Epson 9900 cannot seem to print very fine vertical lines accurately

[bmsorg]

I created a document in Photoshop that is alternating 1px vertical solid black then white lines.  And then sent this to three different Epson 9900's, none of the printers could accurately print the image.  Strange/wavy banding showed up (scan of the print, see first attachment Epson 9900 line test - scanned print.jpg)

One of the three machines was a brand new Epson 9900, just out of the box after head alignment.  This machine performed marginally better than the older machines, but still couldn't print the image accurately.  Tested various papers, platen gaps, settings, image and print resolutions, and RIPs... no combination could print the described pattern above. Including an Epson 4900.  Sent the file to a Canon IPF9100, and the machine printed it PERFECTLY.

Are the Epson machines less accurate in this sense?

You may be wondering why I printed an image like this in the first place? And, also thinking that maybe that this isn't a real world example. The reason it came up is because we were seeing very obvious effects of this issue on a client's file we were printing.  The image was a photograph that had artificial grain/noise introduced to it, but the grain pattern in the print was skewed during printing and the banding/waviness described showed up, making the print unusable.  After no luck at resolving the issue on the Epson machines, I created the test image of the vertical lines in an attempt to replicate the issue.

Has anyone else had this problem with Epson machines?

Regarding my test image, I didn't expect the printer to necessarily print the vertical lines perfectly, and expected the alternating white/black lines to maybe render as an near even gray in the print, but the bizarre waviness appeared instead.

Attachment:
#1 scan of Epson 9900 printing described file
#2 the test file I used to print these (1px solid vertical black/white alternating lines)

*If anyone is interested I can supply full resolution test files and/or the scanned print.

[jferrari]

Tested various papers, platen gaps, settings, image and print resolutions, and RIPs... no combination could print the described pattern above.

Out of curiosity were the various print resolutions you tested equal to 360 or at least a multiple of 60?     - Jim

[hugowolf]

2880 ppi with finest detail set on?

Brian A

[Farmer]

Send a 720ppi image and print it with Finest Detail and if need be, chose uni-directional printing rather than bi-directional.

[bmsorg]

Source test file I created was 300 DPI, didn't test 360 or multiple of 60.  Though client original work was a different DPI.  The banding is such an irregular pattern I'm thinking this isn't the issue.  As the same file sent exactly the same twice in a row, the banding shifts slightly, so it's seems more of an artifact of the printer. But it's an easy test and will give it a try tomorrow.

Did test the below variables, all showed the issue:

• Printer resolution at 720/1440/2880
• Finest detail on or off.
• Sent from Epson Driver, Mirage RIP, EFI express.
• Sent to 4 different 9900s, two of them brand literally out of the box machines. Head alignments done. And a 4900.
• Unidirectional and bidirectional.
• Different papers, and various platen settings.

Canon was the only machine had no problem printing it correctly.

Thanks for the ideas.

[hugowolf]

Source test file I created was 300 DPI, didn't test 360 or multiple of 60.  

The Canon is expecting input at either 300 or 600 ppi, the Epson 360 or 720 ppi, so it is certainly worthwhile trying those resolutions. If the file ends up being sent to the printer/driver in another resolution, then somewhere in the printer pipeline it going to be changed.

Qimage, a printing app (for Windows only) automatically sends the file to the printer in its requested resolution. In Lightroom, it is simply a matter of checking the print resolution box and setting it to 360 ppi.

Brian A

[Some Guy]

Curious as to why your 1 pixel test image shows up on my computer as not staggered at all when viewed at 3,600% in CS6?  Just straight lines, one pixel wide and one pixel apart and zero staggered.

I may try and make one on my computer and try it.  I've been questioning the sharpness out of my Epson as well.  I suspect there is some ink bleeding going on too with mine  and maybe in part searching for a maximum Dmax too.  Might dial back your density a lot and see if it makes any difference.

But yeah, your printed one does look sort of bad at 3,600% in that it is also showing a lot of colors mixed in as well and no whites.  I know mine mixes a bit of color with the black inks, but to what degree it affects the smugginess of the final print is what I too question.  The blottiness in your scanned one is sort of disturbing too if it is in the print and not the scanning process.

SG

[jferrari]

Curious as to why your 1 pixel test image shows up on my computer as not staggered at all when viewed at 3,600% in CS6?  Just straight lines, one pixel wide and one pixel apart and zero staggered.

That's what it is supposed to be. No mention of "staggered." The OP said:

"I created a document in Photoshop that is alternating 1px vertical solid black then white lines."

[Some Guy]

That's what it is supposed to be. No mention of "staggered." The OP said:

"I created a document in Photoshop that is alternating 1px vertical solid black then white lines."

Got it.  

I interpreted it as an alternating 'pixel' instead of an alternating 'line.'  My bad.

Pixel staggering would be an even bigger mess no doubt over the OP's staggered line.  Might be fun though.

___________________________________________________________________________________

Okay, I tried this on a couple of 3880's.  I did both the staggered pixel I made and the lines JPG in the above OP's post.

What I noticed was the OP has a very long format.  Don't know if it was resampled in the Epson driver for a better page fit, or if he printed it that long on a sheet?  On mine, it seemed to compress the lines on a 8.5x11 inch page and really doesn't show a good line, line, line.  More like a blur, blur, line, blur due to the size.  I made mine a 8x10 format and it looks fine other than some bizarre darkening near the last 3/4" of all prints from either 3880 (??).  I did them on a K3 inkset and a K7 piezo both.  The K7 is darker black than the K3, but I didn't print as a QTR software, just out of the Qimage software for all.  I can see the staggered jaggies of the staggered pixel I mode in both, although the lighter Epson K3 inkset seems sharper maybe due to lighter ink coloration against the K7.

I did the same on a Canon and the Canon does seem sharper than either Epson, just the ink seems a bit lighter and more blue too so it might appear sharper.  I see some transport individual line darkening too where it might have slipped a bit in the rollers though, but it is slight for one pixel.   No darkening at the end of the print like the Epson shows above either, other than the OP's JPG showing the irregular lines due to possible re-sampling to fit the page.  I'd redo the OP's image to fit smaller than the test page personally.

Aside, I see the Epson K3 just spit out a clogged nozzle too on about 4 colors as I just did a normal print job and it showed horizontal banding.  Damn Epson!  Why they clog during printing a job is frustrating!    Running an Auto Clean cycle now along with a 2 hour wait before re-testing the nozzle check...

SG

[Wayne Fox]

The file when sent to a Canon is set at the Canon’s internal working resolution.  When sent unmodified to an Epson it has to be resampled. You can see the effect pretty easily by having photoshop resample the file to 360 ppi, the lines will no longer all be just black and white. This is what the Epson driver is having to do as well when it resamples the file. So all kinds of aliasing/artifacting going on.

Before printing to the epson, if you change the PPI setting from 300 to 360 without resampling, you get a document which maintains the exact alternating one pixel black/one pixel white, but now at the Epsons native resolution.  If you print this file I think you’ll find it to be what you were expecting. I get a print from my 11880 with perfect alternating lines when examined  using a loupe.

[jferrari]

The file when sent to a Canon is set at the Canon’s internal working resolution.  When sent unmodified to an Epson it has to be resampled. You can see the effect pretty easily by having photoshop resample the file to 360 ppi, the lines will no longer all be just black and white. This is what the Epson driver is having to do as well when it resamples the file. So all kinds of aliasing/artifacting going on.

Before printing to the epson, if you change the PPI setting from 300 to 360 without resampling, you get a document which maintains the exact alternating one pixel black/one pixel white, but now at the Epsons native resolution.  If you print this file I think you’ll find it to be what you were expecting. I get a print from my 11880 with perfect alternating lines when examined  using a loupe.

I agree 100% with Wayne has said. Could not have said it better myself.    - Jim