Thursday, June 10, 2010


Inkjet RepRap

A while ago Amberish Jaipuria did some preliminary RepRap experiments with inkjets - details are here. And before he left, Ed was always saying, "I think inkjet is the way to go."

It would certainly be nice to be able to lay down waxes, resins, conducting ink and all the rest in a RepRap with the fineness and precision that inkjet could give. We would probably want to use Epson inkjets, which use piezoelectric mechanical pulses to eject the ink drops. All the other systems boil the ink by resistive heating in a tiny chamber to eject it. The piezoelectric systems will obviously be more tolerant of funny polymers and the like, which might not take kindly to being boiled.

Recently I have been looking at continuous ink systems. These replace the normal (and horribly expensive) inkjet cartridges with a tank-fed system that holds 100 ml or more. And they're cheap - the four above cost me £13 in total including postage from these people. You get the tanks, the feed tubes and the hacked recycled cartridges for that. I say hacked, because the cartridges are chipped to report emptiness to the printer - this is how the manufacturers try to prevent you doing re-fills. These report "full" all the time, I think.

Now, with an Epson, the piezoelectric heads are not in the cartridge (it's just an ink reservoir). They are built into the printer. There's a good article about all this by Tim Hunkin here. (Note especially the bit about Epson heads clogging if you leave them unused or let airlocks get in.)

So to the $64,000 question: I have done a good bit of searching for the electronic incantations that need to be sent to the piezoelectric heads taken out of an old Epson printer, and I have drawn a blank. What's needed is the Epson equivalent of this excellent HP Inkjet book by Matt Gililland. I could prod about in the printer with a scope, of course, but it would be nice to have something authoritative.

Does anyone know what pulse-timing patterns and voltages Epson piezoelectric heads need?


Have you asked a local printer ink dispencer if they have cartridges of the style you want in stock or know where to go?
this is a neat idea, i think it would be great for small prints, but i wonder how long it would take to even get to 5mm high though.

but a good idea none the less
Wild thought.
Automotive gas or diesel injectors might be worth looking into also. Older ones work at low pressures. New high pressure direct injection uses piezoelectric and the older style are simply solenoids at the end of a pipe. There's a large hacking community there, especial in the racing community with tons of technical info. I just typeset mathematics, so the following is suspect until scrutinized.

-Air = 1.2 kg per cubic meter so 1 cc of air weighs 0.0012 grams.
-A 4 cylinder 2.0 liter auto engine has 4 500 cc cylinders.
-With the stoichiometric ratio of both gas and diesel engines range of 12 - 15:1 air:fuel
-Each 500 cc cylinder has 0.0012 * 500 = 0.6 grams of air each intake stroke.
-Using a very lean 13:1 ratio and I'm thinking that's 14 parts.
-Each injector then squirts 0.6 / 14 = (0.04285nnn) or about 0.043 grams of fuel each time.
-That gets one gram of material out roughly every 23.3 pulses.
- There's a large range of injector sizes available from tiny scooters to giant diesels. Low or no pressure gravity feed would be best.
I like the engine injector idea!

Dissidence - there are commercial 3D printers that use inkjet, so it must be possible.

Gary T in NZ and Michael (aka Ogre) e-mailed me with some very useful links:

you can find some information about epson heads at:

Volkan Sahin successfully build his own electronics.

I guess you can contact him at:

And maybe this would be very interresting for you:
Don't the commercial machines just dispense the binder via ink jets? The bulk of the material is deposited by roller, so they can build large objects in reasonable time despite the jets having a small throughput.
This looks like the fist generation ink jets i worked with, the tektronix 4-thousand-somthing model in the early eighties. It had tanks. And a technician that came roughly once a month to recalibrate the machine. Bleeding edge technology then.
I looked into driving espon heads a few years ago. Some guys were using the heads for DNA stuff and I got briefly involved.
The heads worked quite differently to what I'd previously thought.
The head has a shift register to load the pattern into and then a pulse is applied to fire all the required nozzles in one go. This mean all the drop sizes are the same. Variable drops size does not mean you can print different sizes at the same time. The pulse shape determines the size. The pulse shape varies from model to model.
It is not just on/off it - it has an envelope with peaks of maybe 40V.
I have epson CIS and chip hacks here
eddie (
If it comes down to prodding around on the chip with a scope, this guy I met at a party years ago is probably the first one to call: Bunnie Huang.

He's done all kinds of work to defeat hardware-level security features, and might be willing to contribute to RepRap pro-bono.
This comment has been removed by the author.
I like the Ideas as well! The only problem is the print time. This same technology is used in the Object line of printer ( Objets are almost the slowest build process, taking days to make a part. If you could some home Combine a plastic fill layer and print of that, it would save a lot of time. Z-corp printer use a plaster base to print on top off with the same idea. Now change the Plaster to plastic and use a solvent to glue it together and Bing!

This could possibly best the print, cheapest, and most well rounded printer if it could work right.
Yes - that was essentially the idea: use our existing fast process for bulk fill, then use the inkjet for details plus small quantities of fancy materials like silver ink.
I think the thing to do just after that or in conjunction with is find appropriate semiconductor materials and start printing oleds and fets and stuff like that. I know they use some kind of printing technology, but don't know what goes into getting the layers connected for conductors and power naturally. I think we may head towards the nano quickly, but if we thing building at the macro level is slow, wait till we print with something tiny.
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to
Posts [Atom]