3D Printing Has Evolved Two Filament Standards

We’re far beyond the heyday of the RepRap project, and the Hackaday tip line isn’t seeing multiple Kickstarters for 3D printers every week. In a way, this is a bit of a loss. The rapid evolution of the low-cost 3D printer seen in the first half of this decade will never be matched, and from now on we’ll only see incremental improvements instead of the revolutionary steps taken by the first Prusa, the first Printrbot, and even the Makerbot Replicator.

This doesn’t mean everything is standardized. There’s still enough room for arguing over deltas versus Cartesians, beds moving on the Y axis versus moving along the Z, and a host of other details that make the current crop of printers so diverse. One of these small arguments is especially interesting: the diameter of the filament. Today, you can get any type of plastic you want, in any color, in two sizes: 1.75 and 3mm. If you think about it, it’s bizarre. Why on Earth would filament manufacturers, hot end fabricators, and even printer manufacturers decide to support two different varieties of the same consumable? The answer is a mix of a historical choice, engineering tradeoffs, and an absolutely arbitrary consequence of what 3D printers actually do.

In The Beginning

Although filament-based 3D printers have been around for 20 years, these machines were outrageously expensive and usually tucked away in the engineering departments of large companies and universities. It was not until [Adrian Bowyer] started the RepRap project that the age of home made hot ends dawned.

These hot ends were, by any modern measure, terrible. The Thermoplastic Extruder is built from nichrome wire, JB Weld, Blu-tack, PTFE rod, copper plumbing pipe, and an M6 brass stud. Compared to the E3D V6 – an all-metal hotend with a heater cartridge, sensible thermal design, and replaceable nozzles and it’s surprising the first hot ends worked at all.

Five years of development: [Adrian Bowyer]'s Thermoplastic Extruder 2.0 (left) and the e3D V6 (right)
Five years of development: [Adrian Bowyer]’s Thermoplastic Extruder 2.0 (left) and the e3D V6 (right) Source: 1, 2.
 Although nearly everything from the original hot end experiments have been reengineered and reworked, one thing remains: the 3mm filament standard. The first thermoplastic extruders were designed to accept 3mm filament, and one of the first RepRap wiki pages for filament suppliers bears this out. Everything was 3mm, and Makerbot has doubled the price of their filament in five years. No surprises here. 3mm filament was the standard until 2011, when 1.75mm appeared on the scene and started to take over.

Why the switch?

Wade's Geared Extruder [Image Source]
Wade’s Geared Extruder [Image Source]
The earliest home 3D printers used geared extruders nearly exclusively to push filament through a hot nozzle. These geared extruders, like Wade’s Geared Extruder, are essentially a reduction drive. By using 1.75mm filament, the torque required from a stepper motor is three times less than with 3mm filament. This reduction in torque means a smaller direct drive system can be used, and since the drive system is smaller, the inertia of an entire print axis is reduced. This means smaller, faster printers than can also print better at low layer heights.

There’s another advantage of using 1.75mm filament – print speed. Heating less mass will always take less time, and if your goal is to print as fast as possible, you have two options: increase how much power you dump into a hot end, or decrease the size of the filament.

That’s not to say there aren’t advantages to 3mm filament. If you’re printing with large nozzles or high feed rates, you want a larger filament. When you consider something as insane as the PartDaddy, even the biggest filament is out of the question; by that time, you move over to pellet extruders. If you’re using a Bowden setup, 3mm filament will be a little less resistant to bending, though.

The difference between 1.75 and 3mm filament is only a choice in engineering tradeoffs – neither one is better, but each offers a few advantages.

The Current Situation

Despite a few holdouts like Lulzbot, the entire 3D printing industry seems to have settled on 1.75mm filament. New Printrbots are exclusively 1.75mm, and Makerbots use 1.75mm filament. The Dremel Idea Builder 3D printer – available at Lowes and Home Depot – uses 1.75mm filament. If you ever find yourself in a situation where you need to buy filament within the next hour, you had better hope your machine takes 1.75mm filament.

With a well-designed printer, there isn’t much difference between 1.75 and 3mm filament. A properly tuned printer can produce parts of a similar quality with either size filament. If this is the case, then surely the market would standardize on one filament or another, right?

This seemingly obvious consequence isn’t so. In talking with a few filament manufacturers and resellers, the sales volume between 1.75 and 3mm filament is about equal. For one reason or another, an entire community has settled on two different standards.

While this completely arbitrary evolution of two different sizes of printer filament exists, it’s not one that actually matters. The engineering choices that go into designing a printer to work on either filament are simple enough, and most people will keep enough filament on hand so they won’t have to rush out to the store twelve hours into a sixteen hour print. It is interesting, though, to see how we got to this point of two completely different standards for something that should just be one.

56 thoughts on “3D Printing Has Evolved Two Filament Standards

      1. Actually, “3mm” has a spec of 2.85 +- 0.05 (usually)
        At least, we supply it like that, and most other major 3mm suppliers do as well.

        The main reason we still use 3mm at Ultimaker is legacy. We started with 3mm just as anyone else. And before we knew it the company was going so well that switching to 1.75mm when it started to gain traction was simply not viable.
        And at the beginning of 1.75mm filament, the 1.75mm was much more expensive then the 3mm stuff.

        (I think we are the main supplier of 3mm printers right now. Sorry. Wish we all where using the same stuff, would be easier)

  1. I don’t see a problem with the two different sizes, as you mentioned they each have their advantages, I see (possibly) room for more sizes above and below those two, for the very same reasons each has its advantage over the other.
    As an analogy, you can buy motor oil in various viscosities and various containers (i.e. litre, gallon, barrel, tanker)
    One size does not fit all.

  2. The “3mm standard” has fragmented into the 2.85mm camp (primarily Ultimaker and Airwolf) and the 3.00 mm camp (the direct-drive guys, generally).

    3 mm presents some challenges with smaller-volume spools: as the spooling diameter gets smaller, the thicker filament has a greater curvature on the spool relative to the filament diameter — that can cause problems with payout filament and the feeding of the filament into the printer.

  3. I think many of the “3mm” filaments are actually 2.85mm diameter. It’s just shorthand. One possible advantage, though, of 3mm filament is that for flexible filaments it has fewer feed issues since it is stiffer. This would be especially true in Bowden systems (from what I’ve heard… I do not have direct experience).

  4. When I designed and built my 3D printer, I went with 3mm to be able to share filament with my girlfriend’s printer. To make my hotend, I bought parts designed for 1.75mm and replaced the M6 stainless steel insulator with a M6 set screw drilled out to 1/8″ (3.175mm), and that’s the only change that was necessary. My printer uses a bowden tube, so I suppose using 3mm filament is better anyways.
    I’ve noticed some local stores discontinuing their stock of 3mm filament, which is inconvenient for owners of 3mm printers, though I suppose a conversion wouldn’t be too difficult.

  5. At first, I thought that the two sizes might have been a result of metric conversion, but neither 1.75, 2.85, 3 nor 3.1 mm come within mils of some common fraction of an inch. But then, converting back the other way, 1/8″ works out to a 3.175 mm, which only leads to the conclusion that involves both ancient aliens and a gov’t conspiracy.

    1. The ~3mm standard exists because that’s the diameter of filament in the spools of industrial plastic welding rod that were appropriated by the first RepRap experimenters. It wasn’t specifically selected for 3D printing.

      The 1.75mm standard is the same size as Stratasys commercial FDM machines have been using for decades. Presumably they determined somehow that this was the right size to use in their machines.

    2. Also, only the U.S. has the problem of converting everything to an antiquated “customary unit” used as standard by only ~6 countries in the world. (Half of them under trade sanctions by the U.S.)

      The rest of us try to stay in decimilised measurements based on fundamental and observable constants of the universe, not our distal limbs, tyvm.

          1. How about every time New Orleans springs a leak? Some crazy French guy thought a mud flat a couple of inches above high tide level was a great place to found a city.

          2. I refuse to argue with people in the UK about the metric system, and how we don’t use it in the US.

            Instead, I’d rather invite them down to the pub for a 568 ml and talk about how many stones we weight and how many miles per UK gallon our cars get.

        1. The first definition of a meter is a quarter of a millionth Of the Earth meridian. It took years to measure and nobody can say that is arbitrary. Anybody can see it and reproduce it. It’s actually a great idea if you think of it.

          As it turns out it’s easier to observe cesium and distance travelled by light in a vacuum than to go to Paris and make an exact copy of a platinum rod.

          As for the US system, what is your excuse? It’s so outdated and illogical, I don’t know where to start. 12 inches in a foot? 3 feet in a yard? How many to make a fathom? Ounces? Pounds? Liquid ounces? Gallons? Wait, US or imperial? I mean, to what extreme are you willing to go to defend your unhealthy conservatism? Even the Brits quit that nonsense. Are you still a colony, frozen in the 18th century?

          1. Our ‘excuse’ is that the educational system can’t/won’t even teach human reproduction or evolution for fear that THE SOUTH WILL RISE AGAIN. (They’re close to adding things like gravity to the list.) You think anyone wants to even approach the subject of completely changing our systems of measurement, across the board? They’d be able to see the goddamn smoke from orbit!

            So, until common sense returns to these lands (Hahahaha, yeah right.) sensible people will be forced to use sensible units in secret, while those around them continue to measure things in barleycorns or whatever.

          2. Well, halves, thirds, sixths, and so on are easy to produce on a line or on a circle using simple compasses, and the original yard can be recreated in a clear night or two with a pendulum and observation of the stars to probably better than 1% accuracy. For example. It’s not completely brain-dead.

        2. You’re right. There is absolutely nothing arbitrary about either of those numbers. They’re based on universally observable, experimentally repeatable properties of materials that are fundamental to physics. That’s the whole benefit – the repeatability of our units of measure are determined solely by the accuracy of our observations of consistent physical phenomena. As the accuracy of our observations increases, we only become more and more certain of the repeatability of those units.

          Unless someone comes along and changes pretty much the entirely of our system of physics, the units are stable.

          1. It WOULD be possible to make things more reasonable by changing the base units further. We could make a decimalised time standard such that caesium oscillates one time per “Cecond” and the equivalent of the second was about one GigaCecond.

            The meter is defined by the emission wavelength of a common atom, so it is not the distance travelled that is used to define it (and the speed of light is now a fixed constant in the metric system, so even though it is part of the function, it is irrelevant to the definition)

            Honestly, let’s compare how they are derived;
            Customary Units; 1 yard = 0.9144 meters (based on the metre since 1893)
            SI Units (“metric”); 1 meter = 1,579,800.762042 wavelengths of helium-neon laser light in a vacuum.

            Customary Units; 1 (avoirdupois) pound = 453.59237 grams
            SI Units (“metric”); 1 gram = 1/1000th of a prototype kilogram, proposed to be replaced by the U.S.-developed Watt Balance which would define the kg in terms of Planck constant, which would become “fixed)

            Customary Units; 1 emperors second = 1 SI second
            SI Units (“metric”); 1 second = 9192631770 periods of caesium-133 radiation

            Customary Units; 1 Fahrenheit = 1/180 the interval between water freezing and boiling at standard atmospheric pressure (which is 101325 kg/m.s^2 ie Pascals)
            SI Units (“metric”); 1 kelvin = 1/273.16 the triple point of water

            The list goes on, but basically US customary units are just an inefficient abstraction layer of SI units, and as per multiple federal laws, Metric units are the preferred standard in the U.S. (Omnibus Trade and Competitiveness Act of 1988, amongst others)

            The only way a distant civilisation could recreate US units would be to
            1) create SI units based on (except for mass) universal and observable constants
            2) Apply arbitrarily defined multiples upon them, dividing by non-base-10 numbers, and thus reducing their mathematical utility

            U.S. Customary units ARE defined by arbitrary definitions, and the point is they are REALISABLE THROUGHOUT THE UNIVERSE*

            Except for the kilogram. For now…

          2. @AussieLauren: a gram used to be “the absolute weight of a volume of pure water equal to the cube of the hundredth part of a metre, and at the temperature of melting ice” (wikipedia). So maybe it would not be that difficult to recreate: there’s water on Mars after all!

          3. keep in mind, if we’re using light, a foot, is almost(but not quite) the distance light travels in 1 nanosecond. annoyingly close. 11.8 inches. 29.98 centimeters. just to adda little more confusion. also, where do seconds and such come in? why are we basing our units of measurement off of base 12 still?

  6. The article contains an error – thinner filament prints faster due to its surface area/volume ratio, not because of “less mass”. Extrusion speed is measured in mm^3/min.

    The 3.00mm/2.85mm thing is just a name difference. Most of the parts used in 3mm extruder and bowden systems (tubing, fittings) are 3mm nominal so if you buy 3.0 +-0.1 filament there is a chance it will jam. To avoid problems, 3.0mm filament is produced undersized to allow a margin of safety.

    Of course, sometimes you get guys with no 3d printing experience asking extrusion shops to run actual 3.00mm filament and then you start to run into problems.

    1.75mm filament does not have this problem since it uses metric 2.00mm fittings and tube.

    1. I wanted to say the same. Damn! It’s really the faster melting of thinner filaments, the reduced distance until the core is reached from the outside. If one prints at the same speed with either material, the ammount of extruded plastic has to be the same or the walls are going to be thinner.

    2. I would ditto the reduced depth to the core as being helpful. Easier to extrude helps too. It lightens up the carriage too, if you’re talking about putting the motor on the carriage rather than a bowden setup, requiring a lighter motor.

      It seems the machines that use 3mm exactly rather than nominally (~2.85mm) are relatively rare.

      From talking to people that sell materials & parts, the impression I’ve been getting is that sales of the hot ends and filament bias themselves towards 1.75mm lately. I don’t remember some of the specific sellers, and I probably shouldn’t name them anyway. I think figures ranged from 2:1 to 10:1, in favor of 1.75mm.

  7. 1.75mm is universally a better choice, if your printer can handle it. Less force required to extrude it, less unwanted force applied to the extrusion head by the stiffness of the filament, and fewer problems using up the end of a roll because it’s coiled so tightly.

    Oh, and also, the article has a conceptual mistake in it. If your printing requires a large volume of material to be deposited rapidly, you usually want to use the one with the smaller diameter. The bottleneck is never getting enough material into the head quickly…it’s getting the material to melt quickly enough as it passes through the head. Your extruder drive can push filament into the head way faster than it can melt and get out of the way. Thinner filament, with less volume relative to the surface area of the extrusion chamber, has an advantage in melt rate.

    1. * I will say that historically, 3mm may have been a better choice — first because it was cheaper (easier to get within tolerance), and second because a given absolute variation in diameter would mean less of a percentage change in volume, so crappy 3mm filament would be more reliable than crappy 1.75.

      Now that there’s an extensive supply chain set up, both types of filament cost the same and the tolerances are where they should be, so the only reason I can think of to use 3mm is if you have a bowden setup.

      1. 3mm doesn’t have to be as exacting, but as long as you have a close fitting guide tube, 1.75mm flex filament prints fine. You can’t have very much gap between the hobb and the guide going to the hot end. I used a PTFE sleeve and shaped it to fit close to the drive pulley. With that little improvement, I could print fine and even do lots of retractions with impunity on 1.75mm flex.

  8. I think Adrian originally came up with the 3 mm standard, you’d have to ask him why, but when we were printing at 8 mm/s, it seemed to work pretty well.

    I was trying to get my extruder to print a little faster, and spent some time measuring the forces various extruders could exert on the filament. My final version got up to 15 kg on 3 mm filament; details are on an old version of the wiki page here:

    That’s around 3000 psi in the extrude head; that’s a lot of pressure. Dropping the filament size to 1.75 mm gives you the same nozzle pressure for 1/3 the extruder force, so you can use a much smaller and lighter stepper/gear combo. Plus you get 3 times the extrusion resolution, which was a bit of a problem when we were using direct drive extruders.

    Good to see that old extruder on Hackaday, thanks for the shout out Brian!


  9. I also thought the 3mm size was “chosen” due to availability; you could buy 3mm plastic welding rod. (If you had a service contract with stratasys you could buy their filament, too; at only 1/10th the cost of gold)
    On the other hand, I recall 1.75mm being driven by the “famestream ” players like makerbot, who organised their own Filament extrusions, in an attempt to have a lower part count.

    Does this match anyone else’s memories?

  10. I kind of wish you’d have separated this article into two separate ones. You start talking about extruders and then suddenly dive into and conclude with filament size. It would be awesome to see a 3D printing series that discusses how each particular part of a 3D printer has evolved over the years, the advantages it brought, and the downfalls. Right now, a good chunk of that is found in the (scatterbrained) RepRap wiki and random forums. It’s hard to tell what is what, what works, and what is the new cutting edge.

    I have been immensely interested in either upgrading (or even just replacing) my old school Makergear Prusa. It WAS the top of the line kit back when I got it, but so many things have changed in the last few years that it is near worthless in comparison. Back then, there were only a handful of options (mainly Makerbot’s odd conveyor printer, UP’s first printer, Makergear’s printers, and a few odd Ebay kits here and there). Now there are ENDLESS printers and no way to tell which way is up unless you have had your nose buried in forums for the last five years.

      1. The 5W means that it doesn’t burst into flames when cooled with ambient air at standard conditions when used as a resistor, not a heater.
        If you attach a heatsink (the extruder) and don’t mind the 200+°C, you can push more power, a lot more…

  11. Don’t draw your own conclusions without evidence. It’s entiely possible that sales of 3mm are for legacy machines only. Just because sales are on a par with 1.75mm it doesn’t mean new machines use 3mm.

    1. We don’t have 563 different filaments, because setting up a quality extrusion line is expensive, which is why all producers supply to various resellers. (I think there are about 15 actual filament producers world wide, the rest is just reselling)

  12. Dont worry, tomorrow HP or Dell will introduce a faster printer with their future and exclusive 0,2 fiilament (it is an example). Now everybody are simply running to take their parcel of terrain.

    1. We played with 1mm filament at Ultimaker (small experiment). Problem is that it did not offer much over 1.75mm, but was harder to produce and harder to make a proper feeder for.

      You can spool it on a tiny radius, so that makes for really cute tiny filament rolls.

  13. It sounds like a big deal, but if you have a RepRap, it’s not.

    From Aliexpress, I buy the E3D clone hotends for about $8 each. I have both 1.75mm and 3mm equipment available, and it only cost me about $14/size (extra nozzles of all size extrusion).

    What this means is that I can stalk eBay for people who are switching from 3mm and get really cheap and good filament deals. Just snagged 2kg 3mm black ABS for $20, including shipping. And it prints great.

    1. We at Ultimaker will be shipping it for sure the upcomming years. As we are still selling 3mm printers right now, and there is no plan to discontinue to sell the UM2. (And end-of-life has never even been talked about)

      So you can always get the stuff from us. I know it’s not the cheapest option. But we are big enough to guarantee availability.
      And I think Colorfabb will also keep their 3mm production running, seeing how much Ultimakers they use for testing.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.