How do I control which cartridge prints on each pixel?

[ChemistryPrinter]

I am attempting to use an Epson Stylus Photo R800 to print chemicals on glass for chemistry research. Since the printer has 8 cartridges this method allows me to quickly mix 8 different chemicals in lots of combinations and proportions by simply printing gradients. However, I have quickly discovered that even when I print what seems to be a pure color (magenta), some ink from each cartridge is deposited including signficiant amounts of black and yellow ink. Is there a way for me to manually control which cartridge is used to print each pixel? For example, I would like to say that one spot should get 1.5 pL from the first cartridge and 3.0 pL from the second but nothing from he remaining cartridges. Then I would like to build up a gradient so that I can know how much of each chemical was deposited at any given point.

The problem is even more complex since this particular printer contains red and blue cartridges in addition to cyan, magenta, yellow. These seem to be redundant, so I don't understand how the printer or computer decides when to print red from the red cartridge and when to print it by mixing combinations of cyan, magenta, and yellow.

If anyone has any advice or ideas about how to gain complete control over the cartridges, please let me know.

 

[Dan@blue-optix]

As far as I'm aware this is impossible, I seem to remember a similar question popping up a few months back.... Although the principle seems like a good plan, this really was not what your poor R800 was designed for - or capable of. Unless these chemicles your using have the same chemical make up of the ink you'll find the printhead will block and cease to function after a few days.

With regards to what colour is printed and when - this depends on several factors like Paper setting, Quality setting and probably application preference etc

If you do find out how to do this then please post and let me know how, I could have a lot of fun playing with printer settings in this depth.

 

[Osage]

Yeeeeeeech----------I am just trying to get my mind around the problem you face.

My initial take is the following-----you need a machine---any machine--that will deposit eight different chemicals in varying proportions onto a media--in this case glass--for chemical
research. Then you get this brain storm-------that sounds exactly like what an eight color inkjet printer does---but minor detail---I need to control this in software to accomplish that task but the basic hardware is already there---to be had cheap in the form of mass produced inkjets of the 8 color subset.

If my take here is correct---I still see many problems.

1. An inkjet printer is designed to process ink of very specific characteristics---viscosity just being one variable--get viscosity very far wrong and it could clog nozzles--or alternately
flow too freely and inconsistantly---and when you add a 1 PL drop to the nozzle it might deliver that 1 Pl liter drop, or no drop, or three drops at some other location. Then one must ask if any of the 8 chemicals will damage the printer--and will not mix with each other before it exits the printhead.

2. If if most of the technical ink delivery questions posed above are solveable--you still have the question of software---and I am guessing that Epson or for that matter and other makers would regard their inkjet software as tip top company secrets--so I really doubt they would be forthcoming about helping you. And I also doubt existing software could be user modified to acheive your goal. Which might make your mastery of software a huge black hole in time with step one gaining an understanding of how the black box works.--and then writing your own software.

Just some of my initial thoughts on the subject---but really an interesting question---but others here may have some other thoughts---and this forum is populated by people of widely
diversed talents--maybe someone else can give you better hope.

 

[ChemistryPrinter]

I have reason to believe that the viscosity and surface tension issues are easily solvable by mixing ethylene glycol and diethylene glycol monobutyl ether into my aqueous solution. Obviously the exact concentrations of these chemicals are important, but we've already done a little work on that. Our work also suggests that our solutions won't be clogging the print head (our solutions are all aqueous like normal Epson inks). As far as I can tell the only really barier is the software. It seems that we would need to completely overhaul the driver, but this part of the problem is beyond me. As you pointed out, Epson probably won't be too helpful sharing their proprietary knowledge.

The worst case scenario is that we can print everything in gray scale and print over the glass multiple times, but clearly the color cartridges would speed up the process.

 

[d87cfv]

I think you'll find that even in Grayscale your printer will use various cartridges, not just the black one. I know that with the Canon, if you check that box all it does is print output in greyscale but it still uses all colours to get the thousands of shades of grey, not possible with just 1 ink....

Again back to the clogging - as i said the solution will have to be identical to the epson ink. There are other characteristics the ink exhibits that are not so obvious - it cools the head as the head prints. Some problems with (not-so)compatible inks are that although they print OK the head will over heat on a long run and then you'll fry the head. Also will your solutions be ok being superheated in a tube so that they rapidly expand and eject themselves out of the head - as soon as you get a few nozzels blocked (well over a thousand per colour on the r800 i think) your quantities of chemical going onto the glass will have changed.

Sorry to sound so negative...are there no machines available that are dedicated to this task?

 

[Grandad35]

ChemistryPrinter,

All printers of this class have their color information input as a combination of Red/Green/Blue (RGB) values. The printer's firmware then decides on the combination of inks to apply to the paper to produce the desired RGB color value. This combination depends on many things, including the particular colors of the inkset, the paper and the printing quality settings. AFAIK, there is no way to bypass this step and directly address each of the 8 individual sets of ink nozzles on the r800.

This is probably done on purpose because the details of selecting which inks to use and the pattern in which they are laid down can change the color that is produced. For example, printing a magenta dot on top of a yellow dot gives a different color than if a yellow dot is printed on top of a magenta dot. Printing a yellow dot beside a magenta dot produces yet another color. Just imagine the number of support requests if a printer allowed the user to drive each set of nozzles independently.

There may be a few high end inkjets that allow this flexibility, but you will then have to use a RIP (Raster Image Processor) to drive the printer.

 

[Osage]

To Chemistry Printer,

Your problem and the question of d87cfv--are there no machines dedicated to this task have continued to resonate in my miniscule mind. After some thought, a wonderful home barbequeded steak dinner, and a few beers have all conspired to reach the unmistakable conclusion that this is really a resource problem. And until we all know more about the resources you can bring to bear on the problem, we will be unable to advise you.

On one hand you may be filthy rich, or some blond bimbo with a well heeled sugar daddy, but its also possible you are a member of the lowest possible social class in society--namely a grad student at a university--perhaps striving for a degree and shamelessly using us to help you do it on the cheap.--and you have received alot of honest answers saying an inkjet printer may not be the way to go.

On a more serious note I will suggest that there well might another low cost option that might involve far less hassle---because you only need an XY positioning table and a set of eight noozles that you can control with a computer. And that the lowest cost largest size option may be with various dealers of used CNC machining equipment--what you may be well looking for would be a CNC semi-obsolete milling machine with a serious uneconomically repairable problem with the spindle--finding such a machine with an extra axis control would likely make the problem of interfacing with a PC much easier. And the software controls on such a machine would be another detail. But from the standpoint of such a dealer of used CNC equipment--having such a machine on their hands would normally make the machine unsellable--so they would likely jump at a chance to unload such a white eliphant cheap.
But flash edit update--you could also just hire the job done at a local machine shop--its not a matter of modifying the machine--its a programming problem that might let the machine shop operator make a few bucks without big expenses or wear and tear--lots of different variants of that idea.

Then you could either arrange your eight noozles to point at the same place--or have them in eight known places--and perhaps then solve your problem on the cheap---you may have to write your own software but you would already have the precise positioning software of the CNC machine usually capable of accuracy as precise as .001" or less--with GOO commands possible for rapid moves. And as someone who dabbles at C and C++ programming--that or other languages can give you powerful controls with looping of the same command giving the user a lot of control with few lines of code.

Just a suggestion there--on a less serious note--if my idea works for you---there should be a moral requirement to name your first born grandchild either Osage or Rumpelstilskin. Or failing that--some chemical help in beer brewing might suffice.

 

[shannrenn]

A new comer say Hello to ever one, just read ChemistryPrinter's topic, if i am not wrong, you need a 8 heads printer with each head print a kind of material, but you want control them by single nozzle.
1. I don't think any existing non-rip printer can do this, you may can try to make a special image which make-up by diffenert color dots.
2. The rip wide format one could control to nozzle, still they only have 4, 6, 8 colors, you may still need do a special image.
3. Even any machine above works, you have a limition of the proportions, say each head with each materil, even you use 3 bite head, the proportions could be 1:8 (if i'm right).
Osage side a good idea, you may just need make a machine lit more complex, in my opinion it is not a printer any more, it is just a machine, this may make things lot of easier.

 

[mikling]

Grandad35 points out something that is very true and is the tip of the iceberg in understanding how to engineer an inkjet printer. There are many choices in how a certain color can be mixed to a chosen shade through various patterns. The chosen pattern will determine how an image is perceived and there are "secrets" as to what produces the best results.

This article: http://www.aardenburg-imaging.com/news.47.html towards the end discusses how a grainy picture can still sharp because of various factors. I think it well worth reading as it will give some insight why ink droplet size alone does not necessarily make for a better image, but other factors are equally important in knowing how humans perceive images.