Ask HN: Why did consumer 3D printing take so long to be invented?
"Today I saw an old paper printer, it has all the pieces for 3D printing except the plastic. Computer sends it data, it moves in 1D to print the image on paper using ink. A 3D printer is just 3 of these moving pieces from a paper printer with a melting plastic and thin pipe and software to connect it all.
All the pieces existed to make a working 3D printer existed even in 1970! and relatively cheaply. So why has it taken so long for [at home] 3D printing to actually become a thing?
Is it because of the internet somehow? Did just no one care in the 1970-2010s? Like there aren't even prototypes from 1970 from garage hobbyists for 3D printing.
What was wrong?!"
- omgsoftcats
214 comments
[ 12.3 ms ] story [ 5197 ms ] threadhttps://futurism.com/expiring-patents-set-to-improve-3d-worl...
Even something that seems as simple as "sous vide" cooking, which is basically a heater hooked to a thermocouple, took a lot of little innovations to make practical to hand to the masses.
And then, there's the general improvement in motors speed, precision, and cost, along with any number of advancements here and there and everywhere to make it practical.
Could someone thrust back to the 1970s and given a fairly substantial budget make some kind of 3D printer? Probably. But it's slow, extremely expensive, and can print various sizes of plastic bricks and spheres and other very algorithmically simple objects, and in rather low quality without many, many years further development. I can think of many ways of bodging on various improvements, but they're all going to have their own compromises, not be garden paths to what we now think of as modern 3D printing. (For example, someone could bodge together a machine that offsets a rod on the printer head, and then in the offset space, has an object to be "copied", by basically banging into the object with a very crude sensor, so there's no generation of geometry at all. But this is going to be clumsy, inaccurate, and full of very complex and disheartening limitations.) You're not going to be printing Dwayne "The Rock" Johnson's face embedded into a toilet [1] or anything, at 1MB of geometry. It will be commercially useless and inaccessible to hobbyists.
[1]: https://www.thingiverse.com/thing:6436138/files
For me this is how meaningful technological progress happens. It is not like someone one day wakes up and suddenly has rocket that reached space caught back on landing by mechanical arms.
"Just adding a fan" is actually an extremely meaningful tech development.
Temperature control is probably hardest part. But in general air fryer could have been done quite a lot of earlier. Maybe materials for the basket is other aspect.
3D printing is not one of them.
People have been doing CAD/CAM since the 01950s. Boeing started using CNC in 01958 on IBM 704s, and MIT's Servomechanisms Lab (working with the Aircraft Industries Association: https://web.archive.org/web/20090226211027/http://ied.unipr....) sent out CNC ashtrays to newspaper reporters in 01959: https://en.wikipedia.org/wiki/History_of_numerical_control#C.... Pierre Bézier started writing UNISURF in 01968 at Renault, who was using it to design car bodies by 01975. The Utah Teapot was created in 01975, and it consists of nine Bézier patches; you could print the whole dataset on a business card: https://web.archive.org/web/20141120132346/http://www.sjbake...
The IBM 704 was a vacuum-tube machine that could carry out 12000 floating-point additions per second and had a failure about once every 8 hours https://en.wikipedia.org/wiki/IBM_704. The Intel 8008 (not 8088, not 8080, 8008) that came out in 01972 could carry out over 79000 8-bit integer additions per second, which is about the same speed. But much faster computers were already available, such as the PDP-8, in wide use for real-time control, and they very rapidly became much cheaper. Any computation MIT's Servomechanisms Lab could do in the 50s was doable by hobbyists by the 80s.
The reason 3-D printers mostly use stepper motors is that they don't require closed-loop feedback control. 2-D printers from the 01970s used stepper motors for the same reason. They were accessible to hobbyists; in the 80s I had a Heathkit printer someone had built from a kit in the 70s.
If you wanted to print Frank Sinatra's face on a toilet, I think you'd probably want at least a 64×64 heightfield to get a recognizable Sinatra; 256×256 would be better than the line-printer pictures we were doing. 8 bits per heightfield point would be 65 kilobytes, which would fit on the floppy disks we were using at the time. This would have been totally feasible, though digitizing Frank Sinatra would have been a nontrivial project, quite aside from printing him.
So I don't think computation was the limiting factor.
Your "basically banging into the object with a very crude sensor, so there's no generation of geometry at all" is called a "pantograph" and it has been a common way to copy three-dimensional objects and engrave letters with a milling machine for 180 years: https://en.wikipedia.org/wiki/Pantograph#Sculpture_and_minti...
I don't think that's a show stopper though. If we had had 3D printers in the 1970's and 1980's, sold at Sears, between the computers and the tools, you'd bring home a paper catalog from the 3d model company, thumb through it for models you wanted to build, send off a cheque (and a SASE) and they'd mail you back a cartridge that you plug into your printer so you can print that model. And then get that catalog in the mail forever after.
As reference on how much home computing power was available back then, the original Apple I came out in 1976 with a "character generator" and was not capable of advanced graphics - the cost of the amount of ram it would take to draw an entire screen from ram would have been astronomical, so characters were stored in ram and the card was in charge of outputting the characters and rudimentary graphics.
Models where severe limited, at most few thousands triangles per scene, but it was enough for talents to make crazy things.
Not exact. First, funny note, Seymour Cray admit, he used Macintosh to design Cray.
Second, existed activity of using Atari ST for 3D design, as sort of companion machine for special model of Commodore, which could process analog video. As I seen, they directly write frame-by-frame on Betacam.
Sure, rendering on ST with 68k was slow even with TV resolution, but they made lot of commercials on TV.
In late 90s, TV commercials made on MMX or Pentium-Pro with 16..32mb of RAM, and use PC frame grabber cards and RAID HDD. And that's history, how SGI become bankrupt - just when Maya was ported to x86 this become end.
Other corps will just use your invention and just skip the markets where you have a patent. And yet their product will find its way there organically.
Many individual inventors that got patents openly talk about how useless and expensive they are.
The most expensive part of home is general the land it's built on.
There are all sorts of restrictions and rules that create this artificial scarcity. Even something as simple as buying a plot of land and parking a trailer on it is not legal in most places except in designated trailer parks. You can get a trailer for next to nothing. And lots of people live in them. But try finding a place where it is legal to put one down and live in one. If it were legal, lots of people would do that. Land plots are scarce and once you have one, you can't just do what you want with it in most places.
I'm just using trailers as an example here. Think prefab buildings and raw material cost. This isn't rocket science. We've been building shelters since the stone age.
There are of course good arguments for this to be made in big cities because of a lack of space. But it's equally frowned upon in areas where there's plenty of space.
They will probably never be the sort of thing that exists in every home, but they could very well be the sort of thing that exists in every home workshop.
So really, for an average hobbyist the idea of a 3D printer controllable from a home PC wouldn't really be possible until like the mid 90s. So you really need to start your look at why it wasn't a thing at some point in history I'd start the digging there, not the 1970s.
However, I doubt most home hobbyists had computers in their house in 00000000001959 capable of manipulating 3D graphics in a meaningful way that would really be approachable to a general hobbyist. Wikipedia suggests 1960s CNC's were often controlled by a PDP-8 or similar minicomputer. When the PDP-8 first came out in 00000000001965 they cost $000,0018,500, almost $000,180,000 today. I dunno about what kind of home you grew up in but a machine like that wasn't exactly a home PC available to spend all night rendering the output for a CNC machine which probably also several thousand dollars. And that's just the computer driving the CNC, not even thinking about the machines and knowledge it took to actually code the designs and curves and 3D patterns.
I'm well aware of computerized CNC machines from at least the 80s. I had family who owned a machine shop. They were not really hobbyist accessible things.
My point about APT is that a "beefy machine" in 01959 was the embedded controller in your keyboard by 01983; it wasn't a beefy machine any more. The PDP-8 was indeed pretty common for CNC control in the late 01960s, and it could run about 333000 12-bit instructions per second, which is about half as fast as the original IBM PC. So, yeah, a machine like that was exactly a home PC by the mid-80s. For real-time control of the 3-D printer, you can get by with less.
There were three big problems for manipulating 3-D models for home hobbyists in the 70s and 80s.
One was computational speed: for games, you need to render the 3-D graphics fast enough to provide an illusion of immersion, and with a few hundred thousand instructions per second (and no multiplier) you were limited to a handful of polygons. Like, typically about 20. See the Apple ][ FS1 flight simulator https://youtu.be/lC4YLMLar5I?&t=93, the 01983 Star wars arcade game (at 2:02), Battlezone (2:14), and Ian Bell and David Braban's Elite from 01984, which was successfully ported from the 2MHz 6502 in the BBC Micro (2:53, but also most of the rest of the hour of the video) to the Z80-based ZX Spectrum (3.5MHz but noticeably slower than the 6502; see https://www.youtube.com/watch?v=Ov4OAteeGWs).
For producing G-code, though, you don't need to be able to handle all the 3-D geometry in 50 milliseconds. You just need to be able to handle it overnight. That's a million times longer, so you can do a million times as much computation on the same hardware. You can't handle a million times as many polygons, because you don't have enough storage space, but you can represent geometry in more expressive ways, like Bézier patches, nine of which made up the Utah Teapot I mentioned in https://news.ycombinator.com/item?id=42080437, or parametric or implicit equations, solids of revolution, CSG, etc.
You do need some kind of user interface for seeing what you're designing that doesn't require waiting overnight to see the results, which I think is what you mean by "a meaningful way that would really be approachable to a general hobbyist". But it's possible I have a different concept of general hobbyists than you do; as I remember it, home computer hobbyists in the 70s were constantly writing things like
and writing programs in assembly language. And machining metal in a machine shop has been a popular hobby in the US for at least a century, using user interfaces that are more demanding than that. So I think hobbyists would have been willing to tolerate a lot of demands on their mental visualization abilities and relatively poor user interfaces if that was the price of 3-D printing.But the user-interface issue gets us to the second problem hobbyists had with 3D in the 70s and 80s: display hardware. The limited displays of the time were hopelessly inadequate for displaying realistic 3-D. The Star Wars arcade cabinet mentioned above used a vector CRT in order to be able to do high-resolution wireframe, because framebuffers were far too small. With color palettes of 2–16 colors, even Lambertian flat shading was nearly out of reach until the late 80s.
Again, though, this is much more of a problem for games than for 3-D printing.
My first experience with CAD was on an IBM PC-XT with a 4.7MHz 8088, roughly five times faster than the BBC Micro Elite is running on above (https://netlib.org/performance/ht...
Yes, we're talking about very different levels of hobbyists here. I'm talking about the people who are frequently doing 3D printing today. They're often not thinking about the actual algebra and Bézier curves and what not. They're playing around in CAD software extruding surfaces and connecting vertices and what not, if they're even going that deep. They're downloading pre-made models off Thingverse and printing them with their printer, tinkering with the physical aspects of their printers. Lets try printing this model in this material, lets mess around with this fill, what if we print it at this angle, etc. They're not digging deep into the code of how the CAD application actually works to draw the curves. They absolutely wouldn't be hand typing out the functions to draw The Rock's face on to a toilet, letting it render all night long, going back to tweak their math, letting it render all night long, rinse and repeat.
> on an IBM PC-XT with a 4.7MHz 8088
> the machines had two video cards
A pretty high-end machine for many home users (~$5k new in 000001983, almost $16k today). I realize there were even higher end machines out there though.
> a version whose 3D functionality was inadequate for any real use
Exactly my point. And when would that machine have come out? The PC-XT released in 000001983. So even in 000001983 decently high-end home computers with $001,000+ (of 1982 money, >$3k today) software suites weren't very suitable for even basic CAD work without many pitfalls and limitations.
So general home users who barely passed highschool algebra and don't have degrees in mathematics trying to use 3D printers wouldn't really have the compute capacity to think about what they're trying to make and modify without getting really deep into it, at least until about the 90s. The 3D printing scene as it exists today pretty much couldn't exist until the 90s. Sure, very smart people with piles of cash could have done it in 000001980, potentially even the mid 000001970s, but not like a go to the computer store, buy all you need for <$000,500 (about $000,060 in 000001970s money), plug in to your cheap $000,200 (about $000,020 000001970s) computer, dial into the BBS, grab some files for The Rock's face on a toilet from your BBS, and start printing in an afternoon.
Which is why I'll always prepend even more digits when replying, because I'm thinking even further ahead than the short-sighted five digit people and even bigger numbers all the time.
https://longnow.org/
This 00is 0Hacker 00News, k?
(But wouldn't outnerding a nerd require not just vel0city's smug sense of superiority, but also demonstrating more than a de minimis level of knowledge or technical skill?)
HN comments are not the right place for personal attacks. :'(
I'm being attacked for my neanderthal inability to understand five digit numbers :( my primitive understanding of time can only really comprehend two digit year values. Today is the year 1924, right? I failed to write my calendar application to accommodate such high year values as your superior intellect is used to reasoning about. Here I thought ISO 8601 was over engineered, turns out it didn't go far enough. We're all just too simple-minded compared to other big-brain bretheren. May we all learn from our five digit betters.
I'm sorry I'm just not smart enough to be at your level :'(
000make 0any 000sense?
* https://www.biodiversitylibrary.org/item/86617#page/385/mode...
* If GenAI is good at poetic forms, could it be asked to generate sentences with a particular mono- vs multi-syllable word pattern? Or would this search better be done in Prolog/with a TTS db?
If I'd attempted to use this to generate an irregular pattern, it'd probably have wound up producing messages decoding to something like "Sammwich Hewix"...
Before that, the precision available without gearing and feedback wasn't sufficient. There were systems but they were order of magnitude more complicated and several orders more expensive.
hard drives use voice coils, a completely different technology. The circuitry that does that evolved and certainly influenced the creation of microstepper controllers: the neat trick they do is treat the stepper motor as a voice coil in between full steps.
Hell, CNC machines existed back then too.
The patents, compute, research access, and dozens of other relatively small barriers created a thicket of challenges and no obvious way to reconcile them, even if you had all the right ideas and inspiration. I think the internet would have been needed in order for all those ideas to come together in the right way.
I mean, towards the end of the decade was something like the ImageWriter, which let you do bitmapped graphics, as a row of 9 dots at a time. At https://www.folklore.org/Thunderscan.html?sort=date you can read about the difficulties of turning it into a scanner. (Like, 'We could almost double the speed if we scanned in both directions, but it was hard to get the adjacent scan lines that were scanned in opposite directions to line up properly.')
The LaserWriter wasn't until 1985 or so. My first hard drive, 30 MB, was a present from my parents around 1987.
By the 1996, laser-based 3D printing based on cutting out layers of paper was a thing, available for general use in one of the computing labs in the university building where I worked.
The result smelled like burnt wood.
When I visited a few years later they had switched to some other technology, and one which could be colored, but I forgot what.
You can look at early calibration settings descriptions and they're still talking about e.g. "The number of X stepper-motor steps needed to move 1 mm for the PIC."
The motors tend to fall to the nearest full step when loaded hard. Most people I have discussed this with believe there is little resolution gain (if any) past 4x microstepping but it certainly is a lot quieter to use x256.
1. Cheap stepper motors and electronics from China
2. Expiration of Stratasys patents in 2009
3. Widespread availability of CAD software and desktop computers powerful enough to run it
4. Reprap project made it easy for companies (and individuals!) to develop their own printers
CAD skills are essential, and it turns out not as hard as you might have thought!
If you think of it as functional/decorative categorisation first of all, obviously some people will overlap but broadly speaking I think people are interested for one or the other, then within the 'decorative' camp you can go a hell of a lot further without and I think it's more obviously reasonable to not care about designing your own models. You never wanted to design your own toys, but there's appeal in printing things not available on Amazon, unofficial merch for a film you like, or whatever.
Not to say there isn't functional stuff (which I exclusively print) on these sites, but often it won't be quite what I want, so yeah I end up in Fusion. (And typically starting from scratch eventually, because for some reason people don't share source, and working with imported STLs is hellish.)
My friend is filling up hard drives with 3D models DMs share.
I'm not saying it /can't/ be smooth, but I contrast it to my dishwasher, which I expect to wash dishes, correctly and successfully, about 99/100 dishes (assuming I rinse them well), and then when it starts to decline in 3-5 years, that it's time for a new one. 3D printers are not to that level of reliability yet, nor are their support systems around slicing/supporting/etc.
Fused Deposition Modeling or FDM (1989, expired in 2009), Liquid-Based Stereolithography or SLA (1986, expired in 2006), Selective Laser Sintering or SLS (1992, expired in 2012), metal processes like Selective Laser Melting (SLM) and Direct Metal Laser Sintering (DMLS) (1996, expired 2016).
A complete set of woodworking or metalworking tools was a lot cheaper than a home computer. And there were entire magazines dedicated to proliferating free or easily obtained schematics/designs. Labor was also cheaper, and people had more time for hobbies.
I would also argue the point that it would have been relatively cheap. We are used to the ubiquity of cheap DC motors and precision parts being a click away. But if you were to rummage through a vintage Radio Shack to cobble together a home printer, I think you would struggle to construct anything precise enough with consumer available parts.
> a melting plastic
Don't sleep on the chemistry of filament. It has to be extremely precise and consistent. We benefit from the massive economies of scale today, but this was small batch stuff 20-30 years ago. And if we are talking about the 1970s the plastics were really primitive by today's standards.
Easy. The printing process itself is not that hard.
It's the model _design_ that is tricky. We needed home computers to become powerful enough to run 3D CAD software, and enough people to get proficient with it.
RepRap started in 2005. Realistically, we could have had it maybe a few years earlier. But not _much_ earlier.
Early machines were industrial machines with huge price tags, proper linear motion systems, complicated extrusion systems and so on. There is a bit of a mental leap to go from seeing $100k+ machine and dreaming to design something that can be built for $200-500.
The problems were: no "reference designs", no tried and true go-to mechanical parts (like cheap chinese linear motion rails), extruders (they were DIYd!) or heated beds (early models were just PCBs) and so on - imo it just took someone to get this rolling and that may have well taken 30 years.
I think Reprap was first publicly shown around 2005. From then on it was taken on by more and more makers and refined. It culminated in the early 2010s hype with Makerbots and its contemporaries but they still cost > $1500 and were far from set-and-forget appliances, like 50% reliable and slow - we had one at work and I got fascinated but it printed at 5-20mm/s so parts would just take forever and often fail due to bed adhesion, clogs, ...
The last 10-15 years then have seen the popularization of 3d printers through the Prusa i3 and its clones (Ender and other cartesians < $300) and steady refinement of reliability through better materials. Then the last ~5years or so significantly bumped up the speeds through better linear motion components, motion systems and input shaping + firmware and ecosystems like Klipper.
Bambu imo got in at just the right time and refined everything that had culminated up to this point into a solid appliance. Imo their genius was more in the industrial design, making it reliable and affordable manufacturing than anything else.
Now you can get the steppers from Amazon for $8, a control board with stepper controllers for $20 with a built in 32 bit MCU. At scale if you're building a lot of them those parts are going to be even cheaper maybe even another order of magnitude. For a while it was difficult to actually even comprehend how much cheaper stuff was getting and what that lets you do. And then you see a resin printer for $140 and realize it's a cast-off I found screen one stepper and some extruded parts.
Once anyone could build their own design, a community of hackers and engineers formed that continuously improved the designs with diverse ideas and experiments. That community is what made 3D printing what it is today. And it was illegal for them to do all of that (in many countries) until the patent expiration in 2008. That’s a big reason why it took so long. I think it’s interesting to consider whether this would have happened sooner if they had never been patented, though perhaps the expired patent created a legal safe haven where no one could take away the basic principles by patenting them. Anyway, patents play a big role in this story!
Edit: Some cool history here: https://3dprintingindustry.com/news/interview-dr-adrian-bowy...
https://bsky.app/profile/tlalexander.bsky.social/post/3laiw4...
I think I have heard that in the country he was living in, violating patents for personal use was okay. My understanding is that things are much more strict in the USA, where I believe you can’t help others violate patents so you can’t publish work the way Adrian did. I know that in the early days of 3D printing everyone wanted a belt printer, but users and hackers on the forum regularly expressed concern over MakerBot’s patent and the associated legal risk of violating it. That’s why today’s belt printers have the head at a 45 degree angle. It’s a patent workaround because MakerBot’s patent specified a belt that was parallel to the motion axes. At least that’s what I think Brook Drumm of Printrbot told me.
So even if they didn’t affect Adrian, they certainly had a chilling effect. And I don’t think even under Adrian’s legal regime he would have been allowed to sell the work, so more expensive engineering development was prohibited. We didn’t get cheap 3d printers until companies could mass produce existing low cost open source designs, and that mass production was obviously prohibited by the patents. We didn’t get low cost machines until competition was allowed in to the space. And I’ve heard directly from 3D printer hacker developers that patents affected their decisions not to prototype certain new components. Sorry I can’t point you to a source directly but this would probably have been on the mailing list for the Bay Area Reprap club in 2010-2014, or the Ultimaker mailing list in the same timeframe.
Looks like context missing.
In jurisprudence exists definition of negligible case, meaning, harm is too small to run full-featured juridical machine. Same thing considered for example, when tax regulations ignore tips, because it need much more resources to administer than could gather as taxes.
So, laws usually ignore, when you do something prohibited, but nobody harmed, like if you violate patent but don't tell anybody about this. Unfortunately, this also means, you cannot involve other people or make business on this, so this activity will not scale to level, when we here could talk about affordable 3D printers. BTW, this is opportunity for AI, to use artificial agents instead of people workers, but this is very different context for now, and it is not researched good enough to talk about.
The Constitutional purpose of the US patent system is to promote the progress of science and the useful arts, so discussing how to practice or improve patented inventions is definitely legal. (That's why patents are published in the first place.) That's what Adrian was doing. There's no "personal use" exemption in the US law, but there is an exemption (in caselaw) for research, although a court decision around that time narrowed it substantially. It would be impractical to practice a patented invention for such protected research purposes if the law prohibited you from even discussing how to practice it. But in fact you can even go so far as to patent an improvement on somebody else's patented invention.
There is a prohibition on inducing others to violate a patent, but generally it's applied in cases where someone is selling a nominally-non-infringing product whose only real use is to practice the patented invention.
I agree that, at the much later time that MakerBot not only existed but had turned evil, its patents had a chilling effect. I can absolutely confirm that! But at that point the cat was already out of the bag, and the original FDM patent had already expired, so that particular patent no longer had such an effect.
The question I intended to discuss (because in my interpretation it's what superconduct123 was asking) is why we didn't get something like the Prusa Mendel V2 in 01980 or 01990 or 02000 instead of 02010. Obviously the concerns people had about patents after 02010 aren't the reason; if it was a chilling effect from patents, it would have had to be a chilling effect much earlier than that.
And I don't think it was. We could imagine a history where there was a low-cost open-source design in 01990 or 01995 or 02000 or 02005 which companies couldn't mass-produce because of the patent, but that isn't the history we actually got. Why not?
I think the answer is basically that a low-cost 3-D printer sounded like such a crazy idea that none of the small number of people who even understood the concept of a 3-D printer decided to dedicate the effort necessary to make it happen. Also, everything was much more difficult then. I think Adrian accelerated the timeline by 2–5 years.
But only 2–5 years. Fab@home or the (non-FDM!) Open3DP group at the University of Washington would probably have been successful without him. MIT started their Center for Bits & Atoms in 02001, and they started up their first "Fab Lab" that year. EMSL built their Candyfab 4000 in 02006, which was the first 3-D printer I saw actually operating (at TechShop, if I recall correctly). So the memes were spreading.
But basically I think you could have built a hobbyist 3-D printer in the 01980s if you'd realized it was a good idea. Just as you could have built a submarine in the 01600s. The necessary technology was available. It would have been enormously more difficult, especially before the advent of cheap PIC16s in the 90s. Finding like-minded people on GEnie or FidoNet would have been much more difficult than in the blogosphere. 3-D modeling in 512KiB of RAM at 1 MIPS would have been limited. But you could have done it.
The question I'm most interested in here is: what similar opportunities are we missing today?
Though that has more obvious advantages as a product that consumers want which translates to an more possibility for funding. There's a clear need for a device that gives directions, vs a 3d printer. While I'm very happy with my Bambu 3d printer, it didn't scratch an actual need in my life.
Looking for similar opportunities today, we have access to relatively cheap compute power. A 3d printer in the 80's would not have been all that cheap as a consumer device. Today though, we have 3d printers and relatively cheap (for now) parts to drive them - motors/gears/belts/UCs.
While LLMs have taken up all of the air in the room, they're not there only ML technique out there, and while GPUs aren't cheap, they're fair cheap considering the cost of a Cray, back when they made them.
Looking those things, I wonder the number of parameters you'd need for a system where the AI model is continually being updated as new information comes in. What size pet robot would be feasible, backed by a 4090? 5090? 10090 or whatever it's called in six years? Could you make a robot planaeria that had basic sensors and responded to limited stimuli? Robot drosophila? Could you make a robot snake with cameras for eyes? Of course, where I'm leading is a robot dog or cat with an AI model that learns from input given to it by its operator.
Not practical enough to get much funding given the state of the industry; the market for a $20,000 robot dog is too small. The technology is only somewhat there, or about to be, and I'm sure it'll be a "why'd that take so long" after the algorithms for it are invented/discovered which makes it seem obvious in hindsight.
I'm sure there are similar opportunities (the army rejected a robot pack mule for being too loud, but how fun would that be?) Or a food replicator - there are some 3d printers that do chocolate but nothing with an AMS. or how about a specific type of home robot food machine. How about a machine I put flour and water, a block of cheese, tomato sauce, and a pepperoni into, and out pops a ready to eat pizza or calzone? What else could you make an advanced device for egg ingredients go in, and it cooks a finished product?
Thing is, with the Internet and especially YouTube, I'm sure there's someone trying to build one (I saw the vending machine thats basically that, last time I was in Vegas, but it's way too big and expensive for a consumer product.)
Or a robot arm that plays chess or something else where the arm only needs to manipulate the things in front of it and doesn't need a moving chassis, so you can play chess against someone remotely, without being on a computer. Robot arms have existed since forever but it's only recently that we've gotten enough compute to make controlling them easy for an end user.
I keep going back to robots because parts are 3d printable or source able via Amazon, and since 3d printing's really come into being lately, that's would be the thing to take advantage of, in addition to relatively cheap compute.
Looking just a compute there's a load of stuff that involve helping users automate things on their computer, but that's so obvious a use case that many well funded players are after it. The problem with distributed computing is that the interconnect speed between nodes on the Internet is just too low to make it useful, so since Folding@home, the...
Technically, not speed low, but latency is huge.
For about speed, if you lucky, could build 100Gbit fiber to nearest exchange and this is comparable to speed of desktop RAM (around 40Gbytes, so just 320Gbit). But RAM latency is somewhere around 10ns and on fiber internet practically achievable around 10us to nearest DC (1000 times more).
With 3G/4G/5G latency around 1ms achievable.
Anyway, main question is will to do something. - If person don't want to do, he will not, even if have all need.
> Under EU and UK law anyone can use a patented technology to research improvements in it without paying royalties. So no, it wasn't any problem.
> This is not the case in the US.
I had a Wanhao Duplicator i3 in 2015 and found it required a lot of tinkering and calibration every time I wanted to use it. I ended up selling it as it was so time consuming to get everything correctly set up that it just killed any interest I had in it.
I built a Prusa MK4 this spring; it calibrated itself and printed a great looking piece right from the get-go. The difference is night and day.
I say this as someone who doesn’t want 3d-printing as a hobby. It’s just a tool I want to occasionally use in order to get something else done and the less time I have to spend tramming and calibrating, the better.
PETG is modified PET, so it depend on point of view - (glass semi-full or semi-empty). Any way, in Ukraine people recycle plastic bottles into filament, which is considered by printer software as PETG.
And in disposable forks that liked to advertise compostability. It's no ABS but it isn't and wasn't a rare plastic. I'm not actually entirely sure how common PETG was in products back then, but it's by no means a new invention.
- Home 3D printing is often more of a hobby than a traditional prototyping or engineering discipline. People view it as a skill to have, and a fun use of free time. Note how the cheapest and most finicky ones are popular; they can be made to work well through careful operation, troubleshooting, procedures, customization etc. They are not set-and-forget, and I think the userbase likes that.
- Home 3D printer parts (the motors, frames, electronics etc) are almost exclusively sourced from China. We live in an AliBaba world; that wasn't always the case.
Those AliExpress clone kits were really popular in the community in the beginning
And with waiting 5 hours to print it isn't unreasonable to wait an hour or two for the slicer either.
There's also the problem of 3D modeling and slicing. Again up until quite recently, 3D CAD was out of reach for most consumers. Either due to hardware capabilities or cost of the software. Slicing is its own entire branch of 3D processing and it took time to develop all the techniques we use today that make it fast and reliable. Slicing software could only exist after the printers were common.
As well, I expect the availability and materials science of the plastics we use needed some further development.
As I recall, 3D printers rose to prominence at about the same time and speed as we started getting genuinely powerful personal computers. You really needed a fast CPU, and printing became more accessible as the early I5/I7 generations became cheaply available.
While you absolutely could build an FDM printer with 80s technology, I don't think it could ever be practical or affordable. Even if someone invented all the computational techniques for slicing, the compute available back then was not even close. It would literally take an actual supercomputer to slice your model. It'd take many, many hours on any consumer computer. This would hold true until the early 2000s. At a random guess, I'd say the tipping point would have been around the Pentium 4.
So, same as most technologies we take for granted these days. Enabled almost exclusively by the speed and capacity of computer available to consumers.
Mind you they were nothing like the tabletop consumer ones we have today. They were about the same of a large American refrigerator.
Since it was not really any special or amazing for us to have several of them, I have to imagine that industrial 3D printing capabilities were well established by the point.
Edit: as I recall they were mostly used to make parts which could be given a nice surface finish and then from which silicone molds could be made.
The way inkjet and laser printers work is also quite different from the way a 3d printer works. The similarity is mostly in the gantry, so there was nontrivial innovation required here.
To some extent 3d printing is probably also a reaction to decreased access to domestic manufacturing. It doesn't make a lot of sense to produce most parts in plastic if you can get a cast or milled part quickly and cheaply.
Some of the greatest and most under appreciated technological achievements in the last 40 years have been in materials science and miniaturization.
I agree that the pieces did not exist in the 01970s, but the missing pieces weren't the computation.
https://drive.google.com/file/d/0Bxv0SsvibDMTZ2tQMmpyOWtsRFk...
I do think that the limitations of memory (and disk) will require some ingenuity in 3D printing more than the most trivial procedurally defined objects!
Logically it's a simple project, though the devil is always in the details!
In terms of paging I was thinking of paging in later layers of data from disk, perhaps from a large file, perhaps from a series of files. But CP/M certainly did have support for paging RAM, in 3.x, if your hardware supported it.
(My own emulator, and the hardware I have running genuine CP/M are all running 2.x so no paged RAM there. Shame!)
If you were using a 6502 or Z80 in the late 70s you wouldn't have to write the floating-point routines; you could just call the ones in the BASIC interpreter you probably had in ROM. They'd still be slow.
As for paging RAM, do you mean bank switching? The 8085 and Z80 didn't have MMUs, so you couldn't do "paging" in the sense people normally understand it today.
The FP routines in the Arduino library take about 5 µs for a multiply on an AVR (e.g. Uno).
That's 1000 times faster.
Factors in that include:
- 16x faster clock speed
- typically 1 cycle per instruction vs 3
- 32 8-bit registers can easily hold all the data for operands and result and temporary values, vs Zero-page on the 6502
- single-cycle register to register 8 bit arithmetic vs three instructions needed per 8 bits on 6502
So that's a factor of around 1633 = 144x combined.
All the above apply equally to add/sub and multiply, but also:
- AVR has 2-cycle 8x8 -> 16 multiply instructions, for signed, unsigned, or mixed operands, and for fixed point 1.7 format as well. That's hundreds of clock cycles on 6502 or z80.
Also, the Microsoft BASIC interpreters used a 40 bit floating point format (32 bits of mantissa) instead of 32 bit. That's another factor of 16/9 for multiply.
The 8080/z80 and 8086 have more registers than the 6502, but not enough to implement soft FP keeping values in registers like AVR can.
If I decided, IDK, to write "cnow" instead of "know", or "J" instead of "I", I wouldn't be able to do so consistently. Not in a world that massively uses the other word.
Or it would take me twice as much time to double-check everything that I have typed.
(I'm sincerely hoping I'm missing something here and am going to look really silly asking this question.)
Never underestimate what lengths someone will go through to appear to be unique.
It's possible that they were referring to the RepRap Host Software, which was RepRap's original slicer.
https://reprap.org/wiki/DriverSoftware
The system here was based around a minicomputer (or at least a successor patent of 1978 so described it) so we're talking tens of thousands of dollars for the compute involved in that scheme - but that first 1971 patent must have expired in the 90s by which time inexpensive compute was trivially available to match early 70s minicomputer capabilities.
Excerpts from the exceptionally excellent book "The Inventions of Daedalus - A Compendium of Plausible Schemes" which is sadly long out of print:
https://paperstack.com/img/photos/page%2090.jpg
https://paperstack.com/img/photos/page%2091.jpg
I don't know why they have a record for earliest filing of a 3d-printing patent, but this guy (Wyn Kelly Swainson) was an American graduate student of English Literature who filed the patent in Denmark after teaching himself some basic lithography and chemistry from a library after wondering why no technique existed to make copies of sculptures. He ended up doing research for DARPA and founding an engineering company.
Also, for any random reader who isn't familiar with the cultural history of the west, Daedalus was the mythical designer of the labyrinth of Crete made to contain the minotaur, and also the father of Icarus, who crafted the artificial wings he and his son used to escape Crete after King Minos tried to keep them trapped there after he helped Ariadne help Theseus to kill the minotaur. He was also somewhat responsible for creating the minotaur in the first place, as he built the fake cow costume King Minos' wife used to mate with a bull and birth the minotaur. I guess it became a popular pen name in the early to mid 20th century, because it was also the name (Stephen Dedalus) of James Joyce's self-insert character in A Portrait of the Artist as a Young Man and Ulysses.
[1] https://en.wikipedia.org/wiki/Centronics
As for the question of software for 3D models, see my overview of the history of that field in https://news.ycombinator.com/item?id=42080437.
But the point I was trying to make was that dot-matrix printers predated the availability of microcontrollers, or even microprocessors, and you needed cheap microprocessors to make hobbyist 3-D printers.
Yes, one could make 3D printer without PWM, but it will be extremely slow, or even cannot do some things.
With modern microcontrollers this problem resolved by using interface just for high level commands (all low level control performed by microcontroller), but on early machines used computer as controller and they have to deal with this.
I have some experience with modern FDM and SLA, and I seen many cases, when FDM was severe limited with microcontroller PWM range.
May be. Problem is by definition of G-code, which is much more than 8bit (as I remember, there somewhere between 12bit and 18bit if consider just integer numbers, but for example diagonal lines and arcs are calculated with floats). Yes, I know, digital 8bit CPU could calculate 32bit floats or even 64bit floats, but it is slow and need additional RAM. All these calculations are much easier to do on 32bit computer.
> some external chips for PWM
Additional chips are bad for economy by definition - because size of PCB grow and because additional pins mean additional costs also. That is why first designs was dumb, without any controller at all and driven by computer - to avoid additional chips.
In modern designs microcontrollers used for convenience - so now printer could run without working computer (for example from file on flash).
You wouldn't necessarily have to interpret G-code on the 8-bit microcontroller itself, although it's about the same difficulty as interpreting BASIC on it. Keep in mind that keeping the motors of a 3-D printer busy only requires a few speed changes per second, maybe 10 at most. By contrast, the 8051 in an Epson MX-80 printed about 80 characters per second and had to actuate the 9 hammers in the print head with a new set of voltages 9 times per character, for a total of about 700 deadlines per second.
When Don Lancaster was trying to figure out how to build 3-D printers and other "flutterwumpers" in the 01990s, his idea was to use a bigger computer to rasterize curves into a set of one-byte step commands that would be interpreted by an 8-bit microcontroller, for example in https://www.tinaja.com/glib/muse140.pdf, as I mentioned in https://news.ycombinator.com/item?id=42080682. His first explanation of this concept may have been his July 01993 "Hardware Hacker" column in Electronics Now (previously in Radio Electronics and Modern Electronics) https://www.tinaja.com/glib/hack66.pdf where he's exploring the possibilities opened up by Parallax's BASIC Stamp:
> One potential use for the BASIC Stamp is shown in figure four. I’ve been doing a lot of work with the stupendously great PostScript general purpose computer language. In fact, this is the only language I use for all of my electronic design, pc layouts, stock market analysis, schematics, Book-on-demand publishing, and just about everything else.
> All the camera ready figures you have seen here in Hardware Hacker for years have been done by using nothing but my word processor and PostScript. Device independently.
> The only little problem has been that PostScript I/O tends to end up a tad on the skimpy side. Usually you only have three choices: Dirtying up otherwise clean sheets of paper or plastic; writing files to the hard disk; or returning your data back to a host for recording or other reuse.
> The BASIC Stamp can instantly let you extend the genuine Adobe Level II PostScript to any personal project or machine of your choosing!
> Assume you’ve got a homebrew machine that has an x-axis and y-axis stepper, an up/down mechanism, and a "both steppers home" sensor. This can be a vinyl signcutter, engraving, or embroidery setup. Or an automated printed circuit drill, a wooden sign router, or a Santa Claus machine.
["Santa Claus machine" was Theodore Taylor's 01978 term for a 3-D printer; it's what Lancaster consistently called them in his columns over the years.]
> You could use two of your BASIC Stamp lines for RS423 serial comm with your PostScript printer. Use two lines for both x-axis stepper phases. And two lines for those y-axis stepper phases. One line for pen or drill or whatever up/down. And a final line that zeros only when both steppers are in their home position.
> The hidden beauty here is that all of those fancier PostScript fonts and the level 2 tools immediately become available for use on your own custom homebrew rig. At unbelievably low cost. With zero royalties!
Normally when people do diagonal lines and arcs on computers without FPUs, they don't use floats; they often don't even use integer multiplies per step, but rather strength-reduced algorithms like Bresenham's algorithm and the DDA algorithm Bowyer talks a bit about in
Something You have not done yourself or something for what You have not paid to someone else to do.
Life is complicated, it is very typical when appear opportunity but nobody use it. Because to use opportunity need 3 things:
1. The will. 2. Enough qualification (or enough IQ and time to learn). 3. Free (!!!) resources or cheap borrowed funds.
If only one thing from list is missing, it become extremely hard to use opportunity. For example, I live in Ukraine, and even without war, regulations in country are so prohibitive, that many services are not enter country or enter with severe limitations, like Paypal only enter with private accounts, but business accounts are not available in Ukraine.
Yes, You understand right, idea cost nothing, but ability to implement ideas worth billions if find fertile environment.
And yes, I spent lot of time to find ideas and I could share with You high level information on how to find (or generate) ideas. Just let me know if this theme is interest for You.
But well, as you ask, for about prohibitive regulations, we in Ukraine made (and making) two things.
First, we are practically motherland of anarchy, and was few multiple years periods, when country live without government and without regulations. You could read about deregulation in books of
https://en.wikipedia.org/wiki/Daron_Acemoglu
and
https://en.wikipedia.org/wiki/Ludwig_von_Mises https://en.wikipedia.org/wiki/Murray_Rothbard
also I myself like Ostrom, Elinor (1990). Governing the Commons: The Evolution of Institutions for Collective Action.
Second, when government become powerful, we read prohibitive as it is - we just avoid made business in prohibitive niches. Because of this appear paradox, government and armed forces said that country need rockets, they even allocated large budgets to buy rockets from private companies, but nobody make rockets as they are prohibited.
In some cases, possible to make virtual business or semi-abroad - it is possible to create business in US or in EU (or other countries with adequate regulations) and hire people remote.
As conclusion, I mean, the best way to deal with prohibitive regulations is just flee abroad, to country with more adequate regulations, and you could from abroad support deregulation struggle if you wish. But if it is hard to flee, you could try to open business abroad.
For other opportunities I writing email.
Well, must admit, this is not only way, but most effective.
Also possible to use opportunity in educational environment or at community or in government organization. But from organizational science, all of these are just special type of business - educational business, or business owned by community , or business owned by government.
And yes, many people are practically running business, but don't think it is business. And this is very serious problem for economy, because they use this opinion "not a business" as consideration to avoid vital business transactions, and thus killing opportunities.
Yes, I agree, providing more info you become more vulnerable.
No, at concurrent market appear more priority matters.
You sure hear about Queueing theory, and may be you know, when execution unit wait for something, this time subtracted from max performance, and in worst case EU could wait 100% of time, so performance will be zero, no matter how fast EU could run ideal case.
How this apply to economy - business do many things without guarantees, only based on trust ("business is all about trust" - you could make big table with this phrase and hang it on the wall as icon).
In low tech business, typical to invest large resources to stored things, so people could just pay money and got product immediately, and do not wait until it will arrive from fields.
In high tech business appear additional dimensions, investments made into r&d and into prototypes of products, before mass production.
Services are just genuine trust, as in many cases you cannot know if service suitable for you before use it, so you should buy small amount of service or service business should spent some amount of it's resources to make gifts to try.
Any way, in many cases, need significant amount of borrowed money to start business or to scale business, and the more trust have subject, the cheaper will be borrow for him.
Statism is usually considered State as intermediary, who in theory should guarantee deals and even borrow cheap money, but this scheme is very limited and vulnerable for not good enough persons, because State itself must got resources somewhere to store them and to have something to share. If state have no spare resources it cannot guarantee anything. Not good persons use lot of tricks to force people to give their resources to State, and in extreme case just use violence to pick up resources from their owners (sure, this is bad for trust and killing free market).
Ideal business as I said before, called for trust. For example exists Islamic banking, which is extremely similar to idea of Venture Fund, and invented because for Muslims prohibited to give money for percents. So they create business with acceptor of money and got share in this business, and involve in this business to control that money will be spent wise.
And the last I should say - some people think, business could grow on internal money, just constantly withdraw some small amount from money circulation and reinvest these money into business grow.
Well, this scheme is working, but exists one problem - many business opportunities are limited in time, exists sort of time window (you could read more on this in ballistics or in space science, they use term as usually exists positions of planets, most energy effective). You could use such opportunity only if you will do things fast enough, so create your product or service before time window closed.
And grow on internal resources is very slow, usually it is just few percents faster then inflation, but to finish inside good time window in many cases need large investments, tens percents of turnaround or even more than turnaround, so without cheap borrowed money business will lost opportunity.
But to borrow money need to show clear enough organization; you may hear about Due Diligence, but I think you are not know, that Due Diligence is usually much more strict and deeper than tax reporting. So conclusion - to have more success in using business opportunities, you must share much more information than to just deal with state regulations.
Things looking even worse if we look on insurance and on finance transactions - typically insurance and finance transactions fees are depend on trust, so very high probability, semi-hidden business will constantly lose few percents of turnaround, and will be even more limited in using opportunities.
More detailed on this you could read in books from Acemoglu, Mises, Elinor Ostrom.
It depends on current market of considered niche. For some cases, market price is so low, every additional hole or trace on PCB could kill economy; for others, could include something like SGI or Cray as controller (mean, something severe overpriced) and it will be still profitable.
Using no other parts, at least none others that would not also have been available in the 1970s?
https://m.youtube.com/watch?v=yz_DKXIuL8U
However getting the right thermoplastic filament would be a major challenge, and slicing would require a high end mainframe with graphical terminals. Everything would be prohibitively slow. Also regular 2D printers were much more expensive back then.
I'm not trying to downplay the reprap team and their work did lead to some of the innovations in home 3d printing, but I suspect a lot of the project is like this, losing time by ignoring things that already existed and trying to reinvent things using cheap household items by going down unnecessary rabbit holes. Thank god that people serious about actually making a shippable product actually got involved at some point or it'd all still be theoretical.
The point at which people serious about making a shippable product got involved was already after it was no longer theoretical.
Honestly that's the normal way to do product design, start with the expensive version that works and then slowly redevelop it to use cheaper parts, or make big enough purchases that the cost of the parts naturally comes down.
I get that you were involved in the project, and won't hear of any criticism, but people serious about shipping instead of tinkerers would have done things differently, and there is no reason to deny or get upset about that.
Evidently the normal way to do product design failed at producing consumer 3-D printers, as it invariably fails at big innovations, so in those cases people have to be tinkerers to be serious about shipping. Clayton Christensen has a book about this you might be interested in reading.
Humanity is not one subject. It is multitude of objects and subjects. The best analogy is loose coupled network (each active node have ~10..100 connections, and billions passive objects with 1..3 connections).
Because of this, we could talk only about some subjects, who directly interested in use opportunities and have all need for this activity.
Even if use relatively conservative approach and consider only formally declared countries as subjects, we will have about 200 subjects (plus some active and successful humans - for example, in business usually considered 80% value of company is good CEO). If we look slightly deeper, we will see some trans-national subjects, like G7, G20, ITU, and many regional subjects. And you should understand me right, before We (I and You) create some at least voice agreement, we cannot consider any subject "we".
Yes, exists "invisible power of market", and some other similar things, like libido, but they all are unconscious powers, and you cannot expect for them to do conscious decisions and conscious things.
Therefore the diversity of humanity, the ITU, libido, and so on, are irrelevant to my claim.
Ex ante, there was nothing distinguishing the proposition that a usable 3-D printer needed precision-ground leadscrews from the proposition that a usable 3-D printer needed a heated build plate. By trying hard to do without, we found that the first one was false but the second one probably true.
https://en.wikipedia.org/wiki/Toshiba%E2%80%93Kongsberg_scan...
Yup I miss that too. Maybe all the appliances came only in avocado green or harvest gold, but they held together and did not spy.