Latex 360 color consistency is not there

[Andriy]

Hi,

I've been having issues with both of latex 360s not printing a color continuously.
If it's 1 continuous color, it will change throughout a 30 foot run almost guaranteed.
If it's a multi-color print, the color will change enough towards the end of the roll making it impossible to run full jobs and leaving me with resorting to the solvents that have consistent color.
I have tried switching every single print-head but it keeps doing it.
Color calibrations before matching and doing the run, print head tests are all great.
Also it doesn't seem to matter on the material, anything from IJ180 to Photo Paper does the same thing.
Doing a color calibration when the color is off doesn't get me back to the original color either which leaves me with adjusting the color to match the original print now.

As nice as these printers are, I can't seem to rely on them for color which limits production greatly.
Any suggestions would be welcome.

Thanks,
Andriy

 

[Scyan]

Hi, what firmware version are you running in the printer ?

When did you update to that firmware revision ?

 

[Bly]

Our 360s seem more consistent than the 25500s.
Anyways we always print multi panel jobs with colour bars and make sure the panels print in sequence (also alternate tiles flipped) so any gradual shift isn't noticeable.

 

[ColoPrinthead]

Climb aboard and join me on the HP hate train!

 

[BIG EASY DOES IT]

Something is up. I print with my L360 today tomorrow 3 months ago. I get the same colors. There is no shift in colors.

 

[P Wagner]

Our 360s seem more consistent than the 25500s.
Anyways we always print multi panel jobs with colour bars and make sure the panels print in sequence (also alternate tiles flipped) so any gradual shift isn't noticeable.

+1 to these recommendations. The Latex 300 series are indeed supposed to produce more consistent color (with 95% of colors <2 dE 2000 along a 20 meter long continuous print run).

 

[Morkel]

Hi,

I've been having issues with both of latex 360s not printing a color continuously.
If it's 1 continuous color, it will change throughout a 30 foot run almost guaranteed.
If it's a multi-color print, the color will change enough towards the end of the roll making it impossible to run full jobs and leaving me with resorting to the solvents that have consistent color.
I have tried switching every single print-head but it keeps doing it.
Color calibrations before matching and doing the run, print head tests are all great.
Also it doesn't seem to matter on the material, anything from IJ180 to Photo Paper does the same thing.
Doing a color calibration when the color is off doesn't get me back to the original color either which leaves me with adjusting the color to match the original print now.

As nice as these printers are, I can't seem to rely on them for color which limits production greatly.
Any suggestions would be welcome.

Thanks,
Andriy

Hey Andriy and everyone else here. New to this site but been printing for 10+ years and designing for 15+

I joined up because I am having a similar but not identical issue. We've had our 360 for close to a year now, but just recently the neutral greys have been way out of wack. Up until then the greys had been very neutral, by far the most reliable in printing true greyscale. They were printing so pink that I assumed a cyan or light cyan printhead had died, except there was no indication of that in the CMYKcm gutter strip. Nozzle test showed everything was firing. Did a colour calibration. Did a full "optimise print quality" which included clean, alignment, calibration. Even changed every single print head and calibrated all over again. Still pink after every attempt. Checked file - same exact file that had previously printed perfectly. I rang our main HP guy and his comment was the standard colour calibration is merely a linearisation of the individual colours, not necessary a calibration of the mix - which flies in the face of everything we'd been told previously. His suggestion was to go through the process of making a new ICC profile. So I did that. It is a fair bit better, but still no where near perfect. His other suggestion would be to turn on pure colours (where if the CMYK colour only uses one channel then it doesn't mix in others to soften or compensate) but that's not suitable for this job as some full colour images fade in to the grey background etc.

This has only happened over a period of roughly one week, which makes me think it may be related to Andriy's dilemma. For the record I had been running firmware 00_06_02.10 but have just today manually updated to .12 as per discussions I found elsewhere. Note though that both the .10 and .12 update came with the "...this firmware contains print quality improvements... you'll need to recalibrate after a few prints...etc", however we'd had the .10 update active for a good 3-4 weeks and had printed anywhere from 30 - 50 full rolls of different media since then (regularly doing colour calibrations) so I don't think that all of a sudden these "quality improvements" would force the grey balance out after that long, certainly not as suddenly as it has.

Has anyone else ran in to a similar situation? From rummaging around some other knowledge bases it appears that I'm not the first one to have this problem, but no one has come up with a solution (or even elaborated on a potential cause) from what I can find.

 

[jasonx]

Could the internal spectro be dirty or not operating correctly? If its not reading correctly then nothing with be perfect. Just an idea and haven't experienced this problem as of yet.

 

[player]

Can you roll back to previous firmwares?

 

[Morkel]

Could the internal spectro be dirty or not operating correctly? If its not reading correctly then nothing with be perfect. Just an idea and haven't experienced this problem as of yet.

It would certainly make sense, especially as it came on so sudden. I'll put it to the tech and see what he says.

PS. Just remembered though that the colours changed progressively through one of these batches of grey prints. If the spectrometer was out it would only change the colours after a calibration, which obviously I did not do mid-print and it was all printed as one 30-odd metre batch.

 

[greysquirrel]

If your firmware is current...ink and heads new...(I hope you agitate the ink cartridge before installing) I would look at two other possibilities...verify your electricity. And Imead at the breaker box, while printing. 220 is supposed to be more stable, but if the power to your printers fuctuates while heating/maintaining them, etc..then your color will absolutely shift.

HP did have a run of bad ink...if you have multiple sets on your shelf...you may have some bad ink...HP should replace...call the 800HPINVENT and they can tell you if your ink is an issue and swap out for you...

but my guess is power is spiking...not enough to trip a sensor or fuse...just enough to f with the heads and ink temps...

 

[P Wagner]

It would certainly make sense, especially as it came on so sudden. I'll put it to the tech and see what he says.

PS. Just remembered though that the colours changed progressively through one of these batches of grey prints. If the spectrometer was out it would only change the colours after a calibration, which obviously I did not do mid-print and it was all printed as one 30-odd metre batch.

Neutrals and near-neutrals are the colours that are most likely to experience color shifts, as opposed to more saturated tones. The fact that the color shift occurred progressively through one long print run would tend to point to an issue with the printer, and the printheads themselves are the most likely culprit for this type of mid-print colour shift in neutrals. Other potential issues could be ink starvation on one or more colors (Cyan/ lt cyan or Yellow)

Another thing that can be done to attempt to isolate the issue would be to build a CMYK-based file that consists of various near neutral color patches, and print them two times (I have attached a file that I use for this purpose.) The first would be to print using the normal RIP settings that you would use for production print runs, typically using color correction. The second would be the same file with all color correction disabled (color management off) in the RIP. Focus on the colours of the various neutral patches on each. If you have a spectro, you could measure them (spot reads), but if there is a significant difference between the two, it should be quite visible to the naked eye.

If the problem is printer-centric (ie printheads or bad ink), the colors should be non-neutral (tinted pinkish or greenish) in each print. If one is neutral and the other is not. then other factors are in likely in play.

Paul

 

[Morkel]

Neutrals and near-neutrals are the colours that are most likely to experience color shifts, as opposed to more saturated tones. The fact that the color shift occurred progressively through one long print run would tend to point to an issue with the printer, and the printheads themselves are the most likely culprit for this type of mid-print colour shift in neutrals. Other potential issues could be ink starvation on one or more colors (Cyan/ lt cyan or Yellow)

Another thing that can be done to attempt to isolate the issue would be to build a CMYK-based file that consists of various near neutral color patches, and print them two times (I have attached a file that I use for this purpose.) The first would be to print using the normal RIP settings that you would use for production print runs, typically using color correction. The second would be the same file with all color correction disabled (color management off) in the RIP. Focus on the colours of the various neutral patches on each. If you have a spectro, you could measure them (spot reads), but if there is a significant difference between the two, it should be quite visible to the naked eye.

If the problem is printer-centric (ie printheads or bad ink), the colors should be non-neutral (tinted pinkish or greenish) in each print. If one is neutral and the other is not. then other factors are in likely in play.

Paul

Thanks for the help Paul. I don't think it's anything to do with ink starvation or damaged/malfunctioning print heads as I replaced them all and the colours are still inaccurate (even after new colour calibrations etc). As in, if I restart the print process it doesn't go back to how it was at the start of the first print run, it still prints the off colour that it did at the end. Like it's accuracy is just getting worse and worse. It's like the printer, after 9 months of being perfect, just decided to go "I'm going to drain what's left of your sanity" and is just toying with me.

To make matters worse, at the end of last year the colours were all a pinky hue. Now it's all over the place. Coming back from holidays this morning, the first thing I did was throw some mono on the printer, print a greyscale test with the printer as-is (ie, with the colour calibration as it was last year). They're slightly purple. Next I reset the calibration to factory, just to clear it out and return it to a base point, and printed that result. Still purply. Then did a new colour calibration. Still purply.

From there, I loaded up the polymeric that was giving me the most grief last year, the one that showed up the most pink. Still a bit pink. Reset to factory - and it's then gone waaay blue. New colour calibration - and still really blue. I don't even??!?!?!?!? I thought the point of doing regular colour calibrations was so that different medias and ink/head/environment variables were compensated for and the end result was comparable? I should note too that these two medias I have deliberately paired up as my go-to Mono & Poly options as they've always printed extremely similarly. And the two medias were actually created from each other - I originally made the Poly one from a generic Polymeric SAV media that came with the machine, then when that was perfect I copied it and made the Mono version.

For another comparison, I then went and printed the same test on the Poly again, but this time with ICC profiles off, so that the greyscale areas would only use black ink (I know that it's not necessarily a neutral-grey-based black, but it's pretty good).

I have attached a pic of these different tests for comparison - apologies for the bad lighting and angle, in hindsight they didn't need to be so big. But anyway, at the top you'll see the best neutral result with the ICC profiles off. The three tests on the right are on the Mono, the three on the left are the Poly, with notes showing which ones are the "as per 2015", the reset to factory, and the newly calibrated ones.

Here is a shot of just the darkest areas of the 3 different results (Poly profiles ON, Poly profiles OFF, Mono profiles ON). Note that the different angle shows just how blue the poly result is (left) compared to one with just black ink (middle).

And thanks for that file - I have printed it, with the results below. As you can see, when the ICC profiles are on, the neutral colours are still the slightly blue colour of the polymeric results, but are very close on the top & bottom row. When the ICC profiles are off, however, you can see that the neutral colours at the bottom are even bluer. Does this qualify as your "other factors are likely in play" scenario?

 

[P Wagner]

OK, based on the results of that testing using my file, I think that we can draw some conclusions:

1. With color correction off, the lower row of near neutral midtone colours has a bit of a color cast toward cool (blue). The reality is that the CMYK build value in the file is C50 M40 Y40 K0 in the fourth swatch from the right (lower row). The second swatch from the right on the top row is C0 M0 Y0 K50. You can open the file in AI and see all of the colour build values for yourself if you like.

So with Color Management Off, you should get a slight blue cast on the bottom neutrals but not on the top row neutrals. Once color management is properly enabled, the colour cast should be minimized and much more neutral grays being produced on both rows, and that is fairly consistent with what you are seeing here.

So based on that, my first suspicion is that the linearization data that you are using is not quite correct. When you build a new media profile on the printer, it is important that you start without any baseline linearization that the printer is working off of. HP refers to this as the Color Calibration for the media. When starting to build an entirely new media profile, the Color Calibration should report as N/A (not applicable) and not something else, such as OK or Obsolete. Once you go through the process of building the new Media Profile (Media Preset) in it's entirety, you should be starting with new baseline linearization data that has been defined for that media. If you 'clone' from a different media, you are taking the old linearization data as the baseline, and not starting from scratch.

While the printer is in a correct linear state, your grays should print neutral without a colour cast, as long as color management is used. When things drift (change a printhead for example) a quick Color Calibration on the media should bring it back into that baseline state and get colours largely back to where you started.

In my testing on cast and calendared medias, all neutral tones returned to within 3 dE00 after performing the color calibration, with most under 2, and that's not too shabby for about 10 minutes of machine time and about one linear foot of media. Happy printing!

Paul

 

[Morkel]

OK, based on the results of that testing using my file, I think that we can draw some conclusions:

1. With color correction off, the lower row of near neutral midtone colours has a bit of a color cast toward cool (blue). The reality is that the CMYK build value in the file is C50 M40 Y40 K0 in the fourth swatch from the right (lower row). The second swatch from the right on the top row is C0 M0 Y0 K50. You can open the file in AI and see all of the colour build values for yourself if you like.

So with Color Management Off, you should get a slight blue cast on the bottom neutrals but not on the top row neutrals. Once color management is properly enabled, the colour cast should be minimized and much more neutral grays being produced on both rows, and that is fairly consistent with what you are seeing here.

So based on that, my first suspicion is that the linearization data that you are using is not quite correct. When you build a new media profile on the printer, it is important that you start without any baseline linearization that the printer is working off of. HP refers to this as the Color Calibration for the media. When starting to build an entirely new media profile, the Color Calibration should report as N/A (not applicable) and not something else, such as OK or Obsolete. Once you go through the process of building the new Media Profile (Media Preset) in it's entirety, you should be starting with new baseline linearization data that has been defined for that media. If you 'clone' from a different media, you are taking the old linearization data as the baseline, and not starting from scratch.

While the printer is in a correct linear state, your grays should print neutral without a colour cast, as long as color management is used. When things drift (change a printhead for example) a quick Color Calibration on the media should bring it back into that baseline state and get colours largely back to where you started.

In my testing on cast and calendared medias, all neutral tones returned to within 3 dE00 after performing the color calibration, with most under 2, and that's not too shabby for about 10 minutes of machine time and about one linear foot of media. Happy printing!

Paul

Thanks Paul,

The bolded is exactly what we have done. When I say "generic", I mean I created a new Media from scratch, using the "+ New Self Adhesive Vinyl" and then selecting "Generic Self Adhesive Vinyl" (apologies, Polymeric was what I named the new media but the generic starting point was simply SAV). From there we went through the heat settings / ink % test (eg, 100%, 110%, 120%), then did the Colour Calibration, then made the ICC profile.

I did though copy this as the basis of the Monomeric media, but the colours were almost identical anyway and when I did a new Colour Calibration they were pretty much a perfect match. These were the first two medias that I made and, like I said, they'd worked perfectly for 9 months or so. Whenever the greys were just slightly out, a new Colour Calibration would have them spot on. It was only in December of last year that they started to go bad.

In your opinion, is it worthwhile me creating a brand new media again, putting in the same temp/speed/ink%/etc settings, and seeing if that new media prints correctly and holds its values?

 

[P Wagner]

Thanks Paul,

The bolded is exactly what we have done. When I say "generic", I mean I created a new Media from scratch, using the "+ New Self Adhesive Vinyl" and then selecting "Generic Self Adhesive Vinyl" (apologies, Polymeric was what I named the new media but the generic starting point was simply SAV). From there we went through the heat settings / ink % test (eg, 100%, 110%, 120%), then did the Colour Calibration, then made the ICC profile.

I did though copy this as the basis of the Monomeric media, but the colours were almost identical anyway and when I did a new Colour Calibration they were pretty much a perfect match. These were the first two medias that I made and, like I said, they'd worked perfectly for 9 months or so. Whenever the greys were just slightly out, a new Colour Calibration would have them spot on. It was only in December of last year that they started to go bad.

In your opinion, is it worthwhile me creating a brand new media again, putting in the same temp/speed/ink%/etc settings, and seeing if that new media prints correctly and holds its values?

Yes, at this point, my recommendation is that you start a new media preset (media profile) from a new baseline starting point (New SA Vinyl) with no existing linearization data (color calibration = N/A). If you want to use the same ink load and pass count as the existing preset, that is fine.

Paul

 

[Morkel]

Yes, at this point, my recommendation is that you start a new media preset (media profile) from a new baseline starting point (New SA Vinyl) with no existing linearization data (color calibration = N/A). If you want to use the same ink load and pass count as the existing preset, that is fine.

Paul

Done. And the results of making a new media is awesome. The photo below does not do it justice, but the new media's greys are almost spot on to the pure-black grey, if anything just slightly cooler which is what I'd prefer. Still very neutral. And using your test print, the mixed CMYK grey is also very balanced with the profiles on.

So my next question is, how come the media preset has ended up like this? It used to be perfect. In fact, this new media is exactly how the old one always was. I thought Colour Calibrations were supposed to always bring it back to a base level. So why has that base level changed???

While it looks like it's fixed the issue for now, my concern is that (1), I'm going to have to do this for every media, and (2) repeat this process every 6 or so months.

Appreciate the help again Paul.

 

[P Wagner]

Done. And the results of making a new media is awesome. The photo below does not do it justice, but the new media's greys are almost spot on to the pure-black grey, if anything just slightly cooler which is what I'd prefer. Still very neutral. And using your test print, the mixed CMYK grey is also very balanced with the profiles on.

So my next question is, how come the media preset has ended up like this? It used to be perfect. In fact, this new media is exactly how the old one always was. I thought Colour Calibrations were supposed to always bring it back to a base level. So why has that base level changed???

While it looks like it's fixed the issue for now, my concern is that (1), I'm going to have to do this for every media, and (2) repeat this process every 6 or so months.

Appreciate the help again Paul.

I am pleased to see that the new media preset (media profile) has gotten you back to where you started, and would like to stay going forward. The reality is that color consistency over time is a difficult thing to achieve, due to a host of variables that are present. These variables include:

- Environmental Considerations (temperature and humidity)
- Printer Considerations (printheads, nozzle outs etc.)
- File-based Considerations (working spaces, spot color recognition etc)
- RIP-based Considerations (RI, Spot Color Replacements, Filters etc)

If you are looking for consistency over time, it is imperative that process control measures are put into practice.

For environmental issues, make sure that the pritner is located in a controllable environment, where temperature and humidity can be held in a reasonable range.

Printer issues tend to center on media consistency (roll to roll) and printhead performance.

File issues require that users have an awareness of embedded working spaces and workflow settings in the RIP.

Next, a device to measure color is an important part of process control. The measurement device must be able to capture LAB spot color values. Finally, a standard reference file should be consistently used. Something like an IDEAlliance ISO 12647-7 2009 Reference Strip, or the subset of colors in my file (posted previously). I built that file based on the most significantly variable colors in the IDEAlliance document.

The process is straightforward. First, build the media profile (media preset) and verify that it is acceptable to you, by printing a reference file. Then, print the 23-patch test chart, and hold it for future reference. You can measure that file and capture LAB values for each, if you like. At a future point in time, reprint the file - the best way to do this is to use the same data that was printed for the initial reference. In some RIPs, including Onyx, that print data file is typically retained in the Buffered jobs queue. Drag and drop that previously ripped file back into the active queue w/o re-processing. This ensures that the same data is being used for the subsequent reprints.

If you measure the new printed swatch files, the difference between the original color, and the new color, can be defined. When the amount of change (dE2000) reaches a previously-defined threshold value of change, that triggers the new Color Calibration (linearization). That might happen in a matter of days or weeks, or perhaps months, but this approach provides a data-driven process for keeping color constant over time.

Paul

 

[Morkel]

I am pleased to see that the new media preset (media profile) has gotten you back to where you started, and would like to stay going forward. The reality is that color consistency over time is a difficult thing to achieve, due to a host of variables that are present. These variables include:

- Environmental Considerations (temperature and humidity)
- Printer Considerations) printheads etc.
- File-based Considerations
- RIP-based Considerations

If you are looking for consistency over time, it is imperative that process control measures are put into practice.

For environmental issues, make sure that the pritner is located in a controllable environment, where temperature and humidity can be held in a reasonable range.

Printer issues tend to center on media consistency (roll to roll) and printhead performance.

File issues require that users have an awareness of embedded working spaces and workflow settings in the RIP.

Next, a device to measure color is an important part of process control. The measurement device must be able to capture LAB spot color values. Finally, a standard reference file should be consistently used. Something like an IDEAlliance ISO 12647-7 2009 Reference Strip, or the subset of colors in my file (posted previously). I built that file based on the most significantly variable colors in the IDEAlliance document.

The process is straightforward. First, build the media profile (media preset) and verify that it is acceptable to you, by printing a reference file. Then, print the 23-patch test chart, and hold it for future reference. You can measure that file and capture LAB values for each, if you like. At a future point in time, reprint the file - the best way to do this is to use the same data that was printed for the initial reference. In some RIPs, including Onyx, that print data file is typically retained in the Buffered jobs queue. Drag and drop that previously ripped file back into the active queue w/o re-processing. This ensures that the same data is being used for the subsequent reprints.

If you measure the new printed swatch files, the difference between the original color, and the new color, can be defined. When the amount of change (dE2000) reaches a previously-defined threshold value of change, that triggers the new Color Calibration (linearization). That might happen in a matter of days or weeks, or perhaps months, but this approach provides a data-driven process for keeping color constant over time.

Paul

While we don't have a device to measure the output, every other step that you've mentioned is already part of our procedure. The printer is in a fully enclosed room with air-con on 24/7. File creation, colour management & print workflow is standard across-the-board and tightly controlled.

Without anything to measure colour discrepancies, we've been doing calibrations as per the printer's recommendations (eg, when it says it's obsolete), but also whenever we've had a job where the colour is critical. For example a reprint or new work for an existing client that we'd done colour matching for), whenever there are large or noticeable areas of greys or neutral colours, whenever we do colour tests to match a specific colour, etc.

What troubles me is that doing all of this relies on the Colour Calibrations doing their job - ie, measuring the current print output and adjusting it so that what we throw in to the rip looks the same when it spits out of the printer. Why would it stop doing that after X amount of time? Surely whatever base point the printer aims for stays the same, permanently? It must, otherwise this new media that I made wouldn't look correct, right? So what base point is the old media "aiming" for when it does the calibrations?

 

[P Wagner]

While we don't have a device to measure the output, every other step that you've mentioned is already part of our procedure. The printer is in a fully enclosed room with air-con on 24/7. File creation, colour management & print workflow is standard across-the-board and tightly controlled.

Without anything to measure colour discrepancies, we've been doing calibrations as per the printer's recommendations (eg, when it says it's obsolete), but also whenever we've had a job where the colour is critical. For example a reprint or new work for an existing client that we'd done colour matching for), whenever there are large or noticeable areas of greys or neutral colours, whenever we do colour tests to match a specific colour, etc.

What troubles me is that doing all of this relies on the Colour Calibrations doing their job - ie, measuring the current print output and adjusting it so that what we throw in to the rip looks the same when it spits out of the printer. Why would it stop doing that after X amount of time? Surely whatever base point the printer aims for stays the same, permanently? It must, otherwise this new media that I made wouldn't look correct, right? So what base point is the old media "aiming" for when it does the calibrations?

I'm not sure that I am able to offer a definitive answer to your question re aim points. With a halftone large-format printer, the RIP is typically in charge of the color production. In that case, the user has access to a host of settings that are handled by the RIP. That is not the case with the HP Latex 300 series, when building on-board media presets. These settings include individual channel ink restrictions, ink splits, linearization measurements, and total area coverage ink values. When using the on-board profile creation, all of this is the responsibility of the printer, and the machine does not share the data with the user, to my knowledge. This has been done to make the process of building the custom preset faster and easier, and there is no question that they have accomplished these goals on the Latex 360 and 370 printers.