Repurposing Inkjet Technology For 3D Printing

You would be forgiven for thinking that 3D printing is only about plastic filament and UV-curing resin. In fact, there are dozens of technologies that can be used to create 3D printed parts, ranging from welders mounted to CNC machines to the very careful application of inkjet cartridges. For this year’s Hackaday Prize, [Yvo de Haas] is modifying inkjet technology to create 3D objects. If he gets this working with off-the-shelf parts, this will be one of the most interesting advances for 3D printing in recent memory.

The core of this build is a modification of HP45 inkjet print heads to squirt something other than overpriced ink. To turn this into a 3D printer, [Yvo] is filling these ink cartridges with water or alcohol. This is then printed on a bed of powder, either gypsum, sugar, sand, or ceramic, with each layer printed, then covered with a fine layer of powder. All of this is built around a 3D printer with an X/Y axis gantry, a piston to lower the print volume, and a roller to draw more powder over the print.

The hardest part of this build is controlling the inkjet cartridge itself, but there’s prior work that makes this job easier. [Yvo] is successfully printing on paper with the HP45 cartridges, managing to spit out 150 x 150 pixel images, just by running the cartridge over a piece of paper. Already that’s exceptionally cool, great for graffiti, and something we can’t wait to see in a real, working printer.

You can check out [Yvo]’s handheld printing efforts below.

38 thoughts on “Repurposing Inkjet Technology For 3D Printing

  1. Great minds think alike. i was recently brainstorming on using the shredded scrap paper from my paper shredder and some type of “glue” from a modified ink jet cartridge. didnt get very far, looking for low-visc glue turns up a bunch of chemistry stuff i idn’t understand.

    1. Use sodium silicate solution as the liquid, and sand as the powder.

      Once the part has been printed, lightly blow CO2 gas up from underneath, and the solution will turn to silicate. Everything glued together by the solution will be hardened, with the consistency of a cookie.

      Use that directly for sand casting (IOW, you can directly make sand-casting molds this way).

      One advantage is that inkjet inks already contain SodiumSilicate solution, so the cartridges/jets are already designed for it.

          1. I wonder if there can be a design where there is an exhaust to output the CO2 over the localized volume with an intake to suck back in the extra CO2 that isn’t absorbed if the rate isn’t stochiometrically calculated to be dispensesed for what can react for the time to set/harden enough.

  2. In my practical term at a faculty of TU Munic a had the chance of working on a 3D Printing System that used PMMA powder for support and building material and utilized a inkjet printhead to distribute a solvent on the printbed. The machine is used to develop a process to integrate electronic components and conducting paths into the printed part.
    here is a link to a paper about this my supervisor authored.
    https://www.emeraldinsight.com/doi/abs/10.1108/RPJ-12-2014-0172

    @Yvo, think about a procedure for cleaning the printhead as they will inevitable come in contact with the powder and clog … ;D

      1. anything UV cured has a way too high viscosity for the microscopic nozzles of consumer inkjet printers.
        There are actually some specialty industrial inkjet print heads that can handle some UV cured fluids. iirc XAAR and Kyocera make some (and maybe Epson did too). But for one thing such a head cost way more than the complete customer inkjet printer and for another, the resolution is much lower

        1. Correct. The Connex series of 3d printers use this technique. 8 industrial inkjet print heads and two high intensity uv lamps to cure the resin. 4 heads are dedicated to a water soluble support material and the other four can run two different materials.

          The material is much thinner than sla resin, but that lends itself to being able to weld broken parts with a portable if source.

  3. I think the difficulty here is that in regular string printing the material out of the nozzle lends some support before it touches the bed. Like a tube of toothpaste. Inkjet printers spray droplets that have no support till they hit the bed.

    1. It seems to me you are confused about how this form of 3d printing is supposed to work. The print head sprays a binder on a solid material that will make up the finished printed object. The liquid from the print head is only a glue of sorts. I will give you that a 3d printed object printed from material sprayed from a print cartridge would be freaking awesome. Maybe a pair of cartridges that could spray a 2 part epoxy, but, that isn’t what is happening in the current suggested process. :)

  4. Can’t wait to see how this works with a full page width print head. Maybe even offset and stack the print heads and powder rollers to lay down multiple layers in one pass.

  5. Awesome, work. This was one of my ideas I proposed back in 2000 when I interviewed with Pfizer regarding the picoliter dosing for the combinatorial chemistry job. I thought other than thermal issues, is a proven way to low volume dispense. I had a citizen dot matrix printer guts and head at the time and never got around to modifying with an inkjet.

      1. Thanks for the pointer. I am behind the times for sure. Here is a link I found regarding and looks like is newer: https://www.popsci.com/science/article/2013-03/how-inkjet-printers-are-helping-scientists-discover-new-drugs#page-3

        We were discussing they were already performing 1200 well plate operations and wanted me to target 4000 well plates as a vision, goal and mission.

        Reading the latest and greatest looks like there are 1535 well plates on the market also. Looks like the industry is down to 6 pico-liters dispensing also.

        Neat to have advocated, not been involved with and see now on the market 17 years later where came out 13 years later. Cool! Now to advance human tissue 3D printing and real time human body and perimeter spectrum analysis using non-destructive sensing and transmission methods in generally healthy, safe, welfare and well being ways and means.

  6. How do you avoid using the ink already in these HP print heads and still use and properly phyically register these off the shelf print heads to keep everything aligned? How many nozzle activations are they good for?

      1. “this will be one of the most interesting advances for 3D printing in recent memory” … If someone built one themselves. Anything is impressive if you add enough caveats I guess. ‘In recent memory’ must mean for someone who just got into 3D printing and doesn’t know about the Saturn Vs. I think the project is cool of course.

  7. Ok. Nice work, I admit. I do remember that there was an arduino inkjet shield that came with an inkjet cartridge and even a library.

    http://nicholasclewis.com/projects/inkshield/

    Also, if hi is building this from scratch, would it not be easier to use piezo electric cartridges? I saw in France several years ago in Rennes a a printer that could actually print semiconductor materials on glass. Worked great too. The guys made transistors of about 2-3mm2 in size. I remember they used nano material powder in a commercial available piezo-electric cartridge. About 40E a pop.

    1. The problem, as with other inkjet cartridges is controlling them. There is very little published data for consumer printheads. Piezo would be better in a few ways, but I have yet to find anyone successfully controlling anything but an HP45 printhead.

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 )

w

Connecting to %s