My epson4900 developed a noticable magenta colorcast, what can be the cause?

[JRSmit]

Today i noticed that prints with my Epson4900 have a mild but noticable magenta colorcast.
I double checked with other papers and their profiles and also other paper types same magenta colorcast. Checking against testprints made 3 weeks ago show that this has developed in the past 2-3 weeks.
I print using Lightroom, the printer has colormanagement set to Off.
I do not print every day, but about twice a week, and at first i thought i made an error in my image development, or that my monitor was not correctly calibrated, or that my lighting setup for evaluating prints was off.
I calibrated my monitor (Nec PA271 SV) , i checked the lighthing i use for evaluating prints, Solux 4700k bulbs with 12V powersupply, and measurements with i1Pro2 and i1DisplayPro both indicate 4600k, as to be expected and no change wrt previous measurements.

So today i printed a grey wedge (Don Hutchinson) on an A4 size paper and this definitely showed a magenta cast over the grey, most noticable in the middle grey part of the wedge. The softproof in LR does not show this cast.

So I reprofiled one of the papers, and the colorcast is not there anymore for that paper with the new profile.

Epson printers are not known for color drift, nevertheless something has changed in my printer causing it to develop this magenta colorcast, which it did not have three weeks ago. And looking at prints in the past 2 weeks it looks as if it is slowly developing.

[JRSmit]

Added 2013-08-17: In  addition to above this morning i did a complete nozzle clean and then i did some testprints and measured via Colorport with i1Pro2 in spot mode some of the steps in the wedges. I also measured testprints i made some weeks ago, the results are below.

My conclusion is that either something in Windows 7 has changed in the past weeks or the Epson4900 has changed. Note that the computer has not been connected to the internet for some months now, so no unnoticed updates took place.

Even ACPU prints show the colorcast, be it somewhat less, and assuming ACPU completely circumvents any colormanagement in Windows, i would say it is the printer.

But what is the cause?
Or is it Windows 7 with its Windows Color System (WCS) doing something that causes a double profiling?
Any help is greatly appreciated.


----
Profiling: i1 Photo Pro2, #patches=1005, spread over 2 A4 sheets, all charts printed via iProfiler

Prints made with LR4.4 on Windows 7, 64bit, image used Hutcheson 21 step-wedge(step 1=paperwhite) tiff , paper: HP Advanced Gloss. Printer is set tot Color management=Off, media is PGPP, quality=1440.
                                                               
Profile with Profilecharts made august 16:     
Step 1                                         Step 11                                    Step 21
L=97.334 a=-0.595 b=-3.919      L=51.752 a=-0.361 b=-2.586      L=3.697 a=-0.044 b=0.059
Visual check under Solux 4700 (measured 4600k): pretty neutral, perhaps a bit cool.

Profile with Profilecharts made july 19:
Step 1                                         Step 11                                    Step 21
L=96.322 a=-0.437 b=-3.717      L=50.271 a=2.835 b=-2.268      L=3.648 a=-0.137 b=-0.033
Visual check under Solux 4700 (measured 4600k): magenta cast.

Printing with ACPU or Windows Picture viewer: visual check shows a magenta cast, but not as strong as example above

Comparing iProfiler charts printed on August 16 with the one of July 19: the August 16 has on the grey patches a visible colorcast compared to the July 16 ones.


I also measured the grey 12-step wedge on the testprints made as part of the profiling process of some other papers, and on print of same image made yesterday:


Proline Metallix Gloss 245 Printed on
Patch       August 2           Augus 16
1                L=89.743           L=89.595
                  a=0.921            a=0.885
                  b=-10.67           b=-10.645
7                L=49.275           L=49.423
                  a=1.055            a=1.587
                  b=-7.226           b=-7.087
12              L=3.841             L=3.863
                  a=0.041             a=0.096
                  b=-0.969            b=-0.756
Profiling patch charts printed via i1Profiler on August 2


Proline Vibrant Gloss, testchart printed on:
Patch           July 20                August 16
1                L=96.914           L=95.262
                  a=1.082            a=1.381
                  b=-10.67           b=-9.89
7                L=52.838           L=50.937
                  a=0.1                a=2.631
                  b=-4.201           b=-5.188
12              L=4.334             L=4.151
                  a=-0.107             a=0.112
                  b=-0.486            b=-0.534
Profiling patch charts printed via i1Profiler on July 19

From the last measurements on can conclude that the biggest change somehow took place between july 20 and August 2.
-----

[JRSmit]

Can it be room temperature? We had some hot days in that period.?

[hugowolf]

Can it be room temperature? We had some hot days in that period.?

How far out of the 59-77 ºF 'print quality guarantee' temperature range where you?

You have done a nozzle check print and all the yellow and cyan nozzles are firing?

Brian A

[Scott Martin]

Are you certain that some of the light magenta nozzles we're clogged before?

[JRSmit]

How far out of the 59-77 ºF 'print quality guarantee' temperature range where you?

You have done a nozzle check print and all the yellow and cyan nozzles are firing?

Brian A

In the period of about July 20 till last week, the temperature rose to about 95F (35C) during the day, perhaps in a peak somewhat higher.
I missed the 59-77'F (15-25'C) print quality guarantee temperature range. I have found it in the manual, oops.

I print regularely the testprint of wayne fox if i do not have some image(s) to print to make sure all colors are printing.

Last friday after printing some testprints, all of a sudden the next print showed stripes typical for clogged nozzles. Nozzle test indicated LLK, VM and VLM , the normal autom. nozzle clean did not fully resolve the clogging, so i did a full clean. After that no signs of clogging in prints.

Looking at the prints of the last weeks with a loupe there are no stripes that would indicate clogged nozzles.

So it could well be that the environmental temperature was too high and caused the colorshift. Will the colorshift stay or go away after some printing?

[huguito]

I had the same problem with my 9600
The problem was cured replacing the Cyan and Ligth Cyan cartridges.
Probably was just one of them but I replace them together so is hard to say wich one
Hope it helps.

[JRSmit]

I had the same problem with my 9600
The problem was cured replacing the Cyan and Ligth Cyan cartridges.
Probably was just one of them but I replace them together so is hard to say wich one
Hope it helps.

Could be, i am expecting a new Light Cyan tomorrow. Thanks for the advice.

[JRSmit]

In the period of 20-27 July it has been very hot by Dutch standards, wih a peak on July 22-23 of around 33'C  (93'F). So in the room where my printer is placed the temp could have been like 38-45'C  (100-114'F), this is a guestimate based on experiences of inside temp versus outside temp. During those few extremely hot days i did print testprints just to prevent clogging.

[huguito]

Best of luck.
Those weird problems can be a whole lot of fun to chase around

Hugo

[Jakuli]

So how did it go with your 4900, my printer started to print with magenta cast the other day. Nozzle check is perfect, I haven't changed anything with profiles, settings, drivers etc. I went and printed some old photos with same paper, same settings and there's a noticeable magenta cast. Light cyan is about 95% full, cyan about 90%. Haven't tried to replace the cartridges yet.

[JRSmit]

hi, well what solved the problem was blue and yellow ink replacement. in my case the cartridges were quite full, and it took some time to realise it was the ink. so i profiled regularely to get rid of the cast. but it completely went away after replacing blue and yellow and several prints of course to get the new ink into the print head.

[Jakuli]

Thanks for the reply. I'm wondering now which cartridge should I try to replace, light cyan was the latest I changed and that's about 2 weeks ago. Yellow is about 15% left, and that's about 1,5 year old. As it's magenta cast I guess next logical step would be to change the cyan cartridges.

It's hard to swallow changing almost full cartridges and burning 100 euros per shot just to see which one would do the trick when there's a 4 meter long print job already in the trash.

Profiling is bit painful because I don't have an automated system but rather a Spyder print solution with a handheld spectro. I'm not sure if it's good quality despite of the +700 batches, so far I've been ok with canned profiles by Ilford and Hahnemuhle.