While Hack a Day’s modus operandi is serving up hacks from around the Internet, sometimes we feel the need to exercise a bit of editorial freedom. A thousand words is a bit awkward for the front page, so feel free to skip the break and head straight to the full text of this article.
It’s no secret myself and my fellow writers for Hack a Day are impressed with the concept of a personal 3D printer. We’ve seen many, many, builds over the years where a 3D printer is a vital tool or the build itself.
Personally, I love the idea of having a 3D printer. I’ve built a Prusa Mendel over the past few months – Sanguinololu electronics, [Josef Prusa]’s PCB heated bed, and a very nice Budaschnozzle 1.0 from the awesome people at LulzBot. I’ve even made some really cool bits of plastic with it, including the GEB cube from the inside cover of Gödel, Escher, Bach (a very tricky object to realize in a physical form, but not a bad attempt for the third thing I’ve ever printed, including calibration cubes). Right now I’m working on the wheel design for a rocker-bogie suspension system I hope to finish by early August when the next Mars rover lands. My Prusa is a wonderful tool; it’s not a garage filled with a mill, lathe, and woodworking tools, but it’s a start. I think of it as the Shopsmith of the 21st century.
Lately I’ve become more aware of the problems the RepRap and 3D printer community will have to deal with in the very near future, and the possible solution that led me to write this little rant.
Infinite variety in infinite combinations
If I were to start researching what 3D printer to buy or build now, I would be faced with even more options than I had just one year ago. Even if I limited myself to a printer under the RepRap banner, I would need to choose between the Mendel, the MendelMax, Prusa, or Mendel90 derivatives, the Huxley, the Printrbot-based Wallace, this neat folding RepRap, and even one that can print its own case.
Outside of “official” RepRaps, there’s the Makerbot Thing-O-Matic, Replicator, and the deprecated Cupcake CNC, as well as Makergear Mosaic, Ulitmaker, SUMPOD, Printrbot, and the UP! 3D printer, among others.
Keep in mind that this list is not complete and only includes the hardware for molten plastic printers, not laser-cured UV resin or laser sintering printers. There’s also the issue of a dozen versions of electronics boards, a handful of firmwares, and a few types of host software available for each and every machine. In short, the 3D printer community is fractured nearly beyond repair.
What we should do
The father of the RepRap, [Adrian Bowyer], came up with the idea of a biologically inspired self-replicating machine nearly 10 years ago. The RepRap, as originally envisioned, was a physical manifestation of a Von Neumann Universal Constructor. [Bowyer]’s idea of a 3D printer serving as the foundation of a Universal Constructor models biological life and evolution; a machine designed to make parts for itself ‘evolves’ with each generation being more capable than the last. This idea caught on, and the huge abundance and variety of RepRaps and other 3D printers is like the diversification of flora and faunae of the Cambrian explosion.
This diversification is not without evolutionary dead ends; useful add-ons like the automated build platform are compartmentalized and only available for the Cupcake and Thing-O-Matic. Small keypads for computer-less printing are not always interchangeable between electronics boards and firmware versions. Obviously, there needs to be a standardization; like a comet smashing into the Yucatán peninsula, it may be time to wipe the dinosaur designs off the map.
What will happen anyway
In [Adrian Bowyer]’s vision of self-reproducing machines, we serve as hosts to the virus-like RepRaps that copy themselves and are assembled by humans. We are the evolutionary force that guides new generations of self-replicating machines. It’s time for us to start acting like intelligent designers.
What I’m suggesting is simple; have the RepRap Core Team define standards for officially sanctioned RepRaps. Throw a few designs up for review, like the Prusa Iteration 2, the Makerbot Replicator, and the Printrbot, and pick the best half-dozen machines from the lot, and put the RepRap name on them. Repeat that every six months or so, gradually improving each model along the way. Put the faith of the RepRap community behind every machine. Make “RepRap” a generic trademark. Put value into the RepRap name so in 50 years everyone has “a RepRap” on their desk instead of “a fused filament fabrication rapid prototyping device.”
What’s in a name
With a half-dozen designs for 3D printers, what do we gain? The combined development skill of hundreds of hobbyists, tinkerers, and makers from around the world for starters. Hundreds of people working on a hundreds of designs brings is a linear development process, and doesn’t benefit the project as a whole. It’s a piecemeal approach that invests too much time towards reinventing the wheel.
Hundreds of people working on a small handful of designs turns the linear development into an exponential development process. A limited set of possible machines means more improvements per model for every generation. If 3D printers are “breeding” exponentially, it follows they should improve exponentially.
Not only would we get a more directed development, but it would become feasible for someone to hand the models of a set of printed parts to an injection molding factory in China without worrying about ‘next week’s model.’ With a set of standards, RepRap kit makers can invest in the tooling to mass-produce kits instead of relying on a shelf full of printers to make the next generation of kits.
Parallels to Parallax
There is a remarkable similarity to what I’m suggesting and another well-known open source project. In 2005, this was the Arduino. Just take a look at how far the Arduino has come in seven short years: there’s a USB port, a better microprocessor, a huge library of example code, and a reduced parts count between the Arduino Serial and the Arduino Uno. Development on the Arduino still continues, and now there are Ethernet Arduinos and Motor Shields that provide a standard library of hardware and software for others to improve upon.
Nothing about the Arduino is new; the BASIC Stamp had been wrapping a microcontroller and peripheral components in a single package for years before the Arduino. Now, the Arduino is available in every Radio Shack while capable hardware like the Handy Board is relegated to educational pricing and the closets of schoolrooms. All this was due in part to the Arduino team defining a standard board and IDE. Without this, the Arduino would not be what it is today.
The RepRap can easily become the next Kleenex, Xerox, Thermos, or Aspirin. All it takes is a benevolent dictator to decide what is best for all of us. As with the Arduino, development of peripherals and improvements will continue. The best printers will rise to the top, and we’ll finally become the intelligent designers a RepRap allows us all to be.
Alright, rant over. It’s just an idea that I’ve been throwing around for a while. If you’ve got a rebuttal, or would like to add your own thoughts, drop a note in the comments. We now return you to your regularly scheduled build posts.
I couldn’t agree more. Core standards are important to foster creativity.
I agree, this fracturing is actually what has kept me from pursuing a 3d printer. I have other projects I’m working on and don’t want to get stuck with a dead end(unsupported) device.
I agree as well, plenty of very good projects have died the death of a thousand forks.
Perhaps what’s needed is a better way of steering the evolution of a few core machines – so that instead of someone saying “I don’t like X about that, so I’ll re-invent it slightly different”, there is a clearly defined and easy path to feed back the desired change & have it incorporated in the next generation/standard.
The flip side of this is that there’s always someone who wants it to also play MP3’s or serve coffee and those people need to be filtered out and politely pointed somewhere else. I’ve seen a few projects bloat themselves to death by trying to please everyone all of the time no matter how off-base they are.
Kudos for bringing this up, I hope the community takes notice.
Yes yes, we must purify the supreme race of self-reproducing machines…
But seriously, it’s still far too early for that. Give it 10 years and see what evolves before we start pruning the family tree.
This is the truth.
My gut reaction was you shouldn’t stifle the evolutionary process. But then I realised this idea wouldn’t stop individuals from developing new branches of the tree as they do currently.
Just look at the Linux kernel, for example. There are tons of variations out there, new features can be added by anyone but only the best get the official seal of approval. The new official branch is then used for the next round of free-for-all improvements.
I fully agree with this. I was considering building a printer, saw the options, and realised I could spend the rest of my life figuring out which option was best for me. End result – no printer.
I couldn’t disagree more.
The irony of choosing to draw parallels between biology and machine evolution, and then selecting Intelligent Design as the correct path (over natural selection) is just wild, given how the two perform in nature (as only one of them exists).
The only upside to this argument is that by taking the ID approach we limit the machines capabilities to only what we (as a committee) can imagine, preventing them from evolving beyond the limits of our comprehension (if you consider this a good thing).
It wouldn’t necessarily hinder growth and advancement. Much like evolution, there are standards that have to work with the others. Just like your genetic mutation has to allow you to breed with the others in your species before your mutation can become the norm.
I can’t read brian’s mind, but I don’t think he was suggesting we completely stop people from innovating outside the community. I think he was suggesting a more standardized approach to make it more accessible to a greater audience.
then again, I could be wrong.
Are you seriously try to spin an evolution/intelligent design dichotomy when 3d printers are by definition intelligently designed (by humans)? You can think of technology as evolving like a living thing, but it’s really not. There are people behind every step forward, and the hivemind model of group dynamics just doesn’t work.
The larger issue, I think, is that standards need to either grow organically, with people throwing their resources behind whatever they think is the best way forward, or be handed down by some overarching power structure. I just don’t think it’s useful to say “we need a standard” without then proposing some way to actually enact it, or at least suggesting what it might be.
You can discuss which tech is the best in this or that respect, or you can speculate as to what will eventually become de facto standard, but I think the only thing you anyone can actually DO is set out to try to make the standard and hope they don’t end up adding to the problem.
To quote the always useful WikiWikiWeb:
“Standards and methodologies don’t work!” says the developer.
“How would you know?” says the Standards and Methodology Guy. “You have never followed them.”
“Neither have you.” responded the developer.
Hey BFB, I’m not sure if you’re replying to me or the author of the article (I didn’t start the Evolution/ID analogy, that came from the post’s author) but I think you and I are on essentially the same page.
Moreover I’m not saying that creating a standardized design isn’t a good way to make 3D printing easier and less confusing but I don’t think that’s what the RepRap project is about. If that’s what you want as a consumer there are other options (Makerbot comes to mind immediately) if you want to buy a machine based on a standard (although continually improving) design using standardized parts and supported by people who know everything about that particular device.
Oh dear, waiting another 10 years before standardising. Horrible thought. Better get in early and have a roadmap and upgrade schedule. That way you can make the machines have a lifecycle, and an upgrade cycle. It’ll lead to less splintered development hopefully advancing it more quickly by growing a larger more consistent community.
A life cycle and upgrade cycle?
We need to avoid bringing in planned obsolescence. There’s no sense in doing that.
Someone mentioned the linux kernel above, that has a life/upgrade cycle, but is far from planned obsolescence.
OK so a 2.6 kernel might not have all the newest features of a swanky new 3.2 kernel, but it’s hardly obsolete
+1 on the article
As long these “standard” stay open eveyone can help RepRap becoming better and more popular
Else im a big fan of standardization
I agree as well. A friend bought a Ultimaker and spent so many hours and days trying to figure out the best extruders, software revisions, settings, onboard firmwares, etc just to get the advertised quality.
Sooo many variables(well its a complicated machine) but still I think its more important to have proper documentation and an intuitive website that is maintained regularly.
Its a curse that we engineers have issues with documenting stuff properly.
In the Ultimaker’s defense, I assembled mine in a weekend and was printing things almost immediately after assembly.
However the documentation isn’t great, and it’s almost chance whether you’ll get stuck in a pit or happen to have a good experience like I did.
I suggest following the mailing list before or as soon as you place your order, and immediately bugging them when you have a problem.
We didnt have too much trouble assembleing our Ultimate either.
The big issue really was printing in ABS is fiddly and thats what it came with. PLA worked much much better.
As did upgrading the firmware to mendel (use the Build-me thing on the googlegroup) and using Slic3r.
We still need to tweak things though.
In my defense. I have an Ultimaker. Software options sucked.
So I started to fix it.
https://github.com/daid/Cura/wiki
It has made many ultimaker users happy so far.
I think it’s a great idea, there are far too many printers and far too many pitfalls, so far from the natural selection doing it’s thing, it’s pretty stifled. As a potential 3d printer buyer, I am put off by the number of different machines out there and different reports on quality. I@ve hung around with enough enthusiasts that have bitten the bullet and been through many machines before they found something that was ‘good enough’. It’s supposed to be a hobby/semi-pro hobby to allow you to print stuff, can’t do much in the way of printing because you’re caught up in the minutae of such fractured designs.
Indeed. I am in a ‘paralyzed by choice’ mode also. Which 3d printer should I invest my time and money into, given the rapidly changing field?
Makerbot seems a good choice, also the Prusa or the Printrbot… Then, what happens when I buy / build one and updated electronics come out in a month? Or a tweak in the design for a larger print area? Or a totally new model that is faster, or more accurate?
I think the process should start with defining certain ideal goals. i.e. What should be the max size of a print before you start talking about the ‘non-desktop’ model? Will the printhead be developed separately from the base? What MUST this be able to print?
Or, start super easy: All bolts used are M3. All holes are drilled with a 3mm bit. Et cetera…
Your statement just made me realize that this is essentially the same problem that standard consumers face when it’s time to upgrade to a new computer. Which processor do I want? How many pixels in my resolution is good enough? What form factor? Etc.
One day, a Steve Jobs will rise amongst the 3D printer community, and tell everyone what they want, and how they want it.
Then everyone will hate the 3D printer. Except for the fanboys.
Makerbot is already going the way of the way of the Steve – they are pushing for mainstream, heavily investing in advertisement and have only two (main) products – regular or large, that’s it.
I partially disagree with this article, as I am very fond of natural selection. Using the Arduino as a peviously mentioned example: yes, the Arduino is very standardized now, but that isn’t because of the benevolent dictator theory (in my opinion). It’s because the platform was well-advertised, extrememly functional, and simplified. And as far as the standardized board goes, that isn’t necessarily true. On my desk, I have the Arduino UNO, and the Arduino Nano. SparkFun has the Pro, the ProMini, the Mega, the Uno, and 3.3 or 5V. Then there are Netduinos and other branches. I was baffled by which board to buy when I started, because of the amount of options, but that is a good thing, because I can get exactly what I want/need.
This was not supposed to be a reply to your comment.
Evolution dictates that the best for the species survives, eventually. I am in the boat of which machine/language/design to go with.
I used to build my own specialty tools for almost everything, cars, models, electronics, etc. But now I have figures out that re-inventing the wheel everytyme I want to make something was getting way too tyme consuming.
I agree with the concept of ID, to a point. Just like with biologic evolution, if a branch becomes viable, it will survive to be incorporated into the main tree. Like Linox.
As long as it stays open source all the way down to the nuts & bolts, it will continue to develope and grow.
Then comes the idea that if your only tool is a hammer, every problem looks like a nail.
I’m not intending this as a criticism of what you wrote, just sorta stealing your idea and going a bit further or in a different direction.
What I mean is, if XYZ becomes “the standard” and has a limited resolution, or uses only a certain or limited extrusion materials, or maxes out at 10x10x10cm then a lot of people won’t buy/build it because it wouldn’t meet some specific criteria for a build. And XYZ may be incapable of making the next big step without compromising backward compatibility.
Another question to ponder, would a standardized printer be designed around open hardware?
I would not use the term Intelligent Design. Think of RepRap and humans as an ecosystem. They each interact and affect the others “evolution”. Standards are fine as long as they do not constrain an evolving system. We can end up with a situation where the standards “protect” designs that follow the standard no matter what the limitations of the design.
The problem I had with the RepRap is that vendors flat out don’t help out. I bought a lasercut Mendel, got a dead motherboard, shipped it back, made the mistake of leaving a positive feedback on ebay then, and never got the replacement.
Sounds more like a bad ebay experience – are there reports of problems with other RepRap vendors?
Tbh this is a load of nonsense. People all have different priorities and aims which can’t be represented by one single 3d printer to rule them all. Multiple types allows people to go off on different tangents and research ideas, the best of which get copied and incorporated into the other designs naturally. And printers that become too niche will get left behind.
I look around my office and no one is wearing the same clothes, no one drives the same car, no one has the same stuff on their desktop. The problem isn’t that there is too much choice but that people are too worried about making the best choice as opposed to a choice. At some point you have to take the plunge because without learning a bit from experience you won’t understand enough to know which design is best or what extruder to use. A lot of this is really only because it is quite an expensive step to take and no onw wants to get that wrong, maybe we need simpler, cheaper beginner printers or maybe we just need to wait until there are cheaper options altogether and this won’t be so much of an issue.
Sure Bob, I hear what you’re saying, but when you look around your office, there are hundreds of items within your field of view that have all been designed around some core standards. I think what the author is getting at here is having a set of core standards that don’t forbid branching, but that by their nature keep any branches moving in predictable (or at least logical) ways. The “multiple flavors of Linux” analogy is very apt.
Well as was mentioned in the article, arduino. If someone asks what they should buy to get into microcontrollers, I think most of us would suggest a arduino. There are different boards out there that are better for certain tasks, but the arduino is good for getting started. It’s not too expensive, it’s easy to use, and there is a large community behind it if you need help or questions answered.
I think the same can be said for 3D printers. It’s not that there should be one printer to rule them all, it’s that there should be a sorta universal model everyone points to when asked “what should I buy to get into 3d printing?”
As far as what the ‘best’ should be, I think the big thing is price. I have seen kits at various prices starting at the cheap $300-$500 range to much much more. This is similar to micro controllers that cost as low as $5 (ti launchpad) to hundreds of dollars. We want a compromise of functionality and price. Really just a starting point, to learn about 3d printing and what it can do. Why that $5000 printer is so awesome for certain things.
We just need a starting point so I can say “I want to get into 3d printing, what should I buy? How much will it cost? How long till I print my first object?”
Your Arduino statement is not so true. It is true for microcontroller boards overall, but telling someone to “just buy an Arduino” is a daunting task in itself. Which Arduino do I buy? If I go to SparkFun and search Arduino, I get more than 20 results. Now what?
http://www.sparkfun.com/search/results?term=arduino&what=products
Bob, Just a counter point, the automobile really didn’t “take off” until Henry Ford made an affordable “standard”. Now, your office mates drive many different cars, but if Henry hadn’t happened, maybe only 1 or 2 would own one.
let it begin
I’ve been reprapping for about 2 years now (including the build/debug time of my bootstrap kit, urethane molded parts before clonedel)
I think this is totally necessary, and the linux kernel is a good analogy of a solid model. I also think that 6 or so ‘official’ designs is a good direction. It would allow for various combinations of goals. Cost vs accuracy vs speed, etc. We would see a number of design variants (hopefully with some parts even interchangeable between them) that would, if nothing else, help steer people new to the community in a direction that best suits them.
From my own experience, I bought the cheapest bootstrap kit I could find and spent over 8 months ironing out design flaws to get a working machine. (you know, some of us have day jobs) Had I the option of a machine like Wallace, or printrbot… I likely would have had a much easier introduction to 3d printing.
Excellent article!
@Bob, a single machine was never suggested. And a handful of designs designed to satisfy the various needs, along with the community designed upgrades that have made RepRap what it is today would never end up in everyone using the exact same machine. Get real.
Great Article!
I’m not sure if I understand all of your points, or necessarily agree, but you bring up some issues which undoubtedly need addressing.
Inkeeping with the biological metaphor, since the article mentions exponential growth, I might as well mention that as such 3D printers might face a Malthusian Catastrophe! All kidding aside, the greatest check on the popularization of 3D printers is the costs of the various non-printable parts (nutrients, in RepRap parlance).
In terms of designing the next model iteration of the RepRap, in my opinion, the following criteria are either essential or at least would be highly beneficial.
1. Replacing the steppers with servos. Servos of high enough quality are generally more expensive than steppers, but I think they have many more benefits. More importantly, I can’t see why it wouldn’t be feasible to fab servos ourselves. The control circuitry is covered by things like OpenServo. The OS board is generally retrofitted into cheap analog RC servos. I don’t see why building something like one of those is out of the realm of possibility. They’re mostly made of plastic, along with some copper wire and permanent magnets.
2. Figure out a way to use ABS pellets instead of filament. The key is cost; pellets are significantly less expensive and way more available. This may not be practical, I don’t know enough on the subject. However, if it is, I think it would be very helpful
3. Use an extruder design like the Ultimaker uses. It’s objectively the best there is, in the DIY 3D printer scene.
I’m sure there are other things to consider, but those three come to mind.
Standardization is a must: to say this market is fragmented would be a massive understatement. Far from giving neophytes options, the vast array of different printers which require experience to intelligently compare and contrast is bewildering and frustrating to someone looking to build the best there is (for their specific domain of interest anyhow).
The area where I disagree is on the generization of the RepRap name. Not that I think it is a bad idea in theory, but I’m reticent to dilute RepRap project, and it’s philosophy with Makerbot and their ilk. I know they’re popular here and other places, but I’m not a fan. They come across, to me anyhow, as shallow profiteerers who are trying to “hipsterfy” the so-called “Maker Culture” or what have you. What I mean by that, is the cheapening of an authentic community by trying to commoditize and market it. I don’t like places like Hot Topic, to use an easy target, for the same reason: they take things I care about, ride the wave of their popularity, and mass-market them to people who care about being trendy regardless of what the trend is. Making money is not bad. In comparison, I don’t have a problem with say, Target, because Target doesn’t try to sell a faux “culture” along with their clothes. The thing that irritates me about Makerbot, then, is their obnoxious need to try and present themselves as something other than a business. There’s nothing WRONG with making money, but just call a spade, a spade. When Mark Zuckerburg spoke after the Facebook IPO, he made some insufferable claim that Facebook was about “changing the world” and that they “make money to facilitate change, not the other way around.” So, while not as influential, Makerbot Industries reeks of the same sort of delusional egomania. There is a market, clearly, for people who want pre-assembled printers, which is fine. However,
in a similar trend I’ve noticed in the admittedly much more psychotic audiophile community, the customers are panderers and apologists for the company that, at the end of the day, took their money. It’s like some sort of bizarre Stockholm Syndrome or something.
I disagree with your assessment of companies like Makerbot. To call them ‘shallow profiteers’ ignores the enormous amount of work they’ve put in to making 3D printing easily accessible. They run thingiverse, which *everyone* uses to share designs, I believe they do most of the work behind ReplicatorG, they sell decent-quality, fairly polished machines and *gasp* have actual support staff you can call. They even develop curriculums for schools.
You might prefer to live in a world where the only way to get into 3D printing is to spend months tinkering and dealing with awful fly-by-night sellers, and enjoy talking about how you were into 3D printing “before it was cool,” but that’s not something I’ll miss any more than I pine for the days when the only people who had home-computers soldered them together themselves.
steppers are better for this
I’m inclined to agree. And I’d add another bit… that everyone gets over the ‘self reproducing’ marketing hype and focuses on genuinely lower cost with greater utility. If you want progress, people have to HAVE these things, and plastic corner joints are not a good place to focus on bringing the cost down.
Want to impress people? Design one from part sku’s that can be found at an Ace Hardware. Shy of maybe one motor controller board with a usb plug.
Oh I totally agree. They really need to drop the “Self-Reproducing 3D Printer” moniker. It’s patently false in even the most liberal usage of the term “self-reproducing.” Moreover, it’s confusing and misleading because it gives people the false idea that RepRaps can be had for their equivalent weight in plastic.
I think it would be prudent to have the “officially sanctioned” RepRap on the main page, and new models supersede the old models as they prove themselves worthy.
It’s not a problem of too many choices, it is a problem of terrible documentation! Like so many open source projects where a million itches are scratched by a million users, finding the version for you becomes difficult if every variant is not properly explained and compared.
First, the ‘front page’ repraps simply need to be only allowed in with COMPLETE, ACCURATE, and STATIC documentation and parts lists/suppliers all in one place (like a single .zip file hosted on the domain). It doesn’t matter if there are 200 of them, if they all have complete documentation available you can at least have an excellent starting point. I’m guessing there are under 5 which actually meet these requirements as of now.
Then sort them by year (or half year.) If the design has been updated in that time, re-upload it with updated complete documentation (but keep the original in tact in it’s old location!)
Right now many of the designs are just hosted on github and you never know when you are downloading something finished or a semi-beta which may or may not work.
It would also help if the most popular reprap wasn’t actually crap.
Of all the 3D printers out there, the only one I tell people not to bother with is the Prusa Mendel. But somehow, in the reprap community, that one seems to be the most popular one. Sure the Prusa *can* work great, I’ve seen plenty of examples of great parts made on a Prusa on the internet, but in real life, everyone I know with a Prusa is “still getting it working” or “just needs to fix one more thing”.
RepRap does need a new printer design. As far as I can tell, it should be more like the Ultimaker. MakerBot is going that way, and the Ultimaker is proving to be one of the best 3D printers out there.
I’d like to see the entire RepRap community work on designing their own version of the Ultimaker that is even better. I’m already upgrading my Ultimaker to use a Greg’s Hinged extruder because its better. I’d like to see even more RepRap stuff used to improve the Ultimaker.
TBH — this is the exact process that evolution takes. Many, many, many, mutations occur in each cycle of evolution, only those that survive, ahem..um.. survive.
What you are proposing is that we decide the evolution of the 3D printer, rather than let the evolutionary cycle make that decision. It could be one or a hundred, those 3D printers that don’t reach the pinnacle of evolution will die off, just as sub-par game consoles died off, sub-par video formats died off, etc. There is no need for a crusade and ‘technical’ cleansing as it were…
Evolution will happen in its own accord, with out our help. Just go ask a cockroach if he is happy in his own skin, then go ask the influenza virus the same question. One hasn’t mutated in millions of years, the other mutates every seasonal cycle. And yet both continue to survive…
ok…I’m blabbering now..I hope you all get my point. If not I’ll blabber some more
Chosen evolution isn’t evolution at all,…it’s genocide!
inre: sub par video formats.
The Sony Betamax had better resolution than VHS, but it also had Sony behind it.
Many of the suggestions of the article (and in the comments I’ve read thus far) are good; but, BUT I have a somewhat reactionary response:
[snark]
OH no! Making things is hard. Kits are confusing!?!? Please, please, please people with greater intellect than I; make a better 3D printer that just works and I don’t have to think about.
I’ll take mine in a brushed aluminum case complete with a pulsating white LED.
[/snark]
The proposal is a good one. It’s not about limiting choices, it’s about an organization earning brand respect–and doing the community a service–by identifying the best branches of development, helping the good mutations to group together.
The author is proposing a role similar to the one Torvalds performed for Linux: providing a centralized “branding” function. The “official” Linux-branded kernel got a reputation for being mostly the good stuff you want, and not too much flaky stuff. This didn’t keep others from innovating; instead, it meant that innovations built on a common backbone, and could more easily be folded back into the mainline.
The idea of a “small n” makes sense, too. many projects use this idea to capture different “markets”. Linux has a stable and unstable kernel. Debian has stable, unstable, and testing “models”. This doesn’t mean you have to choose one; individual customers can choose to run stable debian plus a specific recent patch they desperately need.
I’m another one of the “paralyzed by choices” crowd. I’d love to be able to easily identify the “debian” equivalent of a 3D printer: the one that’s pretty stable out of the box, not too flashy, but reasonably well tuned for many uses. And I know it’ll continue to be supported with reasonable upgrade paths.
It isn’t the mechanical design that needs standardizing… that is the easy part, and Joe Schmoe can design that stuff. The electronics is what really needs standardizing.
Speaking as someone interested in building a printer, I can honestly say that the electronics, and the documentation of the options/types/pros and cons, is what is really holding back the projects. The information is scattered and there is no site out there that says using an Arduino is better than driving it from a slave computer, or a Sanguinololu board, or whatever.
Who cares what it looks like and how the structure-work is laid out, it’s the brain of the project that needs to be standardized.
If you really want to make a difference in the community, make a web site outlining all of the options and what advantages some have over others.
Electronics should be standardized on a black box (inputs, outputs, function) level only, not on a schematic/BOM level, IMHO. Electronic components ecosystem is changing too fast for latter option.
…and for the love of $DEITY, don’t set ‘640K is enough for anyone’ standards. Imagine how much time, money, and hair follicles would have been saved, if the internet would have started out with ip8, instead of 4. Don’t just make it extensible, make the extensions extensible, then make everything an extension.
+1 for the linux analogy; A large but finite number of distros seems to work there.
The best will become the most popular. The rest will merge. Survivors will fork. It’ll sort itself out. Give it a few years.
Having some standards will help guide development, but this isn’t really intelligent design. In nature, there are, in a sense, standards which guide evolution. But that analogy breaks down when you think of human whim and desire as a force of natural selection – it’s a bit of a stretch to make that argument.
So yes, having RepRap procure standards will center development on the cream-of-crop ideas from the community. Exotic and peculiar printer designs are cool and unique, but I would like to see an official RepRap development group release generations of design improvements in a concise and methodical order.
Otherwise you’ll have to deal with a whimsical “pseudo-natural selection” that is inefficient at killing off inefficient designs. Natural selection is brutal, slow and really good at making some efficient organisms. The “human-whim” variant is too forgiving to poor “mutations” of design and is extremely scatter-brained in developments.
To be honest, I’m not entirely convinced the biological/evolution analogy is the best way to express what the reprap is about.
It’s as if building a 3D printer that can itself spare parts isn’t cool enough. 3D printers are already mindboggingly awesome, you really don’t need to sell it beyond, “look at all the cool crap you can make!”
Actually, I’m sort of reminded of the hullabaloo over BEAM robots in the early 90s. Some guy slapped a ‘biology/evolution’ label on a bunch of logic chips and said, “look! it acts just like a bug!” Really, the BEAM robots were just building robots that followed sunlight and the like with a minimum amount of parts. Biology didn’t really play a part in these robots and the analogy was superfluous.
I agree, but in the mean whilst the wiki could just be improved.
Speaking as someone who is just getting started and has just finished making a set of working electronics it would be nice to just have some form of ranking on the wiki, perhaps have a page where people can state what version they have and then use these to rank them by total number of users.
The documentation for n00bs is just poor and needs more friendly guides. The rest is good and detailed, it just lacks overviews.
The issue I have with RepRap is that, from the sounds of it, the project was not begun with the idea of making a 3d printer than can be used for printing arbitrary things, but to make a machine that can make itself. I don’t care about the machine’s ability to self-replicate and “evolve.” Any evolving that’s taking place is just the product of humans messing about with the design and deciding to change things in response to what the human observes during assembly and use of each generation.
I think perhaps what needs to happen is, 1) we leave the Reprap guys alone in their corner of the internet to fantasise about living machines, 2) people who care about 3d printing for its own sake can devise their own standards under a different name, and 3) the name will become ubiquitous based on its own merit, not piggybacking on the Reprap guys, who aren’t really interested in making 3d printers per-se but self-replicators.
Its true there are a ton of designs out there and each has its limitations but as far as hobbyist grade 3d printers are concerned even though the designs are highly divergent, they are all kinda compatible because they are based on the original reprap principles.
Personally I have gotten annoyed with the limitations of the existing designs in build envelope size. I’m also not a huge fan of the the mchanics in a lot of them. With official reprap style designs the build plate moves in more axis than just the Z and the xy gantry printhead design is sadly not more common.
To that end I have designed yet another 3d printer. But its based on what I call a printing framework, allowing the reuse of the core components to build different size printers with various build envelopes. Not to spam but if anyone is interested the project page is at http://maxifab.blogspot.com
Reuse of components in subsequent versions is very important for refinement (“evolution”) process of physical projects and it should be recognized as A Big Deal.
Several comments praised “the Linux way” of doing development but there is significant difference between open source material objects and open source software – patching the software to get updated version costs virtually nothing, while significant changes in mechanical design may mean making a heavy investment into it.
I’ve had a few conversations with Adrian, and I can tell you he will NEVER sign on to this idea. He is committed to total openness of design and the evolutionary paradigm.
That said I think that a lot of the problems people are having are self-generated; they are trying to get the “best” in an ecosystem where “best” is not even defined, much less determinable. What makes an extruder “best”? Cheapness? Reliability? Mass? Precision? Open Source Design? Is it still the best extruder if it uses a filament size I dislike? Will there be a new and better version next month? And how are any of these things to be determined given that the designer/makers are already moving on to other things? Unfortunatly printers are still expensive enough that the desire not to get stuck is understandable, but really, the best bet is to decide and jump. Sometime in the next few years there will evolve some agreement on the components of a “good enough” system, and someone will start to put them in a box and sell them. But the cutting edge is not a place for people who want certainty and right now this entire field (hobby micro-manufacturing) is one big cutting edge.
Brian,
Very nice article, but I do expect a little more from HaD. Your article puts a fine point on the journey of advanced technology from enthusiasts to mainstream “across the chasm”. It also highlights the difference between a benevolent dictator and caucus. Clearly, it seems you have never worked on standards bodies where it’s amazing anything gets done. Remember the ATM standard (a compromise of 48 bytes instead of 32 or 64) + 5 bytes header.
Again thanks for highlighting important topics we all need to advance the community. This is a complicated and polarizing area IMHO.
“…it would become feasible for someone to hand the models of a set of printed parts to an injection molding factory in China…instead of relying on a shelf full of printers to make the next generation of kits.”
Isn’t the intent of the RepRap team for it to be as self-built as possible, so that anyone who completes a machine can profit from helping to make more? Seems to be an irreconcilable difference in philosophy to what you’re suggesting.
But I’m sure in ten years, Microsoft, Apple, or Google will be selling these. Then you’ll have the standardization and mass-production you desire.
Personally, the FDM machines don’t interest me much. They seem fiddly, limited, and way too expensive for what they do; which I’m sure are all reasons people keep reinventing them.
Using the Thing-O-Matic kit as an example (since it’s the first one I could actually find a complete price on), it costs $1,100. Seriously?
For $224 more, you can get a LumenLabs micRo V3 kit. Which has stronger servos, more accuracy, superior rigidity, and larger build area. It mills plastics and some metals, which none of these FDM machines will *ever* do. You can put an extruder head on it if you want, or anything else you can imagine. And although it won’t be as fast as a dedicated FDM, maximum flexibility from my investment is much more important to me. I can wait longer for an FDM part, if it means I don’t have to buy two separate machines. And quite frankly, the Thing-O-Matic would look like a cobbled-together eyesore sitting next to a sexy micRo.
So when such a vastly superior mechanical platform is available to build upon, the interest in and price of single-purpose FDM machines seems disproportional to me.
If the goal is to get a lot of RepRaps out into the world, I think it makes a lot more sense to capitalize on industrial manufacturing techniques to build the RepRaps.
Small example: You run a company that sells RepRap kits. Right now you’re cutting smooth rod and threaded rod stock by hand, one at a time. You realize, “hey, factories have a machine that will cut all these metal rods 10 times as fast as I can cut these. I should get one of these rod-cutting machines.” There’s nothing wrong with using current manufacturing technologies (like rod cutters, or injection molders for the plastic parts) if the end product remains the same. It’s really just a way to streamline the process, and I’d be shocked if I can’t buy Chinese injection molded set of parts for $10 within a year.
I agree. Use traditional mass-manufacturing, and you can build them cheaper and faster. You can probably also do away with many limitations and compromises in the design, which were necessary to make it more self-replicable.
But if one of the primary goals is “as much as possible, it builds itself”, then a mass-manufactured tool is by definition no longer a RepRap. And I doubt the core team is going to endorse any change this noble, but limiting tenet.
Trying to achieve self-replication using FDM alone is also limiting. micRo makes most of its own custom parts as well, but can do so faster, from stronger materials, and with much tighter tolerances. Although not a goal, micRo is actually more capable of self-replication than RepRap!
So I don’t think the RepRap can be “saved”, as you say. Too many self-imposed limitations, and it will continue to diverge as people struggle against them; rather than eliminating the limitations.
For what its worth, I have a full CNC machine shop at my disposal to use whenever I want for free (I design parts for my work, and program the CNC machines there).
I still bought an Ultimaker, and am thrilled with the parts it makes. Making things in the machine shop requires that I have the right size material on hand (I rarely do, and its expensive and takes a day to get anything), and it takes hours to program the machine, get all the tools together, grab the material, zero it, run the part, make any setup plates or special jaws I need, and then run the rest of the operations on the part as needed. Some parts require 8 operations, 20 tools, and special jaws that take two operations in and of themselves.
If I need one of something new, machining it can take forever. So I only do that when I need stuff out of metal, or when its for work.
At home, I use my Ultimaker all the time. I made a sweet coat rack that would be damn near impossible to CNC machine, as well as a cool modification to an IKEA lamp, and lots of other stuff I haven’t posted online. See what I have posted here:
http://www.thingiverse.com/tlalexander
I’ve found that the printed parts can be made to be about as strong as wood. I even made some wall brackets to hold an IKEA nightstand on the wall to hold up my microwave and toaster oven. They’ve been holding great for a couple months now, with no sign of sagging, cracking, or failing in any way, even when I regularly use my toaster oven at 400 degrees right under them.
I have some pics of that here:
https://picasaweb.google.com/106971352551810737566/Ultimaker#5702783249169846306
Sometimes there are still issues with the printing process, but often that is software. Look at this print, which looks like crap towards the top:
https://picasaweb.google.com/106971352551810737566/Ultimaker#5720578138001036690
I did fat layer heights because the part doesn’t need quality, so ignore the fact that the resolution was low – that was on purpose and the machine can do much better if I want (as low as 0.04mm instead of the .35mm on that print). Anyway, see how it’s noodly? Well, the top is noodly, but the almost identical feature on the bottom isn’t. That’s because the hardware is perfectly capable of producing that feature, but the damn software I’m using needs to slow down when the layers get small. Most of the available software supports doing that, but the software I’m using doesn’t. So I’ll probably switch. But the point is that some of these issues with parts are actually software related, so don’t always think these machines can’t do it – the software just needs to be improved, and people are actively working on that. But the technology itself is solid.
To, TL;DR – I’ve got 6 years of experience running a nice machine shop, and I’ve made lots of stuff, but I make more stuff for fun on my 3D printer than I ever did with the shop. That coat rack took me about 20 minutes to design, and then I hit print and walked away. Easy peasy. Machining something even close to that would take me hours at a minimum, so instead I would have just bought a coat rack, which is lame.
(viewed the pictures you posted of the parts and process)
Wow!
I’m smitten!
@Ren
Thanks!
Point of fact, you *can’t* get a micRo, not for a long time now. Lumenlab’s store is still (fraudulently) up, but they’ve admitted that they haven’t been producing or shipping orders for many months. They’ve hardly even been communicating with customers who’ve already placed and paid for an order for about a year now.
Evolution (as used in the context of the entry) about standardizing and branding to reduce confusion, is the process of change in a series of unrelated and unguided transformations to produce “a fused filament fabrication rapid prototyping device.” Wings appear where there were nostrils, so to say.
Intelligent Design – is a proposed goal with specified and guided steps to achieve the goal of RepRap version 5.0.x, or 6.0.x, or 7.0.x, etc. . . . “At 252 degrees activate the cooling fan. At 249 degrees move Z-axis up one step to origination point and begin deposition.”
As can be clearly observed, no fused filament fabrication rapid prototyping device can act, load, and reproduce autonomously despite the best wishes and loftiest of proposals. Somebody designed ‘em. Not even RepRap version 6.0.x is envisioned within the confines of the Arduino Uno perched atop the Makerbot case and just waiting for an uninterrupted evening.
Instead, the concept is but swirling around minds that think “Wouldn’t that would be cool?” The term “evolution” is so badly used when linked in context of the RepRap given that RepRap produces intelligently designed objects. Continued use of this term to describe iterations of a substantially different machine used to produce the same 3D-object is humorous at best.
Servos versus stepper motors, ABS pellets versus filament, touch screen versus committed buttons all read like future implementations which clearly would indicate intelligence is active. The blinding truth is that without very intelligent people specifically engineering very precise changes to the genre it would be nothing more than a complex way to melt plastic.
Of course, all of us have seen unguided mutations of RepRap occurring as can be evidenced by the sweeping herds of Mendel covering the plains in Montana as they convert the sunshine and grass-like substances into nascent Makerbots in a Cambrian explosion of innovation.
As for me, make mine Marvel.
I like the linux analogy as well.But I was was thinking about the early days of personal computing.Befor the IBM Personal computers , it was freakin’ nuts! , Cp/m . msdos , Gem , Unix , Apple , Coleco, Pet,and almost nothing worked with others , and often not even within similar systems.
Self replicating or not. The “best” or not, Until there is a leader , there will be no progress
What a load of crap!
First of all, the idea of ‘evolution of 3D printers’ is flawed, as 3D printers do not mate, they are designed and redesigned. Where in nature intelligent design does not occur, with machines no evolution occurs: you don’t get a new watch from rubbing two models you like together. So with machines, there is a designer and he/she is what hobbles the design. A design can not become smarter than designer, and a lesser follow-up designer will just detract from the design they are changing. A/the problem is that the consumers of these new designs are only mediocre judges on the quality of the design as they are mostly new users.
So instead of artificially limiting the breeding stock for showhorses, I propose we race them! Let’s create a competition where contestants are judged on a few measurables: speed, accuracy and strenght. A jury consisting of the top knowledable people in (and out of) the field are to judge the products of the contestants (instead odd semi-ignorant newcomers).
The gada-prize has spurred a lot of development, but because it is such a long running affair, a lot of teams have stopped before adding much to the knowledge-pool. A competition that takes place once or twice a year (cylce time for a new design) and forces winners to open their design while still allowing winner to profit from it, will generate much more progress, IMHO.
Your idea makes the most sense to me in this entire conversation.
We should be using competition and incentive in a way that inspires new design.
I agree, there is no biological system here to host an evolutionary process. It is purely design changes driven by users needs, those design changes should be measured against each other by experts and incorporated back into the main line. I love the idea of a competition, every other industry has publications or fan websites that do shootouts/comparisons of the top brands..
Ditto analogies to microcomputers in the 1970s. Dozens of operating systems, dozens of chips (remember the F8, TI9900, SCAMP, 1802…). But by the 1980s we were left with x86, DOS and the Mac.
The reason non of the systems worked with each other was that they were commercial closed source projects. Everything is a secret so the competitors won’t steal it. Even now much of the linux drivers are reverse engineered from original closed source drivers. It’s crazy, hardware manufacturers keep shooting themselves in their foot. Maybe android will teach them something.
Openness solves these problems.
Funny this complaint about “hacks” showing up on a site about “hacking”. You can write all you want about how to do this or that but the chaos of the “hacking” community by definition will never have standards. And the standards that are followed were not developed from a hacking community, but by organizations and companies who spent some serious time and money creating them.
Maybe try something like how RFC works with the IETF but for hardware/software? All that shit takes money to develop unless you can get someone to do it for free.
Sounds great and I would love to help, but I really don’t give a fuck about standards. I just want to print objects and not have to fuck with my machine for 4 hrs to get it to print a disgusting looking cube thing.
…annnd this is why there is a problem. The paragraph above is the chaos of the hacking community in a nutshell, which I think is great. (otherwise, I wouldn’t be on this stupid site, and you wouldn’t be able to get money from your trackers; adsense, g-+1, g-analytics, quantcast, sitemeter, wordpress stats.)
I think breaking up the system into modules could help development, but you will still need some kind of moderated central repository of the “best” ideas and solid/finished/working designs.
It begs the question “what is it that you want to do?” Do you want to build an XYZ system to help you model and make parts, or do you want to just model and makes parts?
As far as steppers vs. servos, just how fast can you lay down a material?
Either way, stardardized modules could be one way to increase the qtys and get much better components pricing.
agreed, in all of the facets.
we need some standardization. not (nessarily) to one system, but at least make it like EPP or even ECP, ECP being notorious for many fractured modes and such, but at the current state all the 3D printing is WAY more DISorganized then even ECP, and everyone knows how nonstandard-friendly ECP is.
summarize:
3D printing as it stands is like MKE/SONY/PANASONIC/ect CD-ROM interfaces… the cd-rom connectors from BEFORE IDE …
yeh, each connector its own system/software/commands/eclectricallayout and each needs it’s own drive too… ugh
FYI: schools mostly waited until IDE to buy cd-rom drives … cept that one computer in the corner for converting cd’s to floppys
if you want everyone to know about 3D printing you’d best make it into something non techies can pick out of a brocheure, not something involving a frustrating 3 HOURS picking and researching all the sub-systems!!!!!
some of us find that fun, as do i, but schools, buisnesses, repair shops,,, they just want to buy it, plug it in, fill up plastic resin tank, connect kinect, and hit the “COPY 3D” button! 🙂
car shops would make BIG BUCKS if they were “Just able to use the darn thing.” auto repair tech dont know f***-all about FET’s or how to compile the source downloads,
BELIEVE ME THEY SAY THEY DO BUT THEY DONT!!!
but they do swear like a sailor when a $250 piece of plastic breaks, costing them $250 PLUS S&H
… == super glue, scan+print, sandpaper, DONE! that’ll be $250 but 75% off for you madam! hehehe easy 62$ and it was “SAME DAY DELIVERY” 😀
PS: and of course it’ll be USB or serial+USB, in this case im talking about everything BUT the electrical port.
FYI all you disbelievers in the whole $250 piece of plastic(edit: sh**) thing can go to a car dealership and ASK how much for that shiny switch-holding-front-bezel for the lightswitches,,,
$220 plus S+H … reeealy? it’s true, and thats withOUT the actual switches that get mounted on it!
the only other option, short of printing it, is junk yard! which has it’s own ups and downs, and falls, and slashed arms on pointy and rusted jagged edges…
yeah, but all of you are ignoring the other part of making a 3d printed part–the model.
Models are an art and skill in themselves and I don’t see much talk of the hours needed to make a model before you can print out something useful. It takes time to learn a good modeling package to create a model in a CAD program.
Or all we all just going to print out the low-quality models that are posted to the internet?
Sooner or later you have to make a model to make something the is useful to *you*, or 3d printing remains just a curiosity.
As a 3D modeler myself, geometrix designs really arnt that hard to make. Certainly after a bit of learning you could save money by modeling them yourselves.(as long as your own time is valued at, say, no more then $30 an hour)
But, honestly, anything I make usefull I will slap online for other people to use. I think most modelers with 3d printers will be the same.
The idea that you have to model it yourself or it will be “low quality” is rather crazy.
The only question would be is the replacement part you need popular enough to be made already.
I have been watching the RepRap and other 3D Printers splinter in quite a few directions, some redundant. While that should sort itself out in time like many of the evolving projects of the type. I feel it would also benefit to try and standardize as much as possible, including file types and software.
While you may be thinking, “What’s wrong with an STL file?” It’s not a simple grab and go solution for the future. It would be nice to have a file that was self contained, much like an Epub, containing STL, Description, Documentation, Images, and ideally a precomputed Gcode for common printers and configurations, packed into a Zip store file or similar. Obviously the necessary Gcode could vary wildly for different configurations but it should be left available. The description and documentation would have the included Gcode settings listed. When the interface program loads up the file it would list all the data available and allow a new gcode to be generated if desired, then optionally saved back to the file. Versioning of models could be contained relatively easily since included documentation would not change, just a new complete file would be made instead, possibly with a changelog. With included Gcode the files would get relatively large fast unfortunately, but that’s another reason I figured it would be optional. It would simplify libraries of printable objects greatly. Who wants to always download the files over and over again?
Ideally this would eventually coalesce into a library/printer interface program, again paralleling Ebooks like the Calibre’ software has done.
It will probably start to sprout out of ReplicatorG sooner or later, if no other reason then I suggested it. 🙂
I find it odd that you would try to decide what the RepRap Project should do – and how they should do it.
If you have a great idea and vision for a standardized 3D Printer platform where future designs are voted in by a board… you are free to start a project doing just that (MakerBot anyone?).
Writing a rant like the above about an existing hack project seems very counterproductive to me.
Firat while I don’t mind this at the main HAD page, it’s one of those thing best suited for the HaF forums, although I realize there will be some who would bother with that. As I read the comments the Ford Edsel came to mind. A product of “Intelligent Design” from a company that had been practicing it for decades starting with the model T. Another often use automotive reference is a comparison between the ’55 Chevy, Ford. For that model year Chevy choose to stick with little innovation outside their V8. Ford choose to emphasis safety. Fords sales that year is often cited for setting back safer US auto back years. Like it our not US government dictates(ID)have brought vehicles that pollute less, use less fuel, are safer, and generally last much longer. Another issue is how easily much of the public is manipulated. When Merle Haggard sang a Ford or a Chevy last ten years as they should. Most of his audience where like ” damn straight”, with few saying that’s bullshit, they ARE lasting ten years. In the end “evolution” is slower and could cost more, persons, more money over the long run. While ID will produce a few failures, those failure are likely to cost the consumer less. For the an absolute end, why are we having this discussion? A 3D printer would be nice to have in a personal workshop, but the investment in a
in a 3D printer could be applied to other tools that can create, not every thing a 3D printer can, but can come pretty damn close to doing so. While having day to day utility where you can walk up to them to perform a task, with creating a file the tool needs to work.
Personally, I’m all for crowd-sourced standards, but I don’t think that the actual printer should be standardize. Instead, the standards should be for how the components connect to one another, physically, electronically, and digitally. That would allow anyone to design any component they want, and that component can be connected to any printer, and ‘just work’, both from a hardware and software standpoint.
I’m old enough that one feature of the ‘new’ MCA slots absolutely blew me away: unlike ISA cards, MCA cards were self-identifying!!! You didn’t have to know the exact make, model, version number, and have complete documentation for a card to have a hope in Hell of installing the right driver anymore! HALLELUJAH!!! IBM forced anyone who wanted to make MCA hardware to pay for an identity code, and there were many who protested, but that concept paved the way.
In the same way, I believe that one of the ‘RepRap standards’ should have hardware that self-identifies, and contains firmware that can be upgraded. The controlling software could take that identity and firmware data and automatically know what it needs to to use that component.
From a developer’s standpoint, such a standard would mean they could design one component that would work with multiple printers, saving them a LOT of time and money.
From a buyer’s standpoint, it would mean that they could choose components from multiple vendors, yet have assurance that the whole thing can be assembled and used relatively easily. They could also swap out components for newer/better ones.
One thing a connectivity standard wouldn’t do is stifle creativity; it wouldn’t matter what you built, as long as it played nice with others.
(In case you are wondering, yes, I was thinking along the lines of the USB standard as I wrote this.)
I don’t like “Identify yourself”, it is coarse and obsolete way. That means that host is supposed to know about every manufacturer and every model, or to chase the drivers and other knowledge about the part. Even worse, hackers are excluded from the club, which is very annoying.
Modules should carry all necessary knowledge about themselves with them, e.g. on a serial EEPROM or something. Nowadays large capacity memories are cheap, why not strap a SD card on a module and connect it with a small micro over SPI, load complete documentation and drivers (in some P-code or such, to maintain portability over different architectures), calibration values if applicable, as well as info tables for other modules to query.
While you make a good point about units having onboard storage for documentation/drivers/etc., self-identification is neither coarse nor obsolete, and the ease with which USB devices can be used is ample proof of that. Each component having a unique id is the only practical way to enable every part of a printer to be quickly and easily updated and integrated. (I envision a system like the Linux repositories, where current/past firmware, data, and documentation files are available for every component.)
As for hackers being locked out, please give some credit to those who would create such a standard; there is no reason why a system couldn’t be in place to give a unique id to anyone who asked for one. (The problem hackers have with the current USB device ID is that it costs $2000 to get one.)
Rather than worry about the evolutionary aspect, what about defining some targets that are meaningful to non-enthusiasts who just need to print parts and don’t want to get into the nitty-gritty.
Do we have an amateur 3D printer that even vaguely approaches the ease of use of a 2D printer? Not even nearly. Yet, if such a device were available, a lot of people who just wanted 3D “stuff”, without a detailed knowledge of polymer science or whatever, would be buying it and thereby injecting a lot of money into the manufacture of domestic machines, as well as creating a secondary market for things like downloadable objects.
I hate to use the term “marketing objectives” but, as several others have pointed out, this is a new area with a constant stream of new ideas, thus the freedom for technical innovation should be fiercely protected, not “channelled”, and the only way to reconcile this with the ultimate aim of, say, a 3D printer which requires little or no maintenance, and NO user intervention during a print, is to set objectives that relate to that hypothetical end-user product, and to ensure that the engineers have all the space they need to achieve them.
Product diversity isn’t necessarily a problem – Check out all the different Expresso machines on the market :).
In a number of ways the current situation seems pretty healthy. Firstly a lot of patents on basic stuff are presumably ending up on the open source side of the fence and, secondly, the only people who aren’t put off by the nightmare of choice that Brian despairs of in his rant are likely to be as interested in the printer itself as the product it produces, and therefore far more accepting of all the poking, prodding, cleaning, and disappointments that what are effectively prototypes are likely to deal out.
It seems to me that although Brian’s fears are understandable, affordable “fire and forget” domestic 3D printers are still a fair way off, so why add unnecessary constraints before we absolutely have to?
if you have a look in even the most basic 2d printer from the 80’s you’d see that the technology is not the issue here. if there was a market for these you could pick one up at the supermarket
Actually, it’s kind of funny that you would mention ‘ease of use’ in relation to 2d printers. If you had been around in the late 70’s/early 80’s, when ‘reasonably priced’ dot matrix printers were just starting to be found in homes, schools, and small offices, and had to get one of these infernal contraptions to actually *work*, you’d realize that 3D printing is in relatively the same place; every single printer is different, you have to have both an engineer’s and a programmer’s body of knowledge to get them to work *at all*, and, above all, the patience of *Job* to not just take a sledge hammer to the bloddy thing and have done with it.
Well, aside from this whole debate, I’m glad that everyone else is having the same problems that I am working out which combination of all the different parts to build one of the things from
thats what im taking from this also
Love the RepRaps.
Think it is an evolutionary dead end, however.
the future is going to be Direct Laser Deposition.
bronze, copper, brass, and some new stainless steels availible – all powdered.
the only things holding it back are- gotta do it in a vacuum, and have to figure out a wiper system to layer on the powders. A couple companys have figured it out, and are doing some stunning work.
We are working on asteroid mining, and coming to the conclusion that we are going to be 3D printing most of our ships! Nearly the entire ships can be printed, especially if you can use deposition channels instead of wire.
build the first ship, load a printer, and since most of Luna and the asteroids are aluminum/Mg blends, you just have to sift it, and start printing. after you have a stub, attach it to an arm, and just dip the top in the dirt, and hit it again with the laser.
Just pack a couple extra rocket engine bells/nozzels (you can use Al/Mg for fuel), some optics for nav, a radio, and an iphone for orb mech/nav, and you can have a full fledged miner for space metals.
Talk about bootstrapping!!
They are now doing some cool stuff with molecular building on copper substrates – using a indium tip to push around CO2 atoms to make your own graphene. Basically building superconducting circuit boards (look up molecular graphene.)
start saving all those old non-lithium batteries, they are going to be great sources for powdered metals!
I think we are aiming too low.
I disagree. (I’m a repraper too, and yes the machine i’m building is already outdated, and not even finished yet)
There are two important fundamental concepts that you are missing. It’s ‘free market’ and ‘open source’. Free market is a perfect evolutionary environment, the end user will vote with his money for the machine that suits him best. If he finds the choice too complicated he will look to someone with experience for comparative anaysis, like a test/review article. A savvy user will be a be to print the machine he likes if he makes a mistake in initial selection. Be it consumer products or politics don’t underestimate the intelligence of a n individual.
With open source/hardware ‘standards’ emerge defacto. If a program/algorithm/design is found to be useful by many it becomes a standard (like linux). You can just take the bits that you find useful and build on that, if it works well others will use it too.
I agree with what Glytch sort-of said. To paraphrase his idea…
Let everyone keep doing what they’re doing. 3D printing is still in it’s hacker stage anyway, the Heathkit or Mits Altair stage. 3D printers don’t come in a box, and nobody manufactures them whole.
But anyway, do what Linux does. Anyone can have their anarchic Linux, but if you want a version with a brand name, guaranteed quality, and a bit of support, you buy a commercial distro.
Basically we need a company to make 3D printers you can buy mail-order, and put together in less than an hour, without soldering. That, and some nice software and a few object data files included to start off with. So anyone who can use a screwdriver, can put the disk in their computer and print something.
And most importantly, if it doesn’t work, you can send it back! And the company has tech support! Just so you can be sure you’re not spending hundreds of $ on a piece of crap that you can’t get working.
My vote is that you all just deal with the variety and confusion, and stop trying to control this wild process.
Eventually, a clear choice will emerge. It will have the best features for a good price. Over time, the price will go down, while the looks will improve.
I think that it will become as common and as popular as 2D printers, and there will be at least 50 million of them sold over the next 5 years.
What I want is one where I select a part online, click on it, and a little while later, the part pops out into a tray. And I don’t want to pay more than about $300…
I really can’t see your point here.
Have a look at how many different brands of lathe, end mill, vertical mill and drill press there are.
Even the humble hand power drill has so many different forms.
Way back when Apple presented us with the very first LaserWriter, there was a standard.
It was called PostScript.
Within a year or so of the original LaserWriter there were dozens of PostScript printers.
And a boat load of Non – PostScript printers.
Notably that Epson abortion and HP’s weird LaserJet Series 1, which could take a font cartridge to give it PostScript capability.
Then we had phototypesetting machines like the Linotronic and the Agfa StudioSet 2000.
One thing remained the same…
DEVICE INDEPENDENT OUTPUT!
I can proof on a laser printer and KNOW that besides hairline’s the output will be the same.
If a standard for 3D printing is going to be adopted it should be aimed at device independent output.
The 3D printer Joe Blogg’s builds in his workshop should be able to print the same files as a commercially available unit.
The biology/evolution analogy is a real stretch!
Forget the evolution angle.
Think about lab rats, the strain that all labs use, why?
Because it leads to repeatable results, no matter where you are.
amen
Important discussion, keep it going!
A few points:
1. The evolution/ID talk has use but only goes so far. Some objecting are talking it much to literally.
2. linux and arduino are great examples to learn from.
3. good standards allow for intensified competition and cooperation, good standards select and reward what works while sticking to massively redistributed inventive processes.
4. items to standard: print dimensions, key parts, software, filament, life cycle, upgrade cycle, documentation
5. standard desiderata: easy assembly use repair upgrade, global availability, price AND open hardware values AND self-replication
6. self-replication capability should never be compromised even as external mass manufacturing is adopted. Compare: many use FOSS binaries but the source must be there for anyone to compile. We should stick to freedom a core feature.
7. Expanding on 6: 3D printing is a disruptive technology. Google, Microsoft and the entire Copyright/Patent Industrial Complex will try very hard to coopt, control, regulate or, if all else fails, ban the tech and reap rent. It is our responsibility to resist those forces and stick to open source values. Insisting on self-replication capacity operationalizes that resistance!
There are some nice thoughts. My comment:
We have a lot of “me-too” printers lately which confuses the newbies, the confusion is the only bad thing about that. Evolution is great, my mendel has 89 forks on GitHub, I pull good stuff in, so its pretty tweaked all the time.
What I personally think, now the biggest problem are lots of shady vendors raping our open source project. Some sell stuff like extruders, without even having printer and testing them, etc. Bad quality china plastic is also coming …
I am personally really curious what will happen this year.
Jo Prusa
RepRap core dev
Software standards by all means, but hardware standards no.
There is FAR FAR FAR too much room for improvement at the moment in all sorts of area’s. We need the experiments, the “out there” designs in order to progress.
We arnt yet in a era of minor tweaks being the best way forward, theres still all sorts of big possibilities out there.
For all we know current 3D printers could be like Nipkow’s disc compared to the cathod ray tube.
Maybe someone will come up with a way to print whole layers at a time? or some really sci-fi static based thing that replicates a whole solid object in one go.
The possibilities for both big changes, and minor refinements is endless at the moment.
The only thing I think would help at this stage would be spec-standards for the software to emerge;
-material spec data? (somewhat equivalent to “paper type” in a 2d printer)
-printer driver standard (that is, while each printer is unique they can take the same format input rather then needing machine-specific gcode generated by the user)
Citing the arduino as a good example to follow, while I see it as the worst example of snuffing down the creativity (and a growing number of people agree)… the fact that everything is now Arduino or Not is /bad/.. It forces everything to use one platform, including bloated IDE, libraries etc. In fact I call it the Arduidiot these days.
The fact he quotes the Prinrbot in the best variants to be selected actually makes the whole point: the Printrbot was born very recently of someone who didn’t want a ‘standard’ prusa.. and tossed a few prerogatives outta the window in the process.
* All for the best *
I agree that there’s too much stuff in reprapland, and thats it’s hard to get started, but it was the same with the 80’s micros, and in the end, it solved itself and did so by exploring the deadends.
Locking up standards now would help webshops and patent attorneys, not the community as a whole.
Hmm great post from a conversation POV, but I think the author maybe getting mixed up about what RepRap is.
RepRap as I understand it from Adrian’s talks and writing is a merely a meme with an active community (or evolutionary canvas). It is the idea that counts rather than the individual generations that form physical incarnations, perhaps a reference Ricard Dawkins work would help position it better here.
Similarly rather than likening RepRap to Linux, a better analogy and comparison would, be based around the self replicating GNU toolkit, GCC and all. For extra points you can also draw vitamin and nutrient components ones which we have long forgotten as they have become ubiquitous (a PC and an OS). But the bigger point I am trying to make is RepRap doesn’t equal 3D printer any more than an Apple App equals GNU/GCC and yeas I am aware of the delicious irony of that final comparison.
regards
Al
I’m extremely disappointed and saddened by this article. I understand the positive intent of the author, but the benevolent dictator approach is dependent on a perfect dictator always acting in the best interest of the community and always knowing what’s best. No one fits that description, so that role will always be filled by someone who is in a position to do more harm than good. Let free design continue. The best ideas and methods will naturally rise to the surface.
For example, there may be a temptation to set a standard that machines must be able to print their own parts, as a purist belief without regard to how that affects print quality. This seems like a fine way to relegate the RepRap to the forgotten annals of history when something truly superior comes along that doesn’t follow that dictum. Just an example.
I’m trying to understand why no one in the RepRap community is working on powder based printers. They are patent encumbered, so no one could sell kits, but designs could be published and people could build their own printers if people worked on designs together. Why not move beyond PLA and ABS?
-there are no standarts for most printers:
1.: there is no market:
arduino, wades extruder, MK1 heatbed, t5/ t2.5 belt, nema17 motor, linear bearings, pololus, endstops, 3mm fillament are standart for most printers-
so most of the parts are around in those shops. they are industrie standartized parts and they are the market.
in generel the quality of vitamis is getting better/ more industrial (so results are)
-there is a wired development, things mutate and -evolve in all directions:
2.: things evolve via thingiverse:
downloading free parts to update your printer is essential to have the best possible printing result.
there is no machine that works best out of the box- mostly people add stuff to make them “a little better”- and up and download it there- some spareparts are needed to be printed, cause they wear out, like extruder gears.
cause thingiverse is the only way to share your ideas with the whole comunity, mostly they are found there.
needful would be a rating system to get, which parts are evolved the most/ most useful to your system.
-the whole thing is crappy, i dont knowhow to use it:
3.:firmware, software and hardware:
callibrated printers behave similar- (think of times before callibrated e-steps)
now were having a free and professional printing prepare tool, a printing tool with user interface and g-code preview. all this has evolved from non-userfriendly and buggy printingsoftwares.
-what we need is a system of king:
4.shops, comunity, and presidents:
havin 200 hotends on 350 shops, 16 mendel variants and varous special parts. shops sell sets of those selfinvented prototypers, try to run their business and dont care about other standarts.
but think of a company like samsung (cause they build nearly everything) breaking into the extruder market: builds an exellent wormgear piezo nozzle, selling it for some bucks?
also in the irc are 350 people constantly- half of them designed their very own printer- all of them havin a unique printer, or still work on getting it runnig.
ask those around there to vote for the next generation of repraps, to put on the frontpage of reprap.org.
in general the whole reprap theme is so amazing, that you can hardly stand it evolving it on your own- you think of it, write it down and print it on a weekend, assmble it, make photos and push it to thingiverse. using a rating tool in one of our favourite 3d-model download platforms would push this process to another direction-
in general things evolved very fine in the last year, in my point of view.
Two types of objections that some make higher up in the thread can, I think, be answered.
objection 1: no standardizer (“benevolent dictator”) is perfect. Therefore standardizing will do less good than harm.
That is a fallacy. The truth that no person or standard is perfect doesn’t itself entail that a standard is all things considered worse than no standard. It all depends on the specific flaws in the standard/standardizers and how well the nonstandard alternative works. There are many examples of standards that has some bads but with more and stronger goods and provide overall net benefits. Linux is an example of that.
objection 2: standards will block new, much better designs that think outside of the box.
First, standards are not laws that ban development outside of the standards. They only held concentrate collaboration and ease modulation. Still, concentration in one area might mean less work on other promising fronts. But, second, such risks can be mitigated by building in revision and radical change as goals in the standardizing cycles. Add an alpha track to the standards and host annual prize competitions to stimulate wild and bold revisions.
Speaking as a long-term, former member of the Reprap Core Team, during the first five years of the project while we were trying to get any sort of low cost 3D printer to work and had just a handful of people working on the dream, we tried to have loose standards, viz, extruder, electronics, software & etc.
With the release of the second generation Mendel printer, however, adoption exploded and the community grew enormously. Now we find that there are literally hundreds of ideas for niche printers and innovation being tried out all at the same time.
As well, these ideas are cross-fertilizing at an enormous rate which has made the continued success of the enterprise certain. All you have to do is monitor the commoditisation of printer parts and filament to know that Reprap has succeeded. Stepper driver boards which used to cost $400-500 can now be had for $50. Steppers which used to cost $75 can now be had for under $20.
When you use biological metaphors like Adrian Bowyer rightly does it is, perhaps, better to think of the reprap community as a bacteria community. Individuals are encountering new opportunities and challenges and making known how they addressed them in much the same manner that individual bacteria exchange DNA snippets that code for improved surviveability.
This is the biggest pile of bull..err.
Let me just make sure I understand what you are saying.
Are you saying that the reprap community is responsible for lowering stepper motor and driver prices?
I can’t see what you are taking such offense about. Generally, when an activity such as Reprap creates a strong demand for particular kinds of parts, one typically finds the prices for such parts dropping unless there is some sort of production bottleneck or no excess capacity in the supply companies. Neither is the case with the smaller NEMA steppers produced mostly around Shanghai or driver boards, which a variety of small US producers are creating for sale.
Now if 3D printers used the high torque, high amperage steppers and driver boards that are encountered in conventional CNC equipment I’d say that it would be difficult to assign credit, if that is the right word. There really isn’t that kind of overlap with Reprap 3D printers, however.
Would you care to say more about why you find yourself irritated?
Nice backstep forest
“Al Billings says:
March 17, 2012 at 3:43 pm
I’m trying to understand why no one in the RepRap community is working on powder based printers. ”
One quite successful one has been produced down in Brasil. Unfortunately, the patents for powder printers are still active.
Patents keep you from selling kits. They don’t keep you from designing functioning printers and putting the details on a wiki page or thingiverse.
@Al Actually, it depends on where you are. In the US according to the case law I’ve seen it does prevent you. In the UK, OTOH, it doesn’t.
Are you a lawyer, Forrest? 🙂
Are you?
This isn’t about me.
I take it the answer is “no” then.
Would you care to share what your point is, then? Are you suggesting that only a lawyer might know anything about patent law?
You’re spreading pointless FUD unless you can actually cite case law.
You are, of course, entitled to your opinion.
If you just want to design and print parts, and have support for your hardware, then buy a Mosaic, Replicator, UP, etc.
If you enjoy the process of hacking and improving and tweaking a device, rather than making parts right away, source a bunch of parts for whatever the hot design is at the moment.
The price difference is entirely due to the amount of intelligence and time you’re willing to apply to the project. You can save hundreds of dollars by going for one of the cheaper kits, but unless you feel pretty confident that you could design and build your own from scratch, I’d recommend a prebuilt printer.
No need to mess with the RepRap project. What you’re proposing is already happening. RepRap is the wild, grizzly frontier country. Commercial companies like MakerBot follow RepRap at a staid pace and productize the more useful innovations.
It’s sad a late comer to the scene like Adrian Bowyer gets all this 3D printer credit. There were so many pioneer he doesn’t even nod his hat to.
My college roommate’s “Santa-Claus” machine was printing 8bit color prints with finer detail than the replicator five years before anyone knew who Adrian Bowyer was.
Would you care to tell us more about your room mate’s “Santa Claus Machine”, for instance who made it and what it cost five years ago?
No clue five years ago. Now fifteen years back… There was no official name. “Santa Claus Machine” was pretty common though. Keep in mind this was before social media/blogs/etc.. back when everyone still had a pile of AOL 3.5″ disks.
Most used various types of hot melt glue or fast setting epoxy (slow but awesome). Some added ‘dye jets’. Some experimented with different types of glass like plastics. Mine utilized a very hard DOW hot melt.
You might not be familiar with ‘plotters’ but they were more common back then. It wasn’t too uncommon to buy an old plotter or just ‘bolt-on’ your experimental head and table to a new one. The cost was trivial since everything was hacked together. Hybrid stepper motors similar to the ones used today were commonplace and could be gotten cheap. Drivers were typically custom because 1) it’s cheap and 2)driving a stepper is easy if you’re willing to do some research.
Anything else you would like to know? I don’t know of any websites (haven’t looked). Those that existed were uploaded to sites that no longer exist.
If I recall Don Lancaster had an article..
haha yep.. here’s a ‘recently updated’ page of his. haha.
http://www.tinaja.com/santa01.shtml
Check out the links to books about powders. It was such a hot topic in the 80’s most of the books were published back in the early 90’s.
What does the community think of professional units?
Something like the http://www.stratasys.com and their “uPrint”. Basically a professional 3D printer with soluble support removal and insane detail. You can find a ‘reseller’ on their website that sells reconditioned units at prices close to that of a replicator.
Something like this doesn’t use any proprietary hardware and could be thought of as a cell phone, laptop, paper printer, keyboard, hammer, drill, etc. i.e. a tool that is purchase to be used.
LOL! Trolling for Stratasys! How pathetic. 😀
The stratasys units are indeed VERY proprietary, and the detail achieved by their low cost units is less than the current generation of repraps. I’ve NEVER seen a working one sell for anywhere near $1k, either
Odd. Every credible source (unlike this comment section) appears to state the opposite. Cost of material appears to be the biggest issue.
Although there are many articles I found easily which talk about equivalent plastics that work fine at 1/10th the price.
Never tried any. Just wondering what people think of the used professional units because I see them for sale frequently at between $5k and $2k. Two months back I saw a fully functioning Uprint SE without cleaning thing go at (industrial) auction for $300.
Here’s one I found with a quick search:
http://www.additive3d.com/3dpr_cht.htm
Does anyone know if using alternate plastics void the warranty on new units? I bet it does.
Andrew,
Please explain “low cost units” and “detail”. They only have one unit they call “Easy on your budget” and even that thing is ~$14k.
It has nozzle resolution and x/y accuracy at least one order of magnitude better than the rep rap with a thinner layer height.
I am not surprised you haven’t seen a used uPrint for under $1k. I would be surprised if you have seen one used at all.. considering how new it is. Where have you been looking? I have been lurking on industrial auction sites and I’m amazed how cheaply they go for.
I got a quote for a ‘refurbished’ one at $3k. That’s close enough to the $2k replicator for me considering the soluble supports and material properties that everyone is saying is probably the best of all 3D technologies.
I have had the pleasure of seeing parts made from the uPrint units compared to my reprap and those of some others at Maker Faire NY 2011. (Theirs were Prusa Mendels)The home brew 3d printers could produce surface finishes every bit as good, and better in many cases.
The layer height spec’d by stratasys is .01″ (.254mm) where the repraps are now getting very good results at layer heights at or under .004″ (.1mm)
The specs for XY accuracy are all over the place on both sides so it can’t really be compared in a meaningful way. I will say the heated chamber (PATENTED, PROPRIETARY) allows for some bigger parts to be made with less warping (so better ‘accuracy’ in that respect )
Maybe you can find one for 3K, maybe not. You are still going to have to pay $120lb+ for their material (which has microchips embedded so you HAVE to use their PROPRIETARY material.
Sure, the soluble support is a really big plus, but there is nothing preventing you from using that on the emerging multi-extruder 3d printers. Not to mention, there are hardly enough used uPrints to go around.
Lastly, PLA makes a pretty decent dissolvable support material in it’s own right, and it’s inexpensive.
The uPrints certainly have their use, but they are NOT as far ahead of ‘low cost’ ($3500 and under) 3d printers as you think. As soon as the heated chamber patents run out Stratasys is going to have a LOT of competition.
BTW, the link you included takes me to a site comparing 3d printers. Nowhere there does it say you can find a used uPrint for cheap.
Lol @ “insane detail”
You must have missed the post about Ultimaker print quality.
I think you are also forgetting the DLP printer family into all this like JR’s DLP Printer http://3dhomemade.blogspot.com ( you have already published twice and mine http://www.3dlprint.com.
Cheers!
“Hundreds of people working on a hundreds of designs brings is a linear development process, and doesn’t benefit the project as a whole.”
This would be true if ideas from one line could not be perused by the other lines. But of course such transfers do happen, and they are equivalent to horizontal gene transfer. As Matt Ridley argues, horizontal gene transfer or genetic exchange gives an individual potential access to all the diversity of the entire population.
So I disagree with the notion that diversity leads to a linear crawl. This would only be true in the absence of horizontal transfer.
I agree with Dirk.
It’s far too soon to begin standardizing anything. Whenever, I discuss the Reprap, people tend to think a few short years ahead and mope at how “primitive” it might seem right now. The Reprap is already 5-6 years old and has undergone tremendous development during that time.
Just imagine, what will happen in the next 25-30 years.
It’s better to work on clear guides for popular machines, and few things beats the LEGO-style assembly manual, I saw somewhere, for the Prusa Mendel. It was brilliantly done. *That* should be standard. 🙂
What a terrible idea. People need to be able to improve these designs without the fear that their machine will be seen as inferior if it doesn’t have the reprap badge on it.
The Arduino is not similar at all. For starters, it’s a microcontroller, PCB and a few connectors – much simpler than a 3d printer. Also, there are probably more Arduino variants out there than 3d printers. Not to mention that their standards didn’t do me much good trying to fit the thing onto a breadboard – the headers still don’t line up.