Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: Why are there no open source 2d printers?
288 points by pangoraw on Oct 15, 2020 | hide | past | favorite | 240 comments
In the 3D printing world, there are plenty of open source choices, allowing manufacturers to drive down costs. On the other hand, it seems 2d printing is stuck with legacy companies with completely closed drivers and hardware (you have to buy cartridges from the original manufacturer).

Apart from the nozzle why is it hard to manufacture and/or design?

I worked for a while in the R&D department of HP printer division. As @jacquesm said, good 2D printer costs peanuts. The amount of R&D in color quality, speed and other parameters is huge. There were a lot of teams involved: mechanical, electrical, software, chemical... And because of that investment, there are thousands of patents that the big players are continuously paying each other for. It's a very old market with a lot of legacy. For most of us, a printer is something for home photos, some documents, and so, but that's only a little part of the cake: the money is in professional printing, ads, designers, etc.

Once that is said, it should be possible to work in a general-purpose open source 2d printer. The open community has achieved bigger goals. The biggest problem I can see is the entry barrier: to get a very basic printer, you have to invest thousands of time with a lot of knowledge in different areas, when a basic printer, even from the large companies, is not very expensive.

I think that one of the only chances we have for that to happen is that a company frees its designs and patents and community starts working from there.

What kind of patents? I had tank of an HP laserjet 3 in the 90s, and patent life is 20 years. For the basic functionality, they should all be expired at this point, and the limiting factor at the time was the high cost of memory and compute.

Shouldn't anything relevant have expired years ago? The first laserjet came out in 1984 it seems. Prices have come down, but I haven't seen any real innovation in printers (not that I really need any- I just want them to print)- since 2000.

Do you remember printing in 2000? Literally every printer was suffering from constant paper jams and other mechanical malfunctions. In 2020 a top consumer or business printer will not jam on you.

The business printers in 2000 had slow processors and more ram. It was significantly bad that printing PDFs spent more time processing the file than putting toner on page.

Finally, the interfacing for printers today is fantastic. I know this isn’t about toner on page, but having wifi connection, an LCD touchscreen interface, and them generally being a little smaller has made the experience better.

The only thing that was better about printing in 2000 is that back then printing was more useful because so many people wanted paper copies.

> Do you remember printing in 2000?

The impression I get from HN comments is users of monochrome business laser printers from 2000 are the only people who are happy with their printers :)

The fallacy of "they don't build them like they used to" applies equally to bridges from 1850, cars from 1970, and printers from 2000: the poorly built bridges fell down, the cars with inferior parts rusted, and the printers with chronic issues got Office Spaced. So of course only the well-made examples survived.

Actually the printers they are upset about in Office Space (with the ‘PC LOAD LETTER’ message, not the prop they smash) are the practically indestructible early HP LaserJets.

Seconded, however I don't expect our best aqueducts to last 2000+ years


There also a saying that "Anybody can build a bridge that stands up, but it takes an engineer to build a bridge that barely stands up". The 2,000 year old aqueducts were massively overbuilt, not because they wanted them to stay up forever but because they didn't know how to make them stand up for 20 years without building in a factor of 100 safety margin.

Our aqueducts won't last 2,000 years, but we built them for a fraction of the cost. They'll fall down in 100 years, but we'll rebuild them with something even stronger and even cheaper.

Tell that to the cities who see their infrastructure fall apart after 20 years and can't finance rebuilding it stronger and cheaper.

How little infrastructure works they have if they had to pay for it to be 1000x overbuilt?

there probably is a middle ground. Just because i disagree with the rather radical opinion that claims it's most efficient to just barely fulfill minimum requirement and that throw-away & buy-new should be considered for civil engineering doesn't automatically mean i favor the radical extreme opposite opinion.

I am just stating the very obvious problem.

I think ultimately the problem is that infra is an investment in future growth, but the decision makers aren't accountable on failure.

Got any evidence/examples for that, where the cost of building for 100 vs 1000 years has been worked out?

I don't; I'm just spitballing the numbers. What really happens is that they'd rather tear down a building after 100 years and build a completely different thing. Even skyscrapers generally get torn down in less than that.

There is some infrastructure (bridges, sewer systems, dams) that are supposed to last longer than that. Bridges are often torn down and replaced, and it's probably more expensive than spending twice as much and having it last five times as long (again, spitballed numbers), but that's what fits in budgets. They don't want to discover that traffic patterns have changed and they need a different bridge, or no bridge at all and have to take it down.

That's becoming a real problem for sewer systems, which in a lot of places are reaching expected lifespan, and it's going to be ludicrously expensive to replace.

Incidentally, there are also reports that the Brutalist buildings are so overbuilt that they're hard to get rid of, even when they're bad (such as having insufficient ventilation). Gigantic piles of concrete will be there in 2,000 years, whether we want them or not.

One of the pyramids at Giza was built over the course of 20 years by tens of thousands of slaves.

Taking the (shamefully low) minimum wage in the USA of $7.25/hr and assuming your employees work 40-hour weeks with 2 weeks vacation, that's $14,500 (again, shamefully low but let's roll with it).

Ten thousand of those workers costs you $145 million per year, and for twenty years that's some $2.9 Billion.

It's pretty obvious we could build a pyramid, which is basically just a hill of rubble with worked sides, for a lot less today if we only needed it to stay up for a few years.

For reference, the Bass Pro Shop pyramid in Memphis cost around $100m. The largest pyramid in the UK cost the government of Qatar over half a billion dollars

If you have a basically bulletproof monochrome HP 4000 series or similar that has cheap cartridges that last 5-10k pages, easy/cheap maintenance (a whole new set of rubber rollers costs $30-50, and is basically the only part that wears out), and a low printing volume, I can totally see having a printer for 20 years.

Same here. Happy HP 4000 owner. 1999 technology and it never fails. Maybe get a paper jam once a year. The only trouble I have is the plastic is getting kind of brittle and little odds and ends snap off but the printer still functions exactly the way it did 20 years ago.

A sneaky issue with these printers though is that they use an excessive amount of power even when doing nothing.

I've got a current model monochrome laser printer and I'm very satisfied with it.

Same here. Mine is from Brother, probably their cheapest laser printer. No hi-res colour display or any fancy stuff like that. But it just works.

Exactly. Works amazing, no 3rd rate touchscreen experience. No badly made apps I have to install on all my devices. Just does what I want it to and nothing else.

My previous HP inkjet thing that this replaced was a complete nightmare and is in some kind of purgatory state now where it boots properly maybe 5% of the time.

In the 90s I used an Epson MX-80 dot matrix printer. I think it was created in the early 80s. The thing used paper with holes on the perforated sides, so the sides could be easily removed but also moving the paper along is much easier like this. Actually it worked quite reliably as far as I can remember and some people actually appreciated the unique style of the printings in letters. Eventually I got a more modern printer though but not because it broke. It did support printing of arbitrary files but also supported printing console characters. Therefore it basically needed no drivers for default monospace font printing.

At least for the fun factor it might make sense to start with dot matrix printing. (I think it's still far superior to thermo printers ;-))

Really no comparison to modern printers in terms of reliability. I bought my last printer with 5 years guaranty upgrade. I'd ditch the printer completely if it wasn't still required for some official paperwork :/

I use a monochrome laser printer from 2012, and I'm very happy with it, so you're not far off. :)

I'm running on the same business color laserjet from Dell (relabeled Brother MFC I think) from ~2013. It's never failed me. I feel like every consumer printer I've ever owned was a giant p.o.s. . I regret not going business level on printers much much earlier.

I have a monochrome home laser printer from about 5 years ago that I'm happy with.

>Do you remember printing in 2000? Literally every printer was suffering from constant paper jams and other mechanical malfunctions. In 2020 a top consumer or business printer will not jam on you. The business printers in 2000 had slow processors and more ram. It was significantly bad that printing PDFs spent more time processing the file than putting toner on page.

All of that, including the interfacing are not related to patents.

It's not like there's some patent issue for using USB or wifi in your printer over whatever interface used in 2000 compared.

Or like there's a patent on using faster processor or more RAM in your printer (two things you've mentioned also).

Are you stating your last two sentences as the result of researching these issues?

Because as a layperson, I would not have thought a company could get a patent on clicking a button to purchase an item online.

>Are you stating your last two sentences as the result of researching these issues?

Do you really believe "researching" is needed?

Like, in every other domain and product category, a company can slap a faster processor and more RAM to the next iteration - as they do -, but this is somehow prevented in printers by patents?

I'd say the extraordinary claim that needs extraordinary evidence is not mine...

When you read HN comments it helps to preface them with “I feel like...”

I upvoted you out of greyness because I feel like that is generally correct here. :)

Now that I've helped you out of the quicksand, let's do battle:

"[I feel like] it's not like there's some patent issue for using USB or wifi in your printer over whatever interface used in 2000 compared."

I could certainly try harder to finesse it, but I don't see any sensible place to insert a "YMMV" into any part of that sentence.

Emphasis on "feeling", because there's seldom any thinking going on.

Or at least, you feel like there isn't

>> as suffering from constant paper jams and other mechanical malfunctions.

> All of that, including the interfacing are not related to patents.

This seems like a pretty strong statement, can you elaborate as to why solving paper jams wouldn't involve patented IP ?

The paper jam might be. The rest (speed of processor, amount of ram, interface used) certainly aren't.

I feel like you may not be familiar with how patents work in the US: https://patents.google.com/patent/US20040137855A1/en. This is a patent for "Wireless Mobile Printing", which is more specific than just "Wireless Printing". Both of these concepts are, arguably, obvious combinations of existing technology.

I'm finding myself printing out code more and more these days, highlighting it and marking it up. There's just a different part of my brain that kicks in when I am able to get tactile with something.

I have gotten into the hobby/habit of printing at least one "thing" a day with artistic intent. That could be:

* an anime character printed with a thermal receipt printer * a 4x6 card with an information graphic I rendered with CSS Grid * a shutterfly envelope to family in New England, etc.

most of the time I am starting with an image somebody else made, but there is a lot of judgement involved with fitting the image to paper and process -- it is a bridge between the world of digital images that I work and play in and the real world.

I go through printers the way rock stars go through guitars and what to do with the e-Waste is already part of the product.

I am amazed the the HP Officejet 6600 which just failed on me -- despite the expensive ink, the quality of the work it could do is astonishing.

It stopped picking up paper because something (like a little plastic gear) broke in the drive chain for the pick roller. To be fair all the rollers looked pretty worn -- the printer had been heavily used by a college professor. It's possible we could have fixed it but considering the cost of the next ink refill, I chose to get another printer.

If I were going to salvage the old printer I think I would go for the stepper motors, which would be great for robotics and other mechatronic projects.


I badly want to hack an inkjet printer to print white ink onto transparencies and then put it into a second printer to take a PNG with alpha like


and make a sticker that could go in a window. Commercial kits to do this cost about as much as a good DSLR lens and they are a business expense to people who are making large quantities of swag.

It's a good market case for the "open source 2d printer" however. It's one thing to get white ink compatible with the printer, it's another thing to get the transfer function between "75% transparent" and a certain amount of ink into the printer's brain.

> it's another thing to get the transfer function between "75% transparent" and a certain amount of ink into the printer's brain.

Assuming the printer has sufficient color resolution/depth, you could do gamma correction as a preprocessing step on the host computer. Still obnoxious, though.

While I am not printing much code lately, I do print a lot of datasheet sections that usually force me to go back and forth through a PDF, like complicated clock trees and memory maps (embedded development).

that brings back memories of the school lab in the early 90s. didn't get enough time in the chair on a shared terminal to debug completely, so everyone printed to greenbar and would debug from there. i even got to the point of writing code on college ruled paper by hand, and then transcribing it once i got to the lab. taking typing early in high school definitely paid off. people made fun of me for being the only male student in the class, but i knew it would be a useful skill and put up with the jokes.

I've found the same thing. It's also a great excuse to get the rainbow pack of highlighters!

I’m going to make a note to try that this week.

Make sure you print that note out.

On an abacus running Doom. As the operating system, I mean, you can choose your own printer.

Not at all. Things like the Laserjet 4 were workhorses. No planned obsolescence in sight.

I do still call them the Laserjet 41.3, though.

> In 2020 a top consumer or business printer will not jam on you.

That's just not accurate. I've personally dealt with truly unreasonable numbers of service calls for high end business models over the last 8 years, jams aren't frequent but they are definitely not rare.

Usually in big companies that’s more about the garbage recycled paper that procurement idiots like to buy.

> Finally, the interfacing for printers today is fantastic.

Still in the year of our Lord 2020, once a document hits the Windows Printer Spooler, good luck trying to cancel/abort it.

From an elevated powershell/cmd:

  net stop spooler
  cd *:\windows\system32\spool\printers
  (remove all files in this directory) 
  net start spooler
Stupid, but this fixes like 80% of windows printer problems, most of which IME are random one-off issues with a document getting spooled and the printer not being activated

HP Laserjet were pretty good even pre-2000. I spent some time in publishing. And I'd compare it to nowadays paper feed on a $100-250 ink jet, that has couple of springs (id 3mm, L=15mm) that would just fail and the printer won't take paper any longer. Virtually no printer in that price range has replaceable rubber paper rollers.

Indeed- I only used 2000 as its now the last year for any patents to have been around. But all-in-ones, network printing, duplex, color-laser, were all around in 2000 and any fundamental patents on them should have expired. Wifi integration maybe not, but this seems pretty trivial to "clean room" in this day and age if this is somehow patented functionality.

I remember the paper jams but currently most consumer printers die through built-in obsolesce before the paper jams happen often.

We're trained to spent much more money for business laptops and probably we should for printers, too? Well handled ThinkPads last for many years and when the hardware struggles to execute the software in decent speed they - remain to work. Lenovo (and IBM) make money with reliable hardware.

I still use a 1995-manufactured HP LaserJet 5P. It rarely jams. Works great with a USB->Parallel adapter. The cartridge in there right now is probably a decade old.

I did IT work in a big law firm in 1989. The company had a fleet of HP LaserJet Series II printers (that having been introduced in 1987). They worked fine, same as anything today, with regard to jams and whatever. (Mostly, they were used for cranking out documents written in WordPerfect.) Though modern printers are more affordable and capable, the basic physical user experience is about the same. Modern printers don't even look all that different: vanilla boxes with some slanted aperture on top where paper comes out; drawer for blank paper. If I walked into, say, an insurance agency or whatever and they had a HP LJ II from 1989 sitting on a counter in a corner, I probably wouldn't even notice the antique. :)

> Do you remember printing in 2000...

Yes, and?

> Literally every printer was suffering from constant paper jams and other mechanical malfunctions.

No. I have gone through 7 personal printers and used countless business printers. Total number of jams I've cleared from my personal printers are probably less than 10 in 20+ years.

I've cleared much more paper jams from top of the line printers regardless of their build year. Reliability of a business printer is a function of its maintenance quality it seems.

My most stubborn printer wouldn't feed some papers since they're too smooth but, it started to happen after 7 years and its feed rollers were dry and worn down at that time.

One of my HP printers started to lose cartridge calibration, made funny noises and gave strange error messages after 6 years. Looks like I've worn down its internals. It was a "disposable" model it seems. I've probably used it three times its expected lifetime.

> The only thing that was better about printing in 2000 is that back then printing was more useful

It's still very useful at a personal scale. I've also cut back my printing to save the trees but, reading articles, academic stuff and a good old technical documentation is vastly better on paper, hands down.

> Do you remember printing in 2000? Literally every printer was suffering from constant paper jams and other mechanical malfunctions. In 2020 a top consumer or business printer will not jam on you.

A cheap consumer printer will jam on you and tear the paper to shreds, yet the drivers are still proprietary and you still get ripped off on cartridges.

> ...and you still get ripped off on cartridges.

Yep, it's cheaper to buy a new crappy printer from, say, Fry's Electronics than it is to buy refill cartridges for the last crappy printer you bought that ran out of ink.

In the late 90s we bought two laserjet 8000 printers that were still in use when I left the company in 2014. Outside of the regular service schedule they never required maintenance.

> Do you remember printing in 2000?

I remember printing on my NeXT Laser Printer in 1992 or so. Great experience, fast, reliable, high quality output.

Rare the pages that did not come out at rated engine speed.

Canon engine, tweaked to 400 dpi instead of 300, with the host doing the rendering in DPS (anyone remember machportdevice?) and a custom DMA interface delivering the bitmap directly to the engine IIRC.

No interfaces on the printer itself, no ports, no Wifi, no LCDs, no memory, nada. And none needed. Printer is for printing.

To be fair, my home printer is a Canon MX860. I bought mine around 2009 or 2010 and it has wifi, an LCD (albeit not touchscreen) and the only time I've experienced a jam, I was trying to print on sketchy paper anyway.

The best part is, the ink tanks for it are dirt cheap. Like $20 for 4 complete sets of ink tanks on amazon.

> Do you remember printing in 2000?

Yes I do. No there were no "literally every printer was suffering".

1995-2000 was a perfectly fine era of B&W desktop laser printing (others can chime in about how it was before 1995). Go get some computer history education, or ask an adult in the room.

I'll just point out the existence of the movie "office space" in here as evidence that yes, people were frustrated with printers. That movie came out in 1999, and the most iconic scene is the printer being smashed. That resonated for very good reasons, printers (and drivers) were almost more obnoxious back then than they are now. Please lose the superior "adult in the room" tone.

I mean, they were made because the printer was asking the player character to load letter paper, and they couldn't figure that out.

Ok, I don't know what PC stands for, but load letter means you're out of paper, please fill me up. It's not a blinking clock on a VCR.

"PC LOAD LETTER" was a very common message on the old HP LaserJet Series II printers. This cryptic message on the tiny LCD display was short for:

PC: "Paper Cassette" Load: "out of paper, please load more" Letter: "US Letter size (8.5"x11")"

In some offices you'd be equally likely to see similar codes like "PC LOAD A4" or "PC LOAD LEGAL".

The LaserJet Series II and III didn't have paper drawers, but rather paper cassettes, which you could load with a stack of paper (maybe a few hundred sheets? less than half a ream IIRC), and then swap in and out of the machine as a unit.

Some models (I think it might have been an add-on peripheral for the Series II?) had two cassette slots, so you could load two supplies of paper at the same time. Either you could load the same size twice, in which case it would perform like a backup: if you ran out of paper in the first cassette, it would switch to the second cassette automatically but start flashing a light to tell you that it was time to reload the first one. Or you could load two different sizes, and it would select the right cassette to draw from based on the size of the document being printed.

The Series II didn't have a way to measure the size of the paper, so each cassette was designed for only one paper size. There was an interface where the cassette plugged into the printer which indicated which size paper it contained.

Due to the design of the paper cassette, it was very hard to tell from the outside how much paper was left. There was a tiny window but it was nearly opaque. IIRC later versions of the cassette improved this.

So if you had a model with only one cassette (which was fairly common, I think) and you loaded a "US Letter" size cassette, and it ran out of paper, the printer would refuse to print anymore until you gave it more paper. In the meantime the tiny LCD screen would flash "PC LOAD LETTER".

The printer was a workhorse and I can testify that many offices in my area were using 1980s-era Series II printers daily in 1999, when I had a part-time IT job that among other things involved doing maintenance on said printers. Replacing the rubber paper pickup roller and clearing paper jams out of the fusing unit were probably the two most common trouble cases, IIRC.

My HP Laserjet 4L does not jam at all but instead prints and prints and prints and prints and .... since the 90ies till today and probably years to come.

Screams into void...

Where's my open source dot matrix?

> Do you remember printing in 2000? Literally every printer was suffering from constant paper jams and other mechanical malfunctions.

Paper has a right-side-up, and your printing life improves drastically if you just look at the little arrow before putting the paper in the printer.


> As the front and back surfaces of the paper, as determined during the papermaking process, differ slightly, one side is preferred as the side to image first. The primary determinant of which side to print first is the paper’s curl characteristics.

> If you are using a quality paper intended for digital printing, the ream wrapper will be marked with an arrow that points to the preferred printing side. Print on this side when printing one side only; print this side first when printing on both sides of a sheet.

> Whether this side is to be loaded UP or DOWN in the paper tray has to be determined for each machine (and sometimes for each paper tray) by reading the system's operator guide. Once you've determined the correct orientation, marking each paper tray with a label indicating the correct loading direction helps avoid operator error and lost productivity.

> Determining Curl

> In the event a paper ream is not marked for correct print-side orientation, it may be necessary to determine the curl direction yourself. Do this by holding a 1/2-inch stack of paper by one of its short edges (refer to figure 4-2). Let the paper hang with the long edge parallel to your body. Either the lower edge or the two side edges will be curling slightly toward the center. Observe which way the edge(s) curl. This is the curl side. Load the paper into the tray such that the side opposite the direction of curl is imaged first.

> Note: If the ream had an arrow marking, it would point to the OPPOSITE side. Load into the paper tray in the appropriate direction

> Built-in Curl

> Xerox papers are manufactured with a small amount of “reverse curl,” so that they will be very close to flat after processing – this will facilitate any post-processing that needs to occur, such as binding, trimming or folding. Load according to the arrow direction for best results.

> Loading The Paper Tray

> Carefully unwrap the reams of paper to be loaded, taking care not to bend any of the sheets or otherwise damage the paper. Inspect the paper for any obvious signs of damage (bends, folds, crumpled or wavy edges, tight edges), or defects. Fan the paper as necessary to avoid sticking edges. Do not handle the paper any more than necessary.

> Load the reams into the paper tray one at a time, taking care to observe the correct orientation, as indicated by the ream wrapper arrow.

> When more than one ream is being loaded, it is important to make certain the reams are aligned atop one another. It is easy to wrinkle, bend, or otherwise alter the top sheet of a lower ream when placing another one on top of it. The interface between reams in the paper tray is a frequent source of jams. It is particularly important to avoid loading successive reams inconsistently (some arrow up, some arrow down).

> Observe the paper fill line marked on all paper trays and do not load paper above this line.

There's a bunch more there about correct storage.

I had a burst of creative thought, thinking about getting a typewriter and hooking up a ton of actuators to it so I could just slide it an array of a document to type out. But it'd probably move to fast or slow, jam and then there's feeding it paper.

It'd probably be easier to make a nice block alphabet for a plotter and then just print your documents as biro drawings.

But again, feeding paper seems like a very fiddly problem.


Daisy wheel printers were slow, loud, and had huge limits (no kerning, single typeface, no printing family photos), but the print quality was good. And if you are the kind of person who likes mechanical keyboard sounds, the sound of a daisy wheel printer is pretty cool.


> feeding paper ... fiddly

Printer paper came in long, laser-perforated sheets with tabs. You'd load in the start and one sheet pulls in the next.


The crank you had to put into the front of the printer to get the steam-powered engine turning could jam in the transmission, though, and you had to watch the temperature of your coal-fired ink tank so it didn't over-boil. Those "electronic" printer guys thought they were so fancy.

The problem isn’t building an open source printer, the problem is building a good open source printer.

Yeah and resolution is another issue.

2017: https://www.youtube.com/watch?v=Awxbu8y5cv8

"A reverse-engineered typewriter hack to make it into a printer. Using a simple MOSFET circuit and an Arduino (actually, a Light Blue Bean+ arduino compatible board), I reverse-engineered my IBM Wheelwriter 6 typewriter to print out text and some rudimentary graphics. The GitHub repository is here, and I'll continue to update it with schematics, etc., when I get some time: https://github.com/tofergregg/IBM-Wheelwriter-Hack"

Same user has a similar hack for a 1960s Smith Corona Sterling Automatic 12: https://github.com/tofergregg/smith_corona_printer

You could look into the IBM Selectric - the typing element is a ball so it can't jam, and it's actuated by little cables that might be hackable. Feed it paper off of a roll, and you have your jam-free, continuous printing solution :)

A few years ago I took apart my Correcting Selectric III with exactly this aim. IIRC there are around six levers which can be actuated in different combinations to select and imprint all of the symbols on the type ball. I had a lot of fun selecting them by hand to try to map all of the combinations.

IBM must have had this use in mind because they actually made a variation of the Selectric design that could be used as a serial terminal. We have one in storage at work but I think the mechanism is seized. Wikipedia has a surprisingly long section about modifying the Selectric to work as a computer terminal[1].

[1] https://en.wikipedia.org/wiki/IBM_Selectric_typewriter#Use_a...

I remember it well. All of us computer hackers back in the late 1970s did something like that with a bunch of solenoids. It worked, but you got about 12 cps on a good day, and only constant width fonts. Also, if one of your solenoids was a bit sluggish it would throw off the registration of some of the characters. The advent of daisy wheel technology killed off the Selectric printers. Remember daisy wheels, especially the early metal ones??

OTOH hacking an IBM Executive might have been something. Proportional spacing!! (but a much fiddlier mechanism)

I have a vague memory that there was even a commercial option for that--a device that fit over the typewriter keyboard, that you could buy, and it would drive the typewriter from the computer.

I had for a while a NEC Spinwriter--loud, slow, built like the proverbial tank. My recollection--it's been more than 30 years since I used it--was that it did support proportional spacing.

You need an 1980s Daisy Wheel Printer. Problem solved.

> the money is in professional printing, ads, designers, etc.

And ink cartridges

Ink cartridges is only a $1B global industry [0]. That isn’t a small market, but the global printing industry, with all digital content being transferred onto any 2d mediums through automatic means is probably closer to $1T. Think about every book, every magazine, every coffee mug with a funny saying, every graphic TShirt. In the world.

The ink cartridges are good money, but it’s not where the money is at.

[0] https://www.rfdtv.com/story/42630937/global-ink-cartridges-m...

That actually upsets me more: we're beset by stupid DRM and price-gouging on ink, and it's not even that profitable?

Are we anymore? I've been to an electronics store recently, most printers on display were ink tank models. Even HP ones!

It's not the main source of revenue, but for personal printing most of the money comes from selling ink cartridges. My dad used to work for a printer company and he said that on personal printer sales they usually broke even or sometimes sold at a loss. But all of that was made up with the profit margin on ink cartridge sales.

That model is juuuuuust about dead. Canon/Lexmark/HP are now all pushing megatank/ecotank continuous ink systems. I bought a Canon Pixma G6020 for printing photos and I'm super pleased with it. It comes with enough ink to do 18,000 black and white pages and 6000 full color pages. Refills are around $40 for the set.

Unless the design has changed, there is a pad that absorbs overflow ink from cleaning the print heads. After some time, that pad will be full, and then the printer will refuse to ever print again.

I've now had 2 Pixma G series break that way, with a third well set on that course.

I cleaned one of those overflow pad assemblies once, on an old Epson. I got several more years out of it. But what a disgusting job cleaning it was. It's a great huge tank with many layers of pads. Do they sell replacements? One guess.

They don't sell replacements, although I've heard of people DIY-ing it. However, the far bigger issue in my opinion is that when the printer detects that the absorber is full, it throws an error code (5B00) that prevents the printer from functioning, and can only be reset by the manufacturer.

I have yet to get a full tank out of one, and very much regret buying them.

How is that worth it for them? I pay $40 for a cartridge of about 1/10th the capacity, and the printer isn't even that expensive!

Although I bet I still have to send my document to it six times before it starts printing.

There are lots of DIY plotters, but printers are just way more complicated.

bit of a non-argument this. Existence of cheap alternatives misses the point of opensource, its not about cost. ref: free as in beer vs free as in speech.

I think you're missing the main argument: It would take a lot of money and cross-domain knowledge to design and build a passable printer from scratch. Contrast that with software, where all you need is time and a low to medium power computer. In fact, compared to software open source, all other open source engineering is non-existent (relative to OSS, not in absolute terms).

Why would you design one from scratch? The hardware out there is excellent and inexpensive. Start with that and write the firmware in a similar way that the Magic Lantern group did for cameras.

Embedded systems generally do not allow you to flash custom firmware to them these days, especially for printers as there is money at stake (you could write firmware to support non-vendors Ink/Toner cartridges), or to enable disabled features.

In addition printers contain custom specific IP blocks related to everything from ethernet to scan and print functionality. It is just not feasible for someone to be able to write custom firmware for a printer without access to the hardware documentation for these devices. More sophisticated printers even have multiple CPUs for doing various tasks.

> Why would you design one from scratch? The hardware out there is excellent and inexpensive.

I thought the discussion was about open source printers, not about open source printer firmware? Sure, you could probably reverse engineer the interface to the actuators and sensors, which would allow you to write open source drivers, but you'd still be stuck with the printer manufacturer's cartridges.

The problem is not the design or the firmware. The problem is the production of the actual physical printer. Building real things is not like compiling software. Nothing happens exactly the same every time. The minimal exposure I had to manufacturing has shown me that even simple manufacturing steps are infuriatingly difficult to get consistent.

Sure, but money is a very effective motivational power. If something is expensive, there is a higher need to produce an alternative.

It's a matter of ROI.

There isn't much return or expertise on building an open source 2D printer, as opposed to 3D printers.

99% of FOSS consumers only care about free beer, hence the uptake on MIT like licenses.

If you remove FOSS from the sentence I believe you more.

If I take FOSS from the sentence, then piracy also counts and those don't care about the code in any form.

Where did you derive that number from?

Not parent but if 1% of the people who use my FOSS software [1] did pay for it, I'd be making a pretty good salary. I'm just a datapoint but my contributions don't cover the server cost alone and out of the many companies who did contact me, I've refused almost all of them as I'd be making more money flipping burgers than what they were willing to offer considering the time involvement

[1] https://github.com/mickael-kerjean/filestash

I'm confused, would you be making a pretty good salary or would you be making more money flipping burgers?

I was just looking for an s3 browser. Will give this a whirl

Gut feeling from what gets discussed over here and other channels when someone posts news about commercial tools or GPL dual licensed based stuff.

Hordes of complains about paying for something, followed by MIT like licenses based clones.

Printers are cheap and widely available, which leaves a DIY printer which is going to be slower, more error-prone and more expensive as a very niche idea. There is a clear benefit to buying or building an open hobby 3D printer, whereas that's harder to argue for a 2D one - while there is a lot of crap around, there's enough workable choices, aftermarket inks/toner works, you likely won't be modifying/tuning a 2D printer the same way you maybe would with a 3D printer, ...

Ink delivery is likely the main challenge (although I've seen some low-res attempts), combined with the speed and precision needed for a good printer - reaching a few hundred DPI requires positioning things quite precisely. Laser printers are interesting, but then you need specialized parts like the drum that I'd expect to be difficult to produce in single quantities.

Open pen plotters are a thing, but again not typically used for normal printing duties.

I don't think this is quite right. An open source 2d printer would allow for a lot of things commercial printers don't allow for. For example, I'd like to be able to print on things which are not paper. If I had an inkjet head mounted on something, there's a ton I could do with it, from modifying it to print directly on fabric, to integrating it with other fabrication technologies to make decorated objects. Even a CNC mill or laser cutter with an integrated printhead would be invaluable, both for labeling parts and decoratively (for making pretty parts).

There's a massive growth curve too. If we could find a way to print on plastic, we could integrate this with a 3d printer and make decorated parts. I think this would be multiple stages of amateur R&D, but it would eventually happen (yes, I suspect someone will respond with all the technical issues why it can't work with current technology, ignoring all disclaimers -- I am aware this won't work right now).

I think of tons of other use cases.

I think the problem is as others have described. Making a printhead costs peanuts, but engineering one and NREs are astronomical. Ditto for paper handling, and many other parts of the printer. There used to be an printhead open enough for DIY (you could buy them in quantities of 1, and there was a spec sheet), but it's not sold anymore.

> An open source 2d printer would allow for a lot of things commercial printers don't allow for. For example, I'd like to be able to print on things which are not paper.

Those kinds of printers already exist commercially. The argument is the same: Printing on clothes or PCBs might be cool, but crappy DIY printers that can do that are even more niche than crappy DIY printers that print on paper.

I think the argument is that $200 crappy is better than $2000 commercial for a lot of uses.

If I'm labeling pins and parts on a PCB, I'll take low-quality labels over no labels any day. If I'm labeling how wood fits together on a laser cut, I'll take it. If I'm making educational resources, quality almost doesn't matter.

If I could have a 1980-era printhead I could control, I could do a lot with it.

And if we had that, quality would improve with time. Look how many years it took 3d printers to be useful for anything practical. I expect if we started even with 1980-era inkjet quality, we'd get to nice in 5-10 years.

I also wouldn't underestimate the challenge of moving paper. I once spoke to someone who worked for Kodak (or a similar company - it was a long time ago) who said that they acquired another printer company purely because their paper moving technology was excellent and it would have taken years to develop an equivalent quality of engineering internally. This was before the present day race to the bottom in the printer market.

If you haven't read it already, https://www.newyorker.com/magazine/2018/02/12/why-paper-jams... is a lot of fun.

A bit more than ten years ago I worked for Ricoh on an embedded software that was reporting the state of printer. Had a lot of fun trying to fold papers in different ways to try to trigger jams at different place of the printer to test that we were reporting jams accurately :)

Excellent long-form popular technical(ish) article; well worth the read.

Yeah. printing on a single hand-fed sheet: probably "easy". Reliably going through a random pack of 500 pages, not so much.

Something pen-plotters don't do (they typically want paper to be placed down for them, or work of a roll of paper), and the maybe 3D-printer equivalent of preparing the print bed and removing prints from it is a well-known source of problems and manual work.

I think that's a key point you've made, which is that open-source 3D printers don't do the part that would make them useful for scale production (transitioning the work pieces off the print surface and preparing for the next run).

Nobody would accept a 2D printer that took manual intervention every sheet.

Having taken apart a color laserjet to repair the paper uptake I agree. That was a lot of engineering to make sure the paper moves through the path repeatably.

While I'm sure it's a challenge, after watching 3D printers evolve over the years, I've no doubt that random people will find unique and amazing ways to handle the problem.

Paper printers have been around for ages though, and decades of R&D have been put into them. 3D printers have started to find a foothold relatively recently.

The FDM 3d printer model took off ONLY because the patents around the FDM printing head expired.

How about laser engraving tech? Can a laser engraver be tight focused to 300 dpi and power set (or sensed) to just make a mark on white paper? Does the positioning accuracy also need to be improved to get that 300 dpi?

You'd never run out of toner at least.

Edit, answered at least one question: yes engravers do 500 dpi routinely. Here's one: https://www.troteclaser.com/en-us/knowledge/tips-for-laser-u...

The issue is then speed, you can't engrave at anywhere near the same speed as you can print.

I would not consider an engraver for which you have to call to ask the price as "routine". From what I could find, the cheapest trotec is $16k second hand on ebay (and there is another from their cheap line for $56k second hand). You can hire somebody to handwrite everything you need for that kind of money.

Also, high power lasers are consumables.

I didn't suggest using an engraver. I was suggesting engraver tech for this thing we're proposing here. The real question is can you get a good result marking paper with a laser? Maybe I worded it poorly first time.

Or we could get rid of paper and just print everything on plywood.

> Printers are cheap and widely available

Then how about a printer that can print on a large wall? E.g. with spray paint.

See also: https://www.eff.org/pages/list-printers-which-do-or-do-not-d... (https://news.ycombinator.com/item?id=14501894)

Some of the documents that we previously received through FOIA suggested that all major manufacturers of color laser printers entered a secret agreement with governments to ensure that the output of those printers is forensically traceable.

This and exactly this; 2D printing technology is sealed tight because of those damn microdots. Lulz

I wonder if this is why color printers won't print out black & white print jobs when they run out of color ink, even if they still have black ink.

Yes. But this is unrelated to the question they asked.

I don't think so. The current situation suggests that we really need an open-source printer, because otherwise there is no privacy at all.

The question wasn’t whether there was a need. The question was why there isn’t one. That agreement doesn’t prevent you or anyone else from making your own design.

A common answer to "the question" is that passably good printers are dirt cheap, and excellent printers are highly complex multidisciplinary efforts. Things don't exist in the open source world unless somebody has a need, or passion, to make it happen.

Given the availability and complexity, the question becomes "why bother." Microdots are an answer to that.

Then I'd ask, is there a legitimate purpose? Supposing the end goal is a totally open source chip foundry: yeah, you don't want microdots on your masks. Edit: oh god, and printers that won't do b&w without yellow dye available.

[citation needed]

My printer is out of yellow ink but it won’t let me print in Black and White mode.

I work with inkjet printers in research for my Masters program. I found the three blog posts about printing on this blog to be helpful: https://www.kylescholz.com/

The author describes their work to make a very simple DIY inkjet printer for under $1000. While they are using a nozzle that they purchased, you can make a similar one yourself (check out the book "Microdrop generation" by Eric Lee).

All-in-all it's fairly complicated just to start printing droplets, to say nothing of scaling beyond a single nozzle or precisely moving the printhead.

And if you're one of the lucky 10,000 today, Bill Hammack over at the EngineerGuy youtube channel has this[0] fantastic video explaining how piezoelectric microdot printers work.


You answered your question with "apart from the nozzle"

It is the nozzle. Everything else is very simple to make because it is already done for 3d printers that are more complex than 2d printers(if you do not consider the nozzle).

5 years or so ago I made a 2d plotter with friends at my 3d printers community with the reverse engineering knowledge that we had about a specific cartridge with nozzles on it.

Printing with ink was easy, very easy. But we were interested in using it for 3d print wax, not so easy.

You need to manufacture nozzles, and that requires lots of money. That requires manufacturing plants. Very cheap in volume, but requires volume.

Open source has not volume in the millions, like big companies have, and those companies are not going to sell you the nozzles so you commoditize their professional field like linux did.

Why can't you 3D print a nozzle? Or are there electronics in there too?

... because the channels through which the ink is propelled are only microns wide, and the ink volumes which are distributed per drop are in the picoliter range which requires a complex piezo-electric control system.

If this topic turns out to be the start of a FOSS, I'll join as a (senior) design engineer.

Find one that doesn't have a patent and reverse engineer it.

I imagine the first/original control system was extremely difficult to develop. But since the technology already exists you can reverse engineer it.

After that, you pay a supplier to make parts. In my experience, these suppliers will end up redesigning to match better with their process, but still meets critical specifications.

I'd post my contact info, but I believe I'm targeted by an HN mod.

> Find one that doesn't have a patent and reverse engineer it

At this point, it would be better to start looking at expired (and about-to-expire) patents. A circa Y2k printer would be a perfectly acceptable design, IMO.

I completely agree. But given I got 4 downvotes for offering to contribute to a FOSS, seems like demand is low.

I'd be willing to bet you got those downvotes for an unverifiable and unlikely allegation that an HN mod is currently targeting you.

Glancing back through your posts, I see pretty clear reasons why people would have downvoted your gray ones.

Do you think I need to be an HN mod to downvote you? I don't.

Friendly reminder that rms launched the FLOSS movement because he was incensed when he couldn't program a printer at MIT to do something he wanted.

I read Kernigan's memoir on Unix and one of the things that surprised me is how much of the book is devoted to Bell Labs trying to get their fucking printer to work. These are people like Dennis Ritchie, Brian Kernighan, Alfred Aho, and the like wasting months on the broken drivers and flaky hardware of their document printer.

TIL Richard Stallman's middle name is Matthew.

Thanks for point this out. In my head it's always Richard 'Mainline' Stallman, like a competitive video game handle, no idea why.. I knew it was absolutely wrong but never cared enough to Google it. Now I know.

If RMS is involved, maybe try DuckDuckGo'ing instead of Googling, I'm sure he would prefer it that way

So true. And so sad that there's no open source printers yet. :(

I'm actively working on one of these right now.

What you are talking about I refer to as 2.5 axis machine vs the traditional 3 axis PLA/FDM printer. Aka a plotter. Using an inkjet cartridge or a laserjet toner on a piece of paper outside the context of the printer it was designed for seems foolhardy at best... but what about moving a pen up and down?

Shameless plug, I've been working on a project called Robot Draws You! (www.robotdrawsyou.com). I'm currently using an off-the-shelf machine and the software / cloud hoops it requires me to jump through were enough to convince me to build my own machine. For the proof of concept I'm using a Duet2 board, but eventually I want to write some code that will sit on a raspi and talk to the Duet to allow the machine a more granular drip-feed style control over the "printing" process". More on that later.

"Why is it hard?" The challenge starts with taking in a given SVG file, making sure it scales / fits within the bounds of a given writeable area, and then generating GCODE to send to the printer / plotter. Because there's no extruder, custom GCODE needs to get created to take advantage of the GPIO pins to move a servo up/down to control the pen. The software challenge is replacing the much-hated cloud interface I complain about. It may suck, but it does a lot and it actually works.

The more I use "the cloud", the more I am reminded it does not provide adequate controls/info on:

- The size of the rendered image relative to the writeable area

- The order in which the layers of the file get rendered

- Information about the progress / time left per layer

- Repeatability of failed layers without re-writing entire project

So crazy me decided "I'll make my own plotter UI and hardware!" It's slow going but it's really fun and I enjoy the challenge. The end solution is going to be a mix of hardware and software that allows you to upload an SVG / vector file to a web UI, start/stop/repeat layers and control the order of the rendering. I like to make drawings of people, and also want to use this to make gigantic maps as well.

.. I mean I'm working on my own plotter too, but only because I'd like something larger format than my off the shelf axidraw.

There are tons of OS boards and software already developed around this problem, you don't have to do it all yourself.

A 3D printer needs only a comparatively simple metal novel, and the rest of the printer is just moving that around, some other things like heating the pad to help things work smoothly, and processing a file into a set of movement instructions for the head.

A 2D printer needs to deal with four or more liquids (ink) or fine pieces of plastic (toner). Rather than just heating the ink up, a tiny electrical current is used to squeeze out a drop at a time. Everywhere the liquid touches can get dried up, and needs to be self-cleaned. And then you have to address the color mixing algorithm, calibration, ICC profiles, etc. There are waste ink absorbers, print heads, etc. many of which involve specialty materials that can only be made in a precision factory, which would not be available for open source development.

First, accurate and reliable mechanical manipulation of paper is tricky, to get all kinds of different paper stock to feed reliably without jams and position in the same place.

Second, decent reproduction of e.g. text at decent DPI requires more accuracy in head positioning than what you need for basic 3d printing.

So to me it seems that the big issue is not that a random cheap standard printer would be cheaper, but rather than the home-built version is likely to get worse results than what you can get in store for peanuts.

Big point about OS for me is hackability what if instead of printing ink on paper you want to print fungi spores on jelly.

or names on birthday cakes. cutting guides on wood. UV paint on metal. Jokes on toilet roll :)

thinking there might be mileage in this idea. clearly a mod from a 3D printer design would be where to start. standardise ink delivery / modules for different materials with different characteristics. A standardised printer driver for Linux and other platforms would get a lot or re-us.

Open source Pen Plotters/ Pick and Place machines all already exist and are common in the Maker world

Attaching a ballpoint pen to a 3D printer/CNC machine/laser cutter seems like a good start. It would be low DPI and slow, but fine for drawing and larger text.

A FDM 3D printer is mechanically relatively simple; it's more akin to a 3D plotter: it's a vector thing. Household 2D printers do bitmaps.

There's 3D bitmap printers as well, like the SLA resin printers and laser sintering.

The bitmap thing at a high resolution requires higher precision equipment than a typical 3D printer which makes it more difficult to do as a hobbyist. Not impossible, but a factor.

Yes, I learned a lot about inverse kinematics when I was upgrading my 3d printer to be used as a 2d plotter (simply printing a pen holder and figuring out how to create the GCODE).

Drawing a circle with X/Y stepper motors is more involved than you'd think. You can either resolve every step with a high resolution (command the steppers to move fractions of a millimeter at a time), or approximate the curve and let the printer do its own IK to figure out how to best represent them in real space.

Both have advantages/disadvantages. The high resolution approach can product more accurate lines, but can sometimes require more computation than the little MCU is capable of (which can cause inaccuracy of plotting as the printer fails to keep up)

I'm unsure about the statement. I've repaired many printers mechanically. The hardware is generally repairable and the mechanical parts aren't proprietary.

The software on the printers is whatever you buy. Some of them run on available print languages that you can code up your own driver for should you really want to.

And for the example cited, cartridges, Epson had offered a more expensive printer with a do whatever you want ink setup. And Brother lasers have so far accepted any carts I use without complaint.

As for hard to manufacture the entire printer is a molded plastic to reduce cost (massively) and a fabrication plant made printhead. Mass manufacture only for the huge price-break.

2-d printing is mature, I think everyone has high standards.

When it comes to 3-d printing I see two attitudes around me:

(1) People who are involved with "making" from a blue collar standpoint think that "3-d printing is cool but the quality of the product is subpar" (2) People around the engineering department at my local Uni who 3-d print everything they can

A lot of the 3-d printing market targets type (1) and enthusiasts. If those enthusiasts were inkjet enthusiasts they wouldn't mind getting prints spoiled with an ink explosion 5% of the time.

I think the reason is open source communities are more attracted towards tinkering and experimental technologies, and 3D print is much more interesting than 2D printing. 3D printing always has that idea of a manufacturing revolution behind it, 2D printing was a revolution centuries ago and it's a boring technology now.

Yup. Best case scenario after spending tons of hours and money building a decent open source printer, is you match the abilities of a $50 off the shelf printer you can buy right now.

Mostly the nozzle. The nozzle for an FDM 3D printer could be easily built with desktop tooling from a hardware store. The nozzle for even a cheap inkjet is not within the reach of hobbyists.

Open source 2D plotters do exist.

reading through the comments here the answer seems to be: good printers are cheap and nozzles are hard.

But how about the openwrt-approach? Keep the printer with the good nozzle and starting with the software such a cheap printer runs, trying to get that foss first - jailbreaking your printer if you will. Or swapping the board for a raspberry pi or similar? That would at least help against closed drivers, cartridge restrictions, page counters etc.

There's at least this: https://hackaday.io/project/167446-diy-inkjet-printer

But as others have mentioned, commercial printers are really cheap, especially for how complicated they are to replicate so there's not much motivation to make them.

The printers are cheap (if you also need a scanner price is going up a little). The ink isn't cheap though, its artificially high, and all the major printer manufacturers actively hamper your ability to go for a non-branded ink version. Cause that is where their profit margins lie.

Related: I have a HP 4000 with duplex, which is an absolute unit (beast) of a small workgroup laser printer. It's 20+yrs old, still going strong and parts are easy to get.

BUT, it's compute bound with modern print jobs, and is missing modern protocols like Bonjour.

What if someone open sourced a legacy printer? I'd love to re-brain this printer.

See about replacing the network card on the printer, or updating its firmware. Newer firmwares frequently support Bonjour, and have various security updates.

Replaceable NIC in printers are pretty great, especially when you can buy them for 1/10th of original cost on the used market.

They stopped manufacturing the HP LaserJet 4000 over 21 years ago, and the jetdirect EIO network card (J3111A) not long after that.

If there is a newer network card or firmware that supports this printer, I’d love to hear about it.

Awesome, thanks! I didn't realize that the newer EIO cards were so backward compatible.

Put a different server in front of it instead of letting it hang on the network by itself?

I have a relatively modern LaserJet that does IPP (and Bonjour) over WiFi, but it doesn't do 5GHz WiFi, it doesn't do IPv6, so I have it plugged into a server over USB instead, and the server running CUPS exposes it on the network.

I thought about that. The real issue I’m running into is page rendering time: I think the processor on it is only 50 or 80 MHz, and modern software generates some pretty elaborate print files.

I’ve been meaning to explore server-side rendering, but haven’t got that far down my todo list.

I wish that the intelligence in printers would migrate back to the computer. Having printers with their own complex formats (PCL, PostScript, etc.) only makes them inscrutable and lends itself to proprietary competing formats and printouts which never gets quite right. I think the original NeXT machine was on the right track with their simple bitmap-only laser printer, where the PostScript processing was done on the computer side.

That's interesting that you feel this way, personally I take the exact opposite position. I want printer drivers to be a thing of the past. I just want to pass the printer a file that represents a document and let it handle things for me. Some kind of standard format, probably something in the Postscript/PDF family tree.

Likewise for scanners. I want to just have a HTTP API or something like that where I just "GET" an endpoint and the scanner hands me a file in some reasonable format.

The drivers are always the worst part of devices that allow computers to interact with paper, so let's get rid of them altogether.

> I just want to pass the printer a file that represents a document and let it handle things for me. Some kind of standard format […]

…like a bitmap? Turning documents into bitmaps is easy to do in the computer now. We’re a long way from the time when (due to limits on memory, processing, and bandwidth) you had to encode your document into a complicated command language which you could send to a printer.

> […] the scanner hands me a file in some reasonable format.

Like a bitmap?

I really don’t see how you’re disagreeing with me, here.

Given the size of the damn drivers you’d think this was already the case.

When the manufacturer of your printer offers a "driver-only" download, rather than "complete customer experience" version, it is quite small. Also linux.

This has already happened. The printer under my desk has crap drivers that don't properly handle ligatures leading to the printer being unusable. A certified native Postscript printer would never have this problem.

I would say,it is difficulty to access of technology. Inkjet Printers are either pizo or heat nozel based, both tech works at microlevel and are built with semiconductor based fabrication methods, which is mostly outof reach of enthusiasts.

"In the 3D printing world, there are plenty of open source choices, ..."

The 3D printers you are talking about are very simple. Those are Fused Deposition Modeling and StereoLithography printers.

But there are also 3D printers that work like 2D printers. For example Selective Laser Sintering and PolyJet printers. Just like 2D printers they are very hard to make. Those type of printers are also not available as open source choice.


cough I'll just leave this here. Want some coffee, two sugars, right?

Point being, there are quite a few alternative use cases that commodity printers have going on under the covers that no one tends to talk about all that much. There's forensic watermarking for one but also supposedly certain features hard-coded in where if it detects it in an input, it intentionally leaves it out as an anti-counterfeiting measure. The article for that one was floating around on HN a while ago. I'll see if I can dredge it up.

Making copies is one of those things where there are several opportunities for power consolidation to be had if you look hard enough.

It's like the whole issue with 3d printing of guns. No one in an authoritative position necessarily wants everyone to have the capability to generate at will perfect duplication of information due to the consequences that spells for several entrenched, high relative value use cases.

Because a good 2D printer costs peanuts.

As a relatively new owner of a 3D printer, cost has very little to do with it. I spent way more on upgrading and tinkering with my printer than the thing cost in the first place.

If there was a 2D equivalent, I'd be interested in tinkering with that, too. And yes, that would probably mean spending 5x ot 10x more than a cheap commercial printer to get a printer than underperformed said cheap commercial printer. But it would be mine, and I would understand it, and be able to mess with it. And fix it if it broke. And use weird inks with it. And so on.

Would you do that tinkering if you could just buy a 3D printer cheaply that didn't need upgrading and worked perfectly?

You can probably adapt a printhead from a printer to your 3D printer today if you want a crappy DIY 2D printer and go from there.

I got into it specifically because you can get good printers cheap. There is a manufacturer that offers an open source based printer that was better than it had any reason to be for less than $300.

My printer today looks nothing like the printer I started with, I’ve switched boards, recompiled my own firmware, the whole bit. I was confident in hacking on it for two reasons:

1) It was open source based

2) Replacement parts are inexpensive

I originally started with a Makerbot 2X which was closed source and $2500 (8+ years ago). Mind you, I didn’t have one at home- we had a couple at work for prototyping.

The slicing software was atrocious, the ability to fix it was hampered by what parts they sold, and the replacement parts were generally very expensive.

My $250 printer was better in every objective measure out of the box than the MakerBot was after years of practice and tuning.

This year my friend’s 2X died and a replacement motherboard was going to be almost as much as the cost of my printer new so he just gave it to me and bought a new printer.

With my newfound confidence from my “cheap” printer, I gutted the 2X and installed an open source motherboard and completely rewired it.

It all started with a printer good enough to be liked by the community and open enough to be modifiable.

Nothing like a extruder that is not user serviceable and only prints PLA. When I was told that my office went through several of those extruders, I took some year end left over funds and purchased a lulzbot (Taz5). As far as I know that thing is still doing the work they need it to do.

Luckily the Replicator 2 and 2X were when MakerBot was still trying (or pretending) to be open source. The hotends were standard-ish MK8 (or 10). Everything was pretty serviceable on them as long as you didn’t need to fix the motherboard.

The printers after that switched to the new hotends that are not user serviceable. It boggles my mind why any home user or tinkerer would ever sign up for that crap.


My brother in law teacher, who teaches various middle school classes and groups use them. I asked why they would waste so much money and he had a response that made it all make sense to me. He doesn’t have time to keep them all running all the time (like we had to do with the old ones) and when one goes down he gets replacement parts from MB at a discount, slaps it in, and they’re up and running again. Apparently the latest ones are fairly reliable and it doesn’t happen all that often either.

So I guess I would sum it up as not for me, but I could see why some would.

Good question.

I run Linux (and on a Purism laptop so I can take it apart if I want to), so I guess I've made that choice already in a different context - "yes".

I haven't tried it, but I expect a commercial print head to be extremely specialised and adapted to the specifics of the printer it's in. And of course no documentation on what the connections are or how it hangs together. I think it'd be above my skill level to get this working.

Yes but then you have to buy cartridges from the manufacturer who has the control to say there are empty when in fact they are not. It is a completely closed ecosystem. I believe customers deserve better alternatives.

For many brands you can easily buy off-brand cartridges, and nowadays quite a few sell printers that work of ink tanks and don't even have cartridges.

Just ask your local refill shop what printer to buy so you don't have to go through this. That's what I did when I last bought a printer.

There's a very lively market of refilled / aftermarket cartriges which lowers the incentive to fix these issues.

But are these printers open source thus allowing it to be tinkered?

In the 80s there was DIY 2D plotter Alfi [1] for ZX Spectrum. I had and used one. This year, i visited my friends trying to run current DIY 3D printer, and funny thing is that it was plagued by similar isses like the 2D plotter several decades before.

[1] https://www.youtube.com/watch?v=lq2fGkQO_gQ

It's not so much that it's hard, but rather so far just hasn't been worth it. The average small business owner / consumer spends what, $100-200 for a printer and maybe $100/year on consumables if they print a moderate amount? The pain point just isn't there compared to the 3D printing world in the pre-RepRap days of a little over a decade ago when a printer cost thousands of dollars and $100 in consumables in a week or less.

For something as inexpensive as 2D printers to really get some interest in the 'open' world, they'd probably need to start being as obnoxious as the mobile phone market. Think changing printer languages (i.e. PostScript and PCL) in backwards incompatible ways every year or two and requiring changes to 'new and improved' incompatible consumables (i.e. ink and paper) every so often while cutting off supply to existing customers of the old consumables well before the useful life of the printer has been reached.

Here is a business idea: the printer that just works.


- laser printer that works with non-proprietary toners

- USB plug & play on all major operating systems

- Network printing via WiFi and ethernet

- No 2GB bloatware installation required, i dont want your shitty photo management software, just give me the driver

- Replaceable parts

Buy any Brother printer and use the generic Linux driver. Done.

Well, perhaps not the "replaceable parts", but that is doable with a 3D printer. I doubt Brother would come after you for printing a new output tray hinge or custom-colour button set ... and I doubt any other parts are likely to outright 'break' (my laser printers of various brands have lasted around two decades each). Any other parts, such as the image drum, would require too specialized fabrication to be done at home or at the local fab lab.

I bought an HP3600N about 15 years ago for something like $700. I’m not 100% sure if it has USB because it hangs off my network (which then Ubiquiti devices provide WiFi). It’s worked reliably and 3rd party toner is readily available and works well. If a major component wore out or failed, it would probably be cheaper to replace the unit than the part, but parts are technically available. I’ve had an excellent networked color laser at home for around $75/yr including toner.

Yep, older printers are the best bet. I have an old laser that simply just works (it's connected with a parallell cable, though). Consumables are readily available, and many, many older printers support ESC/P and works with any operating system newer than Amiga OS 2.0.

Brother is probably your best bet.

Pick a Brother that has all of the following features:

BR/Script (their PostScript emulation)



And you will be fairly happy. If any of those are missing, you're buying an inferior line.

Thanks, i'll look into it.

I currently have a Samsung laser toner (M2026W). I bought it because its farily compact and blends in to my little home office corner in the living room. The duplex machines tend to be a bit larger, a pitty.

Printing - when it works - is pretty good, just the WiFi is extremly unreliable and i end up having to re-configure the WiFi from scratch by connecting via USB and using some proprietary Samsung bloatware. Horrible.

> just the WiFi is extremly unreliable

do you have an option to use an Ethernet cable?

I have a multifunction A3 printer / scanner from a major manufacturer with an auto feeder and 3 other trays/inputs. It has USB input, network cable, wifi, internal tanks, fax capability, a colour touch screen....

The hardware works very well. The software is weird and inconsistent. It can do some very useful things if you access the scanning function in one, incredibly convoluted way, but not in other ways. It can scan to a network share, but you have to put the password in every time. It's frustrating because it's so good and so bad at the same time.

I would love to have an open source firmware in it.

On October 23, 2012, I paid $133.76 for a "Brother HL-2270DW Compact Laser Printer with Wireless Networking and Duplex" and I have only ever used $12-15 third-party toner cartridges with it.

I had to buy a replacement drum unit in 2016, that was $22.59 (third party, of course).

I find it unlikely that an open-source DIY printer is going to result in something better or cheaper than what I've got.

The one thing that it would do is to eliminate the search for a well supported printer on linux, as you would know that the open source ones were well supported.

The situation that always seems to happen to me is I need a printer ASAP. When shopping, I might find support for the Brother HL-2270DW in linux, but my local store might have a "Brother HL-2275DW" in stock (made up model number). So I'm wondering if the driver for the 2270 will work on the 2275, etc.

Pretty much any modern printer with a network interface supports IPP.

I have Samsung CLP-310 Color Laser Printer from 2008. It was the last one ever, which you can force-reset so it does count how many times a color cassette been used. I can refill those cassettes from half liter buckets from ebay for €10. It is just incredibly reliable.

One and only problem is that only the XP-windows Samsung printer drivers work truly well. But I have 2008 laptop for that.

I used the open source foo2zjs driver with this printer successfully for quite a while in case that's helpful.

Various Linux drivers work quite ok on text and cartoons. But the Samsung Windows XP driver has some special magic to make perfect photographs. Not even Samsung's own Linux drivers can do that.

I'm way more interested in laser than inkjet.

I casually poked around to see if I could find some old HP LaserJets. Do some teardowns, get some part numbers, maybe transplant a Raspberry Pi for the brains.

I don’t think the nozzle would have to be open source for this to be useful, just a printer that can be modified to print on things other than paper and be cool about whatever cartridges you put it it would be great. Inkjet cartridges are a little messy, but easy to refill at home, whit the printers DRM like check’s being the main issue. This project reverse engineered cartridge control and developed an ESP32 controlled hand held printer: http://spritesmods.com/?art=magicbrush&page=3

In addition to what's been mentioned, printing has much tighter tolerance requirements, e.g., to distinguish between "h" and "b" at a small font size. It actually has to be a very good spatial representation of the design. And how cheap printers actually accomplish this with crude bits of plastic and stamped metal is pretty amazing.

On the other hand, 3d printed objects may only need to satisfy overall mechanical needs, or be suggestive of the shape that they model, to serve a purpose. Most 3d printing that I've seen needs a bit of hand tooling at the end, to really be useful.

It's the nozzle. Let's suppose you really wanted to design your own nozzle... you could use something like MUMPS, which is $5,800 for 15 dice, plus 100 for each additional. You have to wrestle together 15 people to pay 400 bucks (or 20 to pay 300, and so forth) for a printhead that probably won't work. Figure it'll take several iterations and everyone needs to be in for thousands and thousands of dollars.

It's doable but only among a group of very wealthy hobbyists.

People with this kinda money typically gravitate towards high vacuum projects and microwave electronics

This is ignoring all the high-speed cameras and lab equipment for characterizing the electrical and rheological properties of the ink you shoot through it...

A different take on this question: Is there a whitebox/ODM/shanzai print engine that could be combined with an open controller board or RPi? What companies actually make the hard parts of printers?

Not a printer, but there is an open-source printing press: https://openpressproject.com/

Patents are not an issue on having opensource 2D printers but the average hackers don't have the manufacturing capability to make an inkjet print head. There are lots of design on the internet on how to build your own inkjet printer. I think it's not exciting enough to explore in a garage since we can pickup a decent printer from the garbage. I am more excited on parts I can salvage from an office printer than thinking of building a better one.

Because RepRap [1]. Proprietary 3D printers already were on the market.

Because you can print 3D printer with 3D printer. You can make paper templates with paper template.

[1] How he started the worldwide 3D printing revolution / Adrian Bowyer https://www.youtube.com/watch?v=VV0Tjwq7Uc0

3d printing is exciting new tech. Paper print is way over it's most exciting times. This would make great advances in developing a open source 2d printer not be that much exciting.

And that's probably also why there's so much focus on pen plotters they look special like vynil discs. And that triggers interest.

This is an interesting project, a 3D printed 2D CAD drawing machine. Basically a 3D printed plotter!

GRBL Plotter Elegoo: https://www.youtube.com/watch?v=XYqx5wg4oLU

Printer manufacturers don't make money on printers. They make money on ink. Making money on ink requires a closed system to require the ink purchase. There is aftermarket ink, but companies are always working to defeat it.

Maybe if you move the printer head over the paper (like in 3D) rather than drag paper though the device you could make something simple (b/w) with impressive life span. I have no idea if such printers exist.

i think the 2D equivalent to a 3D printer is not a regular printer, but a plotter. and there's a fairly robust community around hacking pen plotters. for eg, https://github.com/beardicus/awesome-plotters

I would check out Axidraw instead, much better!


I don't know about 2d printers, but there are lots of open source/DIY 2d plotters.

Or washing machines, dishwashers, cars or even planes?

I'm afraid the answer is due to the fact that we live under a corporatocracy, and that this is the way things roll round here!

Good luck with anyone who is trying to get a open source version of whatever out there. Unfortunately, for one reason or another, I suspect it won't work out, but I have my fingers crossed for you!

In a world where screens are everywhere who needs printers ?

I doubt I've printed out more than 100 documents in the last 20 years.

There are still a few legacy areas where nothing else will do like legal procedures and shipping labels but for the most part printers just don't seem like a broadly useful enough technology to interest most open source enthusiasts.

There are many open source pen plotters. Welcome back to the 80s!

Because they cost nothing to buy.

what's a 2d printer

You have to analyze the reality of this world. First off open source cannot exist without closed source.

Imagine if everything in the world went open source. Then nobody would be getting paid, everyone would starve because they're giving away there work for free.

Most things in this world are profit oriented products produced as a direct result a capitalism. Open source is an offshoot phenomenon in software arising because software is both easier than other forms of engineering (see thousands of bootcamps) and also easily copyable.

However it should be known that most software developers need to have a job in closed source software in order to pay the bills.

Whenever you see something open source you have to know it's an offshoot phenomenon. These are side projects spawned by intense interest but ultimately still a side project to a person's main line of work which is ultimately profitable. Be surprised that there are 3D printers because it's abnormal. The fact that there are no open source 2D printers arises because there's lack of interest and because there's no profit in open source 2D printers.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact