Ok, I am compelled to point out that the $22,000 printer works ~8 times as fast or does 8 times as much work...depending on how you look at it. Droplets versus mugs.
Maths
(0.2mm/0.1mm)^3
There is not much of a premium for performance in supply costs. The premium in hardware costs is ~4.5x. A lot for a consumer. Cheap if you're racing.
What I was reminded of is commodity servers. You run the models through a cluster of three or four cheap rigs (maybe with different slicing software) to increase the odds of getting a good artifact in 8 hours and to allow taking a machine offline for tuning.
Of course eight 0.2mm resolution printers will no more create a 0.1mms artifact than 9 fertile ladies will gestate an infant in one month.
I wasn't being critical, I was being compelled. Hanselman's analysis is reasonable and not misleading. The $600 product easily solves the 95% case.
And for a consumer oriented product probably the 99.9% case since most will sit on the shelf because it's still printing and how much stuff do people download off the internet and print with their color printers these days?
The idea that we will all be making our personalized LPS (Little Plastic Shit)[1] to our snowflake's content is true. But only because most snowflakes are content without doing it as a hobby. It's Buck Rodgers cool when it's on the screen, but the reality of waiting 8 hours for maybe getting a widget only appeals to some people.
It's like raytracing silver balls above a red and yellow checkerboard. Amazing to see it come out line by line over four hours at 320x200 in 1989. Not worth setting up a Blender scene to do it in the blink of an eye today.
For most people, there are better options than doing it themselves. For a few use cases doing it yourself is the best option. And doing it yourself is a great option for people who want to do it themselves. That's the same class of people who put their recipes on floppy disks in C64 days.
I don't know, do you have data to claim the printer price must scale at least linearly with cubic resolution? There are plenty of much cheaper printers out there with .1mm res. (https://www.3dhubs.com/best-3d-printer-guide)
You picked the most generous aspect for comparison but it still looks rough.
Just because it's the internet, doesn't mean I was being critical. I was just providing some mathematical context.
I used the racing analogy in the price comparison. It's cheap to make up a margin of minutes on the fastest car. It's really expensive to make up fractions of a second.
You could always buy 10 of the $600 printer sure there is more overhead but the effective bandwidth is increased. Unless you just need a single cup, however you could be printing multiple cup designs at the same time and if one machine goes down you're not completely shut down.
If you need a cup, you probably should get one made out of something food grade. In less than an hour I can do so at the thrift shop. And it's only 99 cents.
not completely true, as there are more variables at play. 0.1mm vs 0.2mm is the layer height, but you are missing the extrusion width settings. For all you know, the $22,000 was running twice the extrusion width. In addition, just cause a printer was run at 0.2mm doesn't mean that's the smallest layer height it can do, I would expect the printrbot metal to do at least 0.1mm, if not better.
Not the filament itself, the extrusion width. It's something that you specify in the slicer, and can vary by a bit, between hard bounds specified by the nozzle diameter of the print head. Also, consider how much of that $22,000 is "make it look pretty". They are already trying to do everything they can to make money off the thing, i.e. expensive filament and disposable build plates.
If a part effectively takes a calendar day to print, the build plate increases the odds of an acceptable prototype 0.1%, and net revenue on the completed project will be $5001 per day, then a $5 build plate pays for itself in a properly capitalized design development cycle.
And that's only the case if other features of the system don't add any value that could offset the cost of the plate and the additional day of design labor and overhead is ignored.
All of which is to propose the possibility that there might be business cases where the machine makes sense and the possibility that everyone besides we two obvious geniuses might not be incompetent.
Layer height and extrusion width are separate settings, if the extrusion width is the same, it's only twice as much head travel, not four times. Extrusion width is around .4mm on most FDM printers.
You can't really call it more work either, since the resulting volume is identical.
The only thing that's different between the two prints is Z resolution, and that's a factor of 2 difference.
I enjoyed reading this, as the author stated it is a rather contrived test, there are things the Stratasys printer can print that the Printrbot cannot (anything with complex overhangs and undercuts) but that really wasn't the point I think. There is also the entire 'tools' thing where the tool is often less important than the operator. I know people who can carve a bust out of ice with a chainsaw, but I'd be lucky to make a square block with the very same chainsaw and equivalent ice. The point being that tools are only as good as the training of the operator so tool comparison really does need to expert operators if it wants to be diagnostic.
But what it does say, and I believe it is incontrovertible, is that the "personal 3d printer" market is evolving at least as fast as the "professional" 3d printer market and the overlap of capabilities is growing. For any manufacturing technology its going to be speed, cost, finish, and capability and while personal 3d printers have nailed it on cost, and they are now competitive on speed, they are working on finish and capability. And that suggests to me its a time thing, not a 'new science' thing between when 3D printers start delivering on their potential.
I've got a Makerbot Replicator 1/Dual (the last of the 'opensource' Makerbots and currently replicated by the Flashforge Creator 3D for half the price I paid :-) which was much better than the original Makerbot 'cupcake' or the early reprap machines. There are print heads (a combination of an extruder and a hot end) now available which both improve the finish significantly and the speed. So its really a fun place to be.
And my personal favorite is that its a way to exploit robotics hacking in a much more practical way than running around a room avoiding walls :-).
It certainly raises some good questions, and the Printrbot is a great machine at the price point (I have a Printrbot Simple RepRap clone myself). His looks pretty well calibrated as I've seen a lot of Z-wobbling artefacts on the prints from several Printrbot (Simple Makers, not necessarily the metal one).
Worth nothing is that things like the Z-scar being on the side of the cup instead of under the handle is something that's decided by the slicer (the software that translates the 3d model into printer commands).
Different open source slicers that are regularly used produce different locations for these scars -- and I'm not sure if it's actually possible to position/hint this Z-scar manually in any of them. They often do try to be "smart" about it, but the end result may vary.
So I just took the coffee cup STL and loaded into the latest Cura myself, rotated in 90 degrees so it was standing upright. Without changing anything else, I ended up getting the layer changes on a few difference places over the print (I'm seeing this just by analyzing the G-Code using http://gcode.ws), depending on the layer height. It starts to the right of the handle and after a while seems to sort-of stabilize almost on the opposite side of the handle and on the right side of the actual handle, before the handle separates from the cup body.
All in all, it will likely have come out decent, but certainly not under the handle as you have. This I guess is because Cura will start the new layer close to where it finishes it's previous layer by default. In other words: it depends on infill percent/pattern and model location on the build platform.
I figure this is because you've added support. This causes the print head to get a "closest" start of the new layer exactly under the handle (because that's where the support material are). So going as far as calling it "lucky", no. The support placement actually helps Cura put the Z-scars in a good place on this model. On another model, it might be the exact opposite.
(PS: I use Cura for all my slicing and I like it a lot -- but I don't feel like the software actually gives me much control about the Z-scar placement)
> Printrbot (Simple Makers, not necessarily the metal one).
I've heard from the couple of owners that Metal ones are much better than Simple Makers. here is the quote
"but Printrbot Simple Metal is very reliable, I had one I’ve been carrying around in my car unpacked (I just dump it behind the seat, same with the tablet) and I just put it on the desk and it prints. Every time. Auto-bed leveling does wonders."
And another one from the same guy
"Just to be clear, I’m talking about Printrbot Simple Metal (not the previous wood version – that one is realty bad: it de-calibrates from day to day simply because of humidity "
We bought a Makerbot Replicator for $6,000 and it's been a huge disaster. I'm not sure if the people who bought it didn't read the Amazon reviews or what.
We were having the local college print things for us for $150/pop. For double what we paid we could have gotten a lightly used Stratasys which is the same printer the college uses.
Recent version? My local makerspace has a pair of Makerbot Replicator 2s and they've been great. But I've heard not-so-great things about their newer products.
>We were having the local college print things for us for $150/pop.
Care to expand on this? I've been playing with the idea of buying a couple very nice 3D printers and starting a 3D printing company. I think a lot of people have been burned by the consumer-level printers or simply need bigger or higher quality prints. What are you guys printing? Do you feel $150 is a fair price? What is the turn around time? What materials? Does that include delivery?
My gut feeling is a local company with local delivery and quick turn-around could be more tempting that going with one of the web-based providers who take days.
There's already a few established companies in the "cloud" 3d printer market, even 3D Systems (the 3d printer manufacturer) participates with their Quickparts brand (http://www.quickparts.com/LowVolumePrototypes.aspx). Just go get some quotes and see if you can make a business case doing this type of thing better in some way.
Being cheaper isn't going to win you a ton of business, imho. Being faster or better or more personal, or providing unique materials may win you quite a lot of more niche business, though. For instance, for low volume injection molding, Proto Labs are hard to beat as they will do a fully automated quote and give you flow analysis for parts you upload within an hour or two (usually this takes days). Your business would need to separate itself from the other existing 3d printer providers in order to have good success.
Starting a 3D printing company might be viable. Buying very nice printers makes it less so unless you have customers lined up who demand it and will sign long term agreements.
Buying a fleet of cheap printers [one or two at a time based on demand] makes more sense because the business model is renting and there's a lot more of a market for renting a 3d printer at $5/hour than at $30/hour. More printers means more peak capacity.
We actually just print a handle of sorts with a small enclosure for a circuit board. We were sending out prototypes to be printed, and as you see the cost was fairly significant, so engineering decided to buy their own printer and do it in house, which they haven't had any real luck doing so far.
You should try a MakerFarm kit. My tech unsavvy dad put it together in 10 hours of build and 15 minutes of calibration, and prints almost every day.(and that was before they switched to the easier to assemble v-rail stuff) $600~$700 gets you a good 3d printer
How many products are there in the world that cannot be refined?
Just because a new version or refinement exists, does not invalidate the usefulness of the original, nor does it mean the original users were guinea pigs.
Now when a product has zero utility in it's first version, that's different. But I would argue you should demand your money back at that point.
What's up with the abundance of cheap extrusion printers, while (almost) no consumer-grade SLS (selective laser sintering) machines are available? The latter technology seems vastly superior to me, in terms of freedom to actually print anything you want. Is the technology that much harder?
A lot of the technological hurdles associated with using lasers can be overcome by using a chemical bonding agent 'extruded' onto the particles by an inkjet print head. There's at least one open source printer that's trying this at the moment[1], and HP proposed a similar machine for the commercial sector[2].
The biggest issue, in my experience, ends up being that most non-metal printed parts are relatively brittle and prone to the material rubbing off if the outer layers are even lightly abraded. Ytec3D gets around this by soaking the prints with a liquid glue to solidify the bonds, but between that and the cleanup environment (compressed air, brushes, respirators) it's a hard sell for consumers.
EDIT: It's also important to note that FFF/FDM (extrusion-based) printers are getting much closer to 'freedom to print anything you want'. This year, I feel like multi-extrusion systems are going to become much more mainstream, allowing users to print more complex objects with support material that can be dissolved away later (e.g. ABS with HIPS support).
Harder, much more expensive and much more dangerous to operate. So that's why they're not more common. The last thing you want is for a machine like that to be run in a regular home without proper fume extraction. At the back of those machines you usually see a thing like a very thick vacuum hose connected to remove the (toxic!) vapors.
They also consume a lot of power, usually requiring industrial power hookups.
If only it was possible to hook up on single-phase house mains an arc torch-based plasma gasifier to direct those toxic fumes into. I specifically looked for access to three-phase power when choosing a residence because I don't see power consumption requirements for the metal sintering gear coming down anytime soon.
FYI you can use either a static phase converter or rotary converter to convert from single phase to three phase. A decent 20A three phase static converter will run you about $500. Much easier than trying to find a house with three phase!
Yes, good point, thanks for pointing that out. I did look into phase converters, but at the time was considering continuous usage of three phase power, due to the energy losses. Of course with an infrequently-used 3D metal sintering printer, it becomes much more practical to deploy a phase converter for just the printer's gasifier. Bonus: the well-built digital controlled rotary phase converters can last 30+ years, and from my poor electrical ed-shu-ma-ka-shun I understand these better ones yield a true analog sine wave good enough for even very sensitive electronics.
Thank you very much for pointing that out! Since such a printer wouldn't run all the time, this should be a great fit. Once biogas generators are integrated better with biogas cleaning and upgrading equipment to produce sufficiently concentrated methane to meet manufacturer specs for natural gas-powered equipment, perhaps with liquid nitrogen (highly useful for many sustainability hacking projects), then even the energy extraction externalities of such a gasifier could get quite respectable.
The fumes especially are a good point, hadn't thought of that. SLS is also possible with plastic, though, in which case I guess the fumes are comparable to those produced by an extrusion printer.
Well, almost. The SLS uses very fine grains as a base material, an extrusion printer uses a filament. The grains will vaporize much easier than filament but there is definitely a concentration of unhealthy particles in the air around an extrusion printer and I would not operate one without at least basic precautions (fume hood or something similar).
That's exactly it - last time I checked SLS seemed to be around 5 years behind that of FDM. SLS is also a more complicated process to get good results out of and certain patents for improvements (such as a Nitrogen atmosphere) are yet to expire.
3D printing reminds me so much of the early days of personal computing - it was an expensive hobby for techies, the equipment was hard to configure, and the everyday (practical) uses of the technology weren't immediately clear. But like personal computing, I have no doubt that 3D printing will become an essential part of our daily lives in the decades to come.
Just like today's printer cartridges, the 'print material' used by 3D printers is insanely expensive. That's gotta change for any essential daily lives application.
But today's printers share lots of daily lives applications despite the high cost of supplies. It all depends on the value of outcome and how easily one operates the device.
I'm sure 3d printer tech will end up with ready to use devices, printing metals as well as various kinds of other materials. I may prefer buying a panel and electronics, and print my own metal enclosure for my next monitor (or external disk or electric outlet or back cover of my phone or... you get the idea). Cost will be a factor, but after the utility.
Really? My impression has been that filament is pretty cheap e.g. < $30 per kg. For printing small objects, enclosures, small replacement parts etc, the filament cost has been pretty much negligible. That's why a lot of 3D print shops end up telling people just to buy their own printer if their doing any reasonable volume, because once the capital cost is out of the way, the consumables are fairly inexpensive.
$30 is ridiculous. ABS plastic is around $1,000 per ton (or cheaper if you buy 10 tons). Does it really cost so much to make a wire out of the granules?
That means there's approximately a 30x markup from raw ABS to a spool of 3D printing filament being sold over the counter in a shop to a consumer. I don't know how much filament costs if you buy industrial quantities rather than consumer ones.
To be honest I don't know enough about industrial pricing to understand if it's ridiculous or not, 30x from raw material to something being retailed at a consumer level doesn't seem unbelievable to me.
Markup on a black ink cartridge (the original comparison) is something in the region of 200x, so nearly an order of magnitude higher.
What this suggests to me, is that if you are making a lot of smaller 3d-prints, that don't leverage whatever special capability the UPrint has (size?), then it makes a lot more sense to have a large cluster of a couple dozen Printrbots than a single UPrint.
They reason they both have the "pile of spaghetti" look is that they're both FDM printers. The fact that even the most expensive FDM machines still give that crappy-looking result is because that's all the technology can handle. If he'd compared it with the results from even a consumer-grade SLA machine like the Form 1, let alone professional SLA, SLS or inkjet such as Objet then it would be completely different results.
I'm completly bewildered by the current 3D printing fandom. Every FDM print I've see just looks... off. I certainly am impressed people can make these things casually, but from an aesthetic and even practical point of view, it just looks terrible. I can't imagine using it for some of the more interesting use cases I hear about like replacing a cosmetic part in your car or house.
I spent a bit of time on 3d printing forums and from what I can tell its just figurine printing and hobbyist screwing around. These photos are very telling:
A low price SLA would be a serious game changer. The Kudo3D project will be interesting to watch. I'd much rather spend $2000 on something that prints beautifully than $600 on something that will print "piled spaghetti."
Interesting comment from the Kudo3D person on reddit with some details:
It's pretty amazing how the difference in object quality between the enthusiast and professional systems has evaporated.
I would be interested to hear how frequently the Printrbot gets a clogged extruder vs. the Stratasys, what print quality on ABS looks like between the two units, what happens when you have a more complex shape with a lot of overhangs, etc. Aside from warranty service, what are the advantages of the professional system?
Stratasys has a $3bn market cap, and this post basically says their flagship product isn't much better than something that costs about 3% of the price. Probably not good for Stratasys long term outlook - SSYS.
That printer is definitely not the flagship from Stratasys. It's actually their 'entry level' product and seems to be aimed at companies who want a printer comparable to consumer ones but with perhaps more durability and better support.
A better comparison would be with the Fortus line which is used in commercial applications for functional products.
given how shit the MakerBot product line got after the Stratasys acquisition, I'm not in the least surprised. (Cleaning a clogged print-head on a MakerBot Mini voids the warranty).
Don't know much about 3D printing but seems it's got a long way to go yet if it's taking 7 hours to make a mug with some ugly scaring which costs between $2 and $30 in materials.
If I want to print a mug I'd want it done in 5 minutes and cost $1, otherwise what's the point of owning one from the point of view that one of these will be in every household. Does anyone know if the current materials have any hard limits on them as to how fast they can be manipulated for example?
I believe the application is that you can make exactly one mug for only $2 and 7h. Then you can make exactly one tumbler. Then you can make exactly one plate. Then you can make exactly one coat hook. Then you can make exactly one collar clip that secures the armrest onto the driver's seat in your car. Then you can make a chess piece to replace the one the dog chewed up. And then you can make a bowling trophy.
Any specialized machine can absolutely annihilate the volume printer on the basis of cost, and for many things also on the basis of quality, but you have to produce thousands of units to pay for the machine.
You don't even need it to be a prototype. You just need to be making something that no one else needs--or not enough other people to warrant building a specialized machine to make that thing.
But then, if you do need that specialized machine, maybe there only needs to be one of those, and you have just the right sort of machine to make unique items....
Sure. But making mugs is wrong. You can't compare the output of a machine like this to an industrial process designed to make tens of thousands of something.
Nothing beats a cheap ceramic mug for mug-ness.
And that goes for pretty much any one-off produced with a 3D printer that attempts to re-create a mass produced item. It'll be more expensive, more fragile, less pretty and in general less functional and strong.
That's correct. Making mugs is wrong. Making mug might be correct. If you are making a second one of anything, you are probably better off using the volume-printed item as a physical model, to make a mold, which would then be used to create two whatevers by a different process.
For mugs, that means constructing a plaster mold, then using liquid clay (slip) to fill the mold. The clay is dried, glazed, and fired. If your mug design is not already unique, someone else already does this faster and more cheaply than you can. But if you want 20 mugs shaped like your own head, you're still only going to volume-print one copy.
Or maybe you're making mugs out of food-grade silicone elastomer instead of ceramic. You still only need one printed copy.
Nobody buys a compiler to calculate the values generated by the whetstone benchmark[1], everybody already knows the answers.
The point is not to make a mug, the mug is a benchmark. The point is that it takes a nontrivial amount of time, the model readily available, it is repeatable, and it has some very interesting properties WRT complex curves and overhanging structures.
(Like all benchmarks, we can argue from now to forever whether it is a good benchmark.)
You're right of course, it's hard to imagine it as a household item, although having said that, I know a lot of people who invested a lot of money into 2d printers and ink even though it cost way more than having better quality prints produced commercially. Maybe the appeal will outweigh the impracticality, if you had a uniquely shaped hand and regular mugs and utensils just didn't work, I imagine you'd be looking to get one in the next 12 months :)
Consumer (and office) 2D printing is all about having the document on paper within 30 seconds of hitting "print" though. If it took 7 hours, most people would happily use the print shop.
Yes, but if you can have something in 7 hours instead of 7 weeks, then the 3d printing thing starts to make a lot of sense.
Granted, there are many online services now which will 3d print parts for you with very high quality and about 1 week turn around, so owning your own 3d printer is not always a good financial move for a business, depending on the desired uses.
Not to mention that many printed docs are expected to have a relatively short life span. Handouts for a meeting, shopping lists, address labels, etc.
For the person printing things on the average "cheap" printer, they're not typically concerned with the longevity aspect. Most 3D-printed things seem to be the opposite for mainstream applications. People keep using examples of printing "stuff you can't get at Walmart" and implying the printed parts are expected to be used for many years in an environment that will have some moderate amount of wear and tear.
3D printing not only needs to get faster by an order of magnitude, the quality of the output component needs to increase substantially as well. You need to be able to print in colors and wind up with a finished product that looks close to the quality of an injected molded piece.
My time bandwidth is entirely too sunk into other pressing matters these days to be able to take on 3D printing, but Some Day Soon, I want to set it up in my workshop. These large artifacts that could be mass-produced cheaper are not my target use case, however.
In our current throwaway-oriented manufacturing and product distribution civilization, it irks me to landfill perfectly good objects for want of a very small plastic part. It isn't the gross materials cost compared to the mass-manufactured cost that I evaluate in these situations, it is the logistical tail externalities cost I impose upon my descendents for landfilling objects that could be used with a modest (to my income level) investment in repairing it. This observes the "Reduce, Reuse, Recycle" mantra by reducing the waste stream through repairing.
There is a rant embedded there, on waste streams are a sign of technological backwardness as a riff on Iain Banks' "money is a sign of poverty", but that's for another time.
A couple of examples to illustrate.
A door bell ringer that has a base mounted to the door frame and then the housing slips over the base and latches in with a simple catch in the base. The housing was ripped off the base (toddler Destruct-O-Tron), and upon inspection it was the simple catch that tore off. The catch is the size of a fingernail clipping. It's long gone.
Customer service at the manufacturer was sympathetic, but unable to help; they acknowledged they had the bases, but it just wasn't in their procedures to send it separately even if I was willing to pay for it. If I have a 3D scanner+printer, I could create another base with that catch, instead of ordering an entire ringer and landfilling a working ringer for want of a fingernail-clipping size piece of plastic. When I ordered the replacement ringer, I not only threw away the plastic, but also the PC board (though I salvaged the battery), and contributed my small part to converting nasty bunker fuel oil into emissions and engaging the exhaustively long supply chain it took to create the replacement ringer (including a few drops more oil to create the virgin plastics that went into the new ringer).
That's just me, but multiply by billions of people making similar decisions over the next few decades as a burgeoning global middle class arises from the externalities wreckage wrought by our current globalization implementation, and it gets ugly fast.
Another example. A sippy cup is designed with a small plastic rocker at the top that lets a straw flip up and down [1]. That rocker can be pried off, and eventually is lost somewhere in the house until a few years later when it is found in some nook. Without the rocker, the cup leaks when upside down. The manufacturer has stopped making the product, in favor of a new version with an incompatible rocker; this manufacturer turned out to be willing to ship just replacement rockers, but they only had the new style on hand. Off to the landfill the cup goes, and I order replacement old style cups from dwindling retailer inventories. Unless I have access to a 3D scanner+printer.
The state of technology is sufficient as-is today to let me more than satisfactorily solve these situations without resorting to a replacement purchase, and about 1-3 instances come up every week. Double that occurrence frequency if I got a metal laser sintering 3D printer. I could refine various products with small, custom silicone bumpers/insulators/grips/etc. by creating silicone molds with a 3D printer. The applications only increase as the 3D printing technology advances. These are only the simple, obvious applications; much greater changes and improvements in industrial product design, manufacturing, externalities reduction, sales, marketing, support and distribution lie ahead when we combine these immediate applications to other developments.
You are, of course, exactly right. Right now 3D printing is mostly in the "neato" stage, and is being oversold by a lot of actors. The materials are limited, costly, and prohibitive, and the process is extremely slow and of moderate accuracy. It is getting better, of course, but there are some hard, sober limits on the sorts of materials that can be extruded and modeled like this.
Cheap consumer 3d printing sucks. Keep in mind that if there is something cheap now like 3d printing, it's been around for 20+ years. And it has. Königsegg (The car company) 3d prints their turbos and doesn't do any machining to finish the impeller. Think of how accurate it has to be to spin to 100-150k rpm or more.
What I was reminded of is commodity servers. You run the models through a cluster of three or four cheap rigs (maybe with different slicing software) to increase the odds of getting a good artifact in 8 hours and to allow taking a machine offline for tuning.
Of course eight 0.2mm resolution printers will no more create a 0.1mms artifact than 9 fertile ladies will gestate an infant in one month.