Non OEM Inks - Why aren't they a "perfect match" to OEM? Can they be?

[Nifty]

Grandad and I were having a conversation offline and I thought I'd bring it to the forum for the benefit and comment of the group (family).

In a nutshell the discussion was centered around the idea that it should be possible for a 3rd party ink supplier to match Canon's ink much more closely than they presently do. If they could closely match the Canon ink's colors and print reliably, wouldn't they be able to increase their market share?

On one hand I find it hard to believe that the big ink makers don't have all the technology necessary to make a high quality / perfectly matched ink to the OEM stuff. On the other hand, I wonder, "If they have the technology, then why do all the "BIG" ink manufacturers have such different ink compositions?"

My thought is that it is either an issue of patent, cost, or both. Maybe they "can't" exactly duplicate the ink because of patents and so each manufacturer changes up the mix just enough to skirt the patent law. Another thought is that there are a few elements of the real OEM stuff that is so cost prohibitive that the other manufacturers find their own ingredients that will match "close enough" for the masses.

I just can't believe that with the resources available that these big 3rd party ink companies can't fully backwards engineer the OEM inks and/or do enough trial and error to get something that really is a perfect match. There have got to be some reasons why there aren't perfect matches and why the inks from all these companies are so different.

What do you think?

 

[hpnetserver]

What are the differences? the matter is an extremely subjective one to begin with. Perhaps the differences should be identified and agreed upon first. OEM inks are no doubt excellent but not perfect. Being different from OEM may not be bad. One of the problems I have with the OEM ink has been that I could only produce optimal colors when the cartridges are pretty full. Once the ink inside the sponge gets lower the vibrancy of the colors is significantly reduced. I had to resolve the issue by using 3rd party ink/cartridge and keep my cartridges full.

This is a great subject. Let's talk about the differences first. Then we may have a different view on the differences. Being different may be good or bad. It depends on what the difference is.

 

[CanonPhreak]

I agree, one of my issues with canon inks was that I could not get a good match on green. My greens always came out too yellow. When I swtiched to a 3rd party ink (weink/image specialists) I finally got the green I was expecting.

I also have fewer fading problems with 3rd party inks and papers. weink recommended red river polar papers with their inks which addressed the gas fading problem i was having with canon inks and pro papers.

Truthfully I dont see any greater benefits to oem inks over third party other than the fact that the oem inks are matched more closely to the default profiles provided with the printer, but given that custom profiles can be had for $50-$80, I'd rather save money on inks instead.

What are the differences? the matter is an extremely subjective one to begin with. Perhaps the differences should be identified and agreed upon first. OEM inks are no doubt excellent but not perfect. Being different from OEM may not be bad. One of the problems I have with the OEM ink has been that I could only produce optimal colors when the cartridges are pretty full. Once the ink inside the sponge gets lower the vibrancy of the colors is significantly reduced. I had to resolve the issue by using 3rd party ink/cartridge and keep my cartridges full.

This is a great subject. Let's talk about the differences first. Then we may have a different view on the differences. Being different may be good or bad. It depends on what the difference is.

 

[Nifty]

Excellent posts! I love getting a good discussion going. I doubt there will ever be a complete agreement regarding OEM and 3rd party inks as there is quite a bit of subjectivity.

This being said, I'm going to submit that the goal of 3rd party ink makers isn't to make a better ink, but to make one that will appeal to the largest consumer market. I'm going to bet that the majority of their consumers want a product that will produce the same prints as they received when they first used their printers with the OEM inks that the printer came with.

While there may be better combinations available with different inks, papers and profiling, this will appeal to a much smaller consumer market than an ink that matches exactly what the average users get from the OEM at a fraction of the cost.

 

[Kenyada]

I think the problem is grounded in the nature of invention/discovery. Ink formulations are "mousetraps," and the goal in every research & development lab is to build a better mousetrap. Sure it would be nice to precisely match the OEM ink - and to sell it at a substancial savings. But 3rd party labs are looking for the silver bullet that will "revolutionize" the industry. Besides, OEMs are constantly refining the target formulation, because they, too, are looking to invent the better mousetrap.

 

[Grandad35]

I am pleased that this subject has generated so much interest.

With regard to color casts, This image (a screenshot from a PS "Variations" dialog) shows a color chart and what it looks like when the colors are pushed in any of 6 directions. The point of interest is that colors on opposite sides of the original (Current Pick) are directly related. For example, if your image has a green cast it can be caused by too much yellow and cyan, OR it can be caused by too little magenta. The yellow/cyan relationship to green is obvious, since mixing these two colors produces green. It is not obvious that a problem with the magenta can produce a green cast. In a similar fashion, a blue cast can be caused by too little yellow and a red cast can be produced by too little cyan.

In theory, it should be possible to generate a custom printer profile to accommodate almost any reasonable ink set. The problem comes in when there is a secondary "profile" built into the printer to specifically accommodate quirks in the OEM ink set. If the OEM ink set has any strange characteristics (see below), it is easy for Canon (or Epson/Hp/Lexmark) to take that behavior into account internally in the printer by locally "warping" the printer's response to match the quirks in the ink. Where a profile can correct for "broad" corrections, it is possible to make much more localized corrections with the software inside the printer.

Suppose that a 3rd party bulk ink supplier produces a "perfect" set of inks without any quirks. Unfortunately, the printer's internal software will introduce "reverse quirks" in the colors when it corrects for the OEM ink's quirks. A custom profile can broadly correct for some of these problems, but it will never be as good as the OEM ink. The reason for this rambling explanation is to show why, practically speaking, the way to get the best color is to match the OEM ink set's quirks as closely as possible. A custom profile will then have a much easier time fine tuning the colors and give an even better result than is presently possible.

Hpnetserver - if your colors wash out as the carts empty, the carts are apparently partially blocked. This is not normal behavior. I run my (refilled) carts until they are so empty that the printer refuses to print, and I can't see any difference in the color before and after the carts are changed.

CanonPhreak - If you have a color problem with the OEM carts, this could indicate a problem in your print head. For example, if there is a restriction in your magenta or PM channel, you will get a green cast in your prints (as was discussed above).

Kenyada - While the OEMs are undoubtedly working on their formulations, they will create a huge problem if they change the colors. One of the main reasons given for using OEM inks is "consistent quality". If they changed the colors, they would have to also supply updated firmware for each printer and (probably) supply a free set of the new carts for every printer that they ever sold, as well as replacing any of the original unused carts anywhere in the supply chain, including those that people have stockpiled to avoid frequent trips to the store. Do you remember the recent turmoil caused by Formulabs' unannounced change in their magenta bulk ink? That was nothing compared to what would happen if Canon changed their BCI-6 colors. However, when they introduce a new ink set (CLI-8) they are free to make whatever changes they want with no negative impact. A desire to change the ink formulations may have been some of the driving force behind the new generation of printers/carts.

In addition to my recent posts on what I have seen on various photo papers (http://www.nifty-stuff.com/forum/viewtopic.php?pid=2938#p2938 and http://www.nifty-stuff.com/forum/viewtopic.php?pid=2943#p2943), I have been looking at various inks. I have samples of Formulabs and AIJ (IS?) bulk inks, as well as OEM and G&G (InkGrabber - March time frame) prefilled carts. Samples of each ink were painted (using a Q-Tip) onto Kirkland photo paper. Rob will be sending a similar sample with his Formulabs and IS inks, and Fotofreek will be sending a sample of his MIS inks. The plan is to look at the color spectrum for each ink and to identify which inks most closely match the OEM ink. While it is way too early to report on this (I don't have all of the samples yet), here is a quick look at a few interesting points. If you skipped the previous posts on photo papers because your eyes glazed over, you may also want to skip the rest of this one.

The image below compares two Canon/Formulabs colors - cyan and PM. The Canon charts are in black and the Formulabs are in Red. The color values for each ink are also given above and below the charts, using the CIELab color space. The CIELab color space is a master reference space that is widely used for color matching. The three values given for each color represent the following:
L = Lightness (higher values are lighter)
A = redness/greenness (higher values shift away from green and toward red)
B = yellowness/blueness (higher values shift away from blue and toward yellow)

The Lightness values indicate the "lightness/darkness" of the colors and aren't a big concern for this simple test, since the darkness is also affected by my uncontrolled painting technique. An ink formulator would have to be concerned with the lightness, but they would also have to use an applicator that precisely controlled the amount of ink that is applied to order to get meaningful results. A mismatch in the lightness of an ink would easily be corrected by a custom printer profile.

The spectrum of colors is shown along the bottom of each chart, and the height of the chart shows how much of each color is reflected by that ink. The cyan chart (on the left) shows a peak about midway between purple and blue, just where one would expect it. The greens, yellows and reds are all absorbed by these cyan inks and won't be reflected. This chart shows that these two inks match each other very well, both in the color spectrum and in the Lab values.

On the other hand, the PM chart shows that these two inks are very different. The Canon PM reflects red, orange and yellow (as expected), but it also has a strong peak in the violet range (apparently to shift the color toward green). Note that the "b" color values shown on the chart are very close, but that the "a" values are very different. The higher "a" value for the Formulabs PM ink indicates that it will have a red shift when compared to the Canon PM ink.

Interestingly, the PM in the G&G prefilled carts matches the Canon PM more closely than either Formulabs or AIJ. This type of information will be presented when all of the samples are received and tested.

This preliminary information was included to answer Hpnetserver's "What are the differences?" question.

 

[alchemist]

Let's address the issue of why an exact match is difficult to come by. Normally the formulation is patented, and specifically the dye or pigment dispersion is claimed. Yes I can use another dye mixture to match the colors normally within 1.5 delta E. The problem is the salts normally used to stop intercolor bleeding are also claimed, so the dye set may have to be modified to avoid infringement to get the same performance, but now I have a larger delta E. Exact substitutes can sometimes be found, but they might not have the same water fastness as the oem.

The oem will also normally claim the solvents and humectants used in the formulation and almost any substitutes that will also work well in the printhead. They are allowed by our US patent department to claim large ranges of concentrations of the solvents that make it more challanging to engineer around.
Then even if you manage to work around all "valid claims" you may still get a nasty letter from the oem legal department. Just this week at World Expo, several HP and Epson engineers and legal types were seen walking around. I had two meetings with compatible cartridge manufacturers, Asian and European cancelled due to them not wanting to be served for US infringement litigation. I know the feeling since in the past I have been involved with most of them.

Good news is that I just did a very preliminary test on some darn good Epson pigmented aftermarket inks. As soon as I finish the test I will share the results. The color match is near perfect on Epson matte photo paper.

Latest news from the show was that the oems have scared off some very good quality after market compatible manufacturers from selling in the US. The pacific rim boys keep on violating the patents and selling here. Some of their inks are near exact matches, give great results but no major distributor will touch them for fear of the oem comming after them.

Alchemist

 

[JV]

Alchemist,

I found U.S. Patent 6,695,444 "Multicolor dye set with high light-fastness and high water-fastness" by searching in http://www.uspto.gov/index.html using your terminology "InkJet dye set". The patent includes a detailed description of the invented ink and very interesting reading.

JV

 

[Grandad35]

Alchemist,

I will be making several posts in the next few weeks about ink formulations and (what I think are) some areas for improvement by bulk ink suppliers. This is a classic case of "A little knowledge can be dangerous", so please jump in and comment when I go astray or if you have better solutions. I am not afraid to make a fool of myself, and I will be proving it with some of my suggestions. Don't be afraid to set me straight - it's all part of the learning process.

A question for you. The Lightness value measured by the spectro is dependent on getting a uniform ink coating on the test swatch. Is there a simple way to apply the ink in a uniform and repeatable manner to minimize this difference? At present, I am taking 5 readings from each color, then choosing the values for the median color that seems to give the best overall average of each of the L/a/b values. I am getting "Delta E" values in the range of 1.5-2 when I carefully paint several swatches with the same ink on the same paper, but I have no idea what the differences might be when I test samples prepared by others. Any ideas on how to minimize this sample preparation variation?

 

[inkjet-RD]

OK popping in for some comments,

The OEM manufactures pay HUGE money in designing unique Dye and pigment bases that are then used to define their next ink set. This new structure needs govt approval and might be a year or more in testing.

The NON -OEM ink manufactures have to blend existing materials to compare to the latest inks from the OEM's.

The NON- OEM ink manufactures also have to get ink to work in an arena that the OEM never plays, AFTERMARKET! So actually the aftermarket inks need to be more forgiving to please the customer base while still meeting the OEM target.

The testing methods for ink and paper in always one of test and control. YOu can not use "aged" controls to compare to fresh generated targets. While stored results in a color measurement device is close nothing beats a frest test, current ink, known paper, calibrated tester, and conditions.

I always say Ink at a price point is CHEEP ink.