FNO - Fading in black/cyan/"photo" dye based inks

[JV]

Please see the revised Post #11

JV

 

[Grandad35]

JV,

That was quite a lot of work - thanks for your efforts. Your tests make it clear that the i9900 is the best Canon printer for this purpose.

I printed 5 copies of the "extended nozzle check" and ran some tests to see if the small color patches from the i9900 were large enough to reliably measure the color and to see how much inherent variation there is in using this printer to generate the test patches. After some experimentation to develop a procedure to reliably center the measuring head on the color sample, I measured dE values on each of the 7 colors for the 5 copies in the range of 1.5 or less, with most values being less than 1.0. As a point of reference, measuring the same color without moving the head usually results in dE values of 0.0, and removing and repositioning the head on the same sample usually results in dE values of 0.6 or less.

I am currently of the opinion that the only reliable way for someone like us (without a specialized device to print small ink samples) to generate samples accurate enough to allow color comparisons is to run all of the tests on a single printer, using a single box of paper. I have a spare print head for my i9900 that is perfectly functional, but it has a loose heat sink that makes an annoying noise while printing. If there is sufficient interest, I am willing to print test sheets with various sets of ink carts using this head (so that I don't risk my "perfect" print head). I have full 8 color inksets for Canon OEM, refilled Formulabs and "New Inkgrabber" (G&G). You have 8 colors of BIJC#2 , Fotofreek has 6 color inksets in MIS refilled carts and prefilled Arrow and Hobbicolors carts, and Rob has 4 colors of several other inks. If there is enough interest to run these tests, it has to be a "one-time" thing (every ink to be tested will be printed on the same day, then the current print head will be reinstalled), as there is too much risk in switching print heads/inksets just to test an additional unknown inkset at a future date.

Which inksets should be tested? The final color information is stored in a ".cxf" file that can be downloaded into a free program available from GretagMacBeth that can be used to display/compare the color values and spectrum of each sample.

 

[JV]

Grandad,

I can mail you full eight color inksets of the BIJC#1, BIJC#2, and OA100 carts if you email me your address.

The OA100 carts were recommended by Frank Trotti in http://forums.dpreview.com/forums/readflat.asp?forum=1003&message=14559744 The distributor claimed that OA100 carts used Formulabs ink but I could not substantiate it with your custom printer profile.

Without a printer profile BIJC#1 and OA100 carts are somewhat better than BIJC#2.

It would also be desirable to test Image Specialists ink since it is available in 2 oz bottles ($6/bottle) labled with IS ink numbers from http://www.printerfillingstation.com/index.htm I could contribute $ if there is intrest.

JV

 

[Grandad35]

JV,

Sorry for the delay in responding. I've been in contact with Rob and Fotofreek, and the consensus is that the original goal (developing a test that would make it easy to compare the colors of an unknown inkset to known inksets) isn't really practical unless we can find an easy way for anyone to generate a set of color swatches. Thanks for your generous offer, but we won't be pursuing this project unless we can find a solution to this problem.

Thanks again for your efforts.

 

[canonfodder]

GRANDAD35,

Since this posting was some time ago, have you or any others improved your method of laying down sample swatches of ink?

I have been thinking about methods that might standardize the thickness of the deposited sample. I understand the desirability of having a method that does not use a printer and that will work with minimal ink consumption.

If you have arrived at a satisfactory method, I would like to hear about it. If better ways are needed, I will continue my thinking and trials.

 

[Grandad35]

canonfodder,

Sorry to say, but we never came up with a solution to this problem since last year's posts. The closest that we got were the single ink test patches printed on a special test routine on an i8500/i9900, but these are the only printers that we are aware of that have this capability. Since none of us who have these printers was willing to be the "central guinea pig for ink testing", it never went any further.

If you can come up with a method that anyone can use to "print" their particular inks in a repeatable manner (duplicate tests by several individuals on the same ink giving a delta-E of 1.5 or less) at a density approaching that of an inkjet printer, I would be willing to measure the various ink samples with my spectrophotometer and to post the results. However, when we were working on this project we did look around for such a test device/method - we came up empty.

 

[canonfodder]

O.K. I am thinking. And I will be testing.

 

[canonfodder]

Grandad35

I will be trying to produce ink samples on some form of paper. There are some immediate questions.

What surface quality is best for your spectrophotometer? Gloss, matte, or flat?

Would it matter if the surface rippled, like ~~~~~~~~~~~~ ?

How large should a sample be? Such as the diameter of a circle or dimension of a square, etc.

Should I take these communications off forum and to email? I certainly don't mind everyone seeing what is said, but I don't know if there is any problem with clogging up the forum with stuff that isn't newsworthy, as yet.

 

[Grandad35]

canonfodder,

1. The spectro can handle any of these surfaces, as long as the paper's surface "grain" is much smaller than the spot size (about 2-4 mm diameter) so that any local differences due to the surface are averaged over the spot size.
2. Whatever paper you choose, it should be something that is readily available and consistent over most of the globe. The reason for this is allow anyone to contribute to the database, once you work out your techniques.
3. A large scale ripple (say 0.5 mm or larger) could cause problems in that the color readings could be different depending on which spot is used to take the readings.
4. Even though the spectro's spot is small, I would like at least a 25mm x 25mm sample. This allows me to take numerous readings around the sample to find an average reading.
5. Discussions such as this on the technical aspects should be public - people who aren't interested will just ignore this thread. Discussions on where to send the samples, etc. should obviously be private.

 

[Nifty]

Discussions such as this on the technical aspects should be public - people who aren't interested will just ignore this thread. Discussions on where to send the samples, etc. should obviously be private.

Agreed! You never know what information contained in a discussion may spark the ideas of someone who can provide assistance.