GMG-Epson7900 drifting

[lenicalabrese]

My proofs are drifting daily. When i compare current values on profile editor of charts printed last week and today I get a .72 delta e average of all values.It doesnt seem like much but the the proofs are visibly different and unacceptable for clients. Can I fix this-
The whole story is- We are a retouching studio. A week and a half ago my 7900 and 9900 started printing terribly different. It may have been the 100degree weather a new pump on my 7900 , corrupt gmg files-i dont know. What i did do was start from scratch. i got fresh calibration files and reiterated them using the on board spectros. I usually print to gmg's canned gracolc1 mx4s that i then reiterate further to get closer tolerances and then remove the paper tint. I remade those. got them to like a .45 delta e. But even just using the 7900 im not stable honestly most jobs have been printing ok but on the crucial one - of coarse- the pastel colors keep visibly moving. Now im printing nozzle checks on each job to make sure thats not the problem- Is the the .72 comparison of the eci2002 chart (read with isis) tell me i hit the wall? ps im using gmg semi gloss proofing paper and use epson ink- and about to loose a client/job

 

[meddington]

When you say your proofs are drifting, and your 7900 & 9900 suddenly began printing terribly different...is it BOTH devices that are printing differently, or does one appear more stable than the other? As a long time GMG/Epson user, I'd have to say that consistency has always been a strong point. How often do you recalibrate your mx3 file? Are you using the on board spectro for this, and allowing sufficient time for drying? Is your environment temperature/humidity controlled?

 

[lenicalabrese]

drifting

drifting

hi thanks for replying.
The strange part is that both printers started printing differently than the day before- not matching proofs made even a week before. On July 9th -end of day of coarse - nothing matches We just had a new cap and pump installed on the 7900. We have a really important job that we primarily ran on 79 so i kind of took the 9900 out of the loop. But the 7900 is definetly not stable after creating a new mx4 from a freshly reiterated mx3. We autocal with an onboard spectro every two days to a delta of 3. On tuesday it calibrated at 4 am and used the same calibration on wednesday but prints of the same pdfs did not match both days thats the problem still have a loss of stability after rebuilding things my question is when comparing eci charts and getting a .72 difference is that acceptable? Should the difference be visible in a full color proof? what could be wrong in the proofer or rip?

 

[meddington]

We autocal with an onboard spectro every two days to a delta of 3.

A delta E of 3 on a mx3 calibration is pretty high. Just want to make sure that's not a typo. You should be getting under 1 delta E for your printer calibrations.

my question is when comparing eci charts and getting a .72 difference is that acceptable? Should the difference be visible in a full color proof?

A .72 average delta E between two ECI charts is actually very good, but that doesn't mean that a particular hue won't look different between the two, and delta E doesn't tell the whole story. Flat tints will often show a deviation more readily than colorful imagery. The max delta E value can tell you more.

Wish I could offer more advice, but if your printer calibration is indeed yeilding a delta E of 3, I'd look very closely into this. Ensuring your using the correct mx3, the correct gamut file, etc. You might concentrate on the 9900 first, since it hasn't undergone any mechanical changes as your 7900 has

 

[Ian Mackenzie]

Sounds odd.
GMG RIPs are rock solid.
Unless there are some room climate issues, Epson printers are, too.

What paper are you using?
If you are using GMG paper (very expensive, but very good), there should be no issues.

Epson paper can be a different story. It can experience consistency issues and head strike issues with the new printers.

If you are using cheap 3rd party paper, then that could be the culprit.

 

[lenicalabrese]

im changing inks index colors with high yellow show the greatest deviation delta e 2.2 to 2.65 and its probably over 6 months also putting a dehumidifier on gmg semigloss

 

[TerryWyse]

I'm not clear on the .72 dE......is that for the MX3 calibration or is that a comparison of the MX4 profile to the data set (target values) it's supposed to be matching? I'm also perplexed about the "3 dE" that use for calibration.....I'm assuming that's a max dE value, not the average.

I would recommend setting the calibration values fairly tight and see how it behaves. For the color/CMY tolerance, I'd use .50 dE average and 2.0 dE peak. For K tolerance, use .25 avg. and 1.0 peak (K is just delta L*, not dE, so it should be held tighter in my opinion).

When you reset the MX3 calibration and did a new MX4 profile, did you create new full/limited gamut files for each? Especially on the MX3 when it starts having trouble, I always create a new full gamut that reflects the *current* printing condition of the printer rather forcing the MX3 to iterate the calibration based on an old full gamut/printing condition. If you're using the on-board spectro, use the testcharts for the gamut file that have the greatest number of patches available. If you're using an off-line spectro for the MX4, you have the option of creating a custom chart for the limited gamut file with any number of patches you like. Personally, I would use a minimum of 4,000 patches if I had the choice (if you want help with custom gamut charts, contact me off-list)

On the MX4 profile, I would not accept greater than about .40 dE average (should be closer to .20-.25 dE) and you really should not have any patches greater than about 2-3 dE, especially with a printer like the 7900 on a good media.

Regards,
Terry

 

[lenicalabrese]

.72 comes from the profile editor tool i opened the mx4( GMG's Gracol that i iterated further) - Then compared those current values from a chart printed last thursday with current values of a chart printed yesterday( did this by exporting thrsdays current values and then importing them as target values in gmg profile editor and then measured the newly printed chart as current values got an average of.72 with the highest being 2.6

When i run auto cals on the mx3 cal. file the MAX deltaE for color is set at 3 - GMG suggests 5 -Average delta e is set at .75 -GMG suggests i think 1

 

[Johu]

Hi, which size ink cartridges do you use? If you use the larger (700ml) do you shake them when you replace them and if they last a long time in the printer do you take them out and shake them every now and then?
I have found that shaking the cartridges is one of the biggest stabilizing factor with modern proofers.

 

[gcplau]

A bit of fringe info

A bit of fringe info

It seemed that everyone is talking more on the software then hardware. So, let me share a bit of my experience in terms of the hardware.
#1 the inter-instrument deviation are about 0.3~0.8 delta-E (if I didn't remember wrong the built-in spectro-proofer is oem eyeone, and it should fall in the 0.5~0.8 range)

#2 the Epson x900 we have install in a prepress site also exhibit a higher than expected drifting, and finally we have to advise client to go for a standard head cleaning every 1~1.5 days in order to maintain a stable delta-E; during loose epson cleaning for e.g. once a week, the color drifted to a visually noticeable extent.

#3 if you're using a less common and non-high quality paper such as RIP's suggested fogra-cert grade of paper, be prepared that the inkjet paper itself can cause a lot of delta-E including: the linearity of the paper along it's width, on the same roll and same batch of mfr; to put it simple lower grade microporous paper or cast coated paper can give you more than 3 delta-E already.

#4 Make sure you allow sufficient time for drying (depends on paper) during your baselin/profiling or mx3/mx4; our appointed testing lab indicate that even the very high grade of proofing paper will take at least 5 minutes (I always let dry 8~10 minutes) in order to have the ink on paper stable enough for delta-E measure.

Lastly, I would like to point out that if you are looking at stability and repeatability, we should be looking at printhead, ink, paper and spectro deviation; instead of RIP software. Unless these variables are kept well, the delta-E can fluctuate A LOT to an extent that surprise you.

No much of people look at how much delta-E are responsible by which part, and to a certain extent accepted such variance, but these factors are all adding up to the complexity of changes. No matter how good the RIP software are, their delta-E are sitting on these variance and which I would say may eat up half or even all of the ISO tolerance, if you are comparing proof to proof and day to day basis.

 

[Johu]

To gcplau

#1 the inter-instrument deviation are about 0.3~0.8 delta-E (if I didn't remember wrong the built-in spectro-proofer is oem eyeone, and it should fall in the 0.5~0.8 range)

This is not the case. It is a X-Rite device but not the EyeOne as HP got an exclusive deal to use i1 in their internal measurement device.

#2 the Epson x900 we have install in a prepress site also exhibit a higher than expected drifting, and finally we have to advise client to go for a standard head cleaning every 1~1.5 days in order to maintain a stable delta-E; during loose epson cleaning for e.g. once a week, the color drifted to a visually noticeable extent.

Can you do this automatically from the software? Some softwares have this feature built in in case you want to set it to be done automatically.

#3 if you're using a less common and non-high quality paper such as RIP's suggested fogra-cert grade of paper, be prepared that the inkjet paper itself can cause a lot of delta-E including: the linearity of the paper along it's width, on the same roll and same batch of mfr; to put it simple lower grade microporous paper or cast coated paper can give you more than 3 delta-E already.

I would find a different paper then. A proofing paper which can vary this much is really not good for the job. Why would you by a very precise software if your paper screws it up any how.

#4 I totally agree.

My experinces with proofing systems is that they do not vary this much. But if what gcplau said about the paper is true this is really not a surprise that you can see a huge variation. My experinces are that the variation including all of these factors should be around 0.5 to 0.7. More that this and there must be something wrong with the system or it's components.

 

[prepresscolor]

Color Shift!

Color Shift!

I noticed two things in your question that would concern me. 1. the 100 degree day, do you have this unit in a temperature contrlooed room? 2. the mention of "a new pump"! Was this unit reciently repaired?
If you had the unit repaired you need a new calibration. The new pump is delivering a different flow of ink now.
The 100 degree day may have done nasty things to the moisture content of your paper so the ink laydown changed.
These machines are very stable in the right environment with the proper care. But, change anything and you are looking at a re-profiling job.

 

[Bloodsaler]

I have a such experience like you.Our 7910 can`t maintain consisitency.Each time I do calibration,the result color will change a little,can`t stay the same from a visual look.especially the gray balance part.Our EPSON7800 can get a much better color look.The software we use is EFI colorproof 4.0,original ink,proof paper is easicolorEP515,measuring device is Isis,So I doubt what variation cause the outlook from both two printer can`t stay same,and what cause our 7910 color unstable~
As I know,7910 has a auto cleanning system,should we also have to clean the printhead every 1-1.5 days?