of 3-D printing.
It's only now that good wax and ceramic printers are available, but they're still not right for reasons of cost, scale, or precision. As you mentioned, the amount of hand finishing needed after printing is a big issue if you're working in metal rather than plastic. The couple of reasonably finished CAD prototypes I sent out for Shapeways printing came back almost unusably rough. That was years ago and quality has probably improved. Even now, truly jewelry-quality wax printers start at 4 - 5 figures, the software is likewise expensive and difficult to learn, and you've still got to deal with investment and casting costs.
So I contented myself with lapidary work (stone shaping and carving), which has a much lower cost of entry. We've just finished rebuilding a workshop, and I may revisit the whole 3-D printing process whenever work lets me have the time.
Most professional jewellers use sla printers with castable resin rather than printing wax.
Source: have a friend who works in IT for major high street jewellery firm. Also am professional jeweller.
Castable resin is expensive and the burnout process is a little fiddlier than wax but neither are unbearable.
I don't have the link to hand, will try and remember to edit later today but look up "print wave casting" for a slightly different take on the 3D model to metal pathway.
Thank you! It's been about 5 years since I last looked deeply into the state of jewelry-grade printing technology, and it seems to have changed a great deal. Overwork, arthritis, and lack of workspace got in the way, and I'm just starting to research the state of the art and affordable solutions again. I see Formlabs is all over this, so it's time to dig deeper.
I have all kinds of ideas for multicomponent bracelets. They're ugly expensive for materials and it's very time-consuming to solder hinge parts by hand, while there are some neat snap-together hinge designs from plastic printing that might transfer to metal.
Here's that print-wave technique I mentioned, which is being done on an FDM printer but I reckon with a bit of adapting could be used with SLA and get jewellery-quality results fairly easily.
Formlabs printers are nice but they are super spendy. I have an Elegoo Mars 4 which was less than two hundred GBP with a few bottles of resin and that's good enough for me, although I don't print much jewellery stuff if I'm honest because I have a stack of already-made print forms which covers the majority of casting I do - although when (if!) I have time I would like to do more.
As a semi-regular user of 3d printing who's not in the community or watching news much I was mostly surprised that people have been still trying to push plastic 3D printing as a legit manufacturing process. I mean, I actually used to use it like that, 3D printed amplifier cases, but that was because we sold one piece a month.
It just seems so obvious that FDM was always going to be prototype only aside from very low volumes, and SLA seems like the same with maybe an order of magnitude higher volumes, if you're making small stuff, because it's faster and more accurate, plus it can do things that are difficult with FDM, like overhangs, complicated internal shapes etc. Guy I know produces some accessories for vape pens (stands I think?) that he sells with mild success, for example.
The situation is different with metal powder printers I think, because there you can do stuff like drastically reduce the number of parts when producing complicated engine manifolds for example - those are relatively low-volume anyway and the assembly cost is probably high, so it actually helps. And the parts can be relatively large too iirc. This is already used in some jet plane motors. But the printers are only financially accessible for non-small companies.
But as for FDM, it seemed obvious from the start that most home users will never need them and companies will only use them for rapid prototyping, where their value is very high. I use shared printers in our local hackerspace, and in a community of DIY minded people who like to build things of various complexity, they're used all the time, for non-commercial and commercial projects alike. Spaces like that seem to be where they work the best.
Not sure of the exact process they use, but several of the newer rocket companies including SpaceX also make heavy use of 3D printing to manufacture their rocket engines, where it’s advantageous to be able to for example print plumbing right into walls and blocks, removing complicated and failure prone external parts. It also enables design iteration at a pace that’s uncharacteristic of traditional processes.
Even the cheapest rocket engines are not exactly pocket-change cheap though, and the added reliability and speed are well worth the costs involved.
Fully agreed that small volume production is the only use-case for 3d printing. We use it at work to prototype parts in plastic before we have them expensively made out of metal. At home it is great for making parts custom tailored to my need. When I compare a part printed at home it's quality is below the professionally made parts, simply because tolerances and materials are hard to get right at home. I just don't see how the average person has a need for these parts.
One case where I do see an application is in miniature printing, like for DnD and such games. Those minis are small low volume products when store sold, so buying a CAD file and printing at home becomes cost effective quite quickly.
Metal is quite different as you've said, and others have pointed out. I can give some examples that can give insight into the industry as a whole, but short of what people have already mentioned, theres only a handful of interesting case studies left to share.
All in all, as an engineer in the AM space, other peoples observations are right; 3D printing is great for one offs or low volume production of bespoke parts. Process limitations are what are stopping it becoming more mainstream - its very difficult to automate from at an industrial level in comparison with traditional manufacturing. There are so many variables, that we're unlikely to see it leave this space of niche, small volume production for at least another 2-3 decades.