Graininess in mid tones

[tgreen]

I just replaced the printhead and ink assembly, did a linearization check using the supplied piezography quad for my selenium inkset and the p800. The scanned target was light in the shadows and dark in the highlights. I also noticed what appears to be graininess in the midtones. This is most noticeable in the LK LM and Y channels. It also only appears noticeable on Type2 paper, not on HPR. Any idea what this might be?

grain in lk, lm, y channels

test print for below crops

grain in mid tones

grain in mid tones

lots of grain

Little grain

[walkerblackwell]

It almost looks like the opposite paper side! Did you print on the other side of this specific sheet and test?

[tgreen]

When I can’t tell, I usually lick the paper to determine the sticky/coated side. This is definitely the sticky side. The 10-ch ink separation was printed on two-sided photo paper.

[walkerblackwell]

Was the voltage of the new head read into the firmware? If not, then all the nozzles are firing at the wrong times. The voltage is read in by the QR code on the outside of the printhead. T2 seems to show this issue more than other papers. I can get you out a replacement box too . . . it may be heat/cold related issue during ship as well.

-Walker

[tgreen]

Yes, the voltage was set, I just double checked. I did all the alignments after the install, as per the service manual.

Is it possible this is an over-inking issue? When looking at the 10-channel separation, the LK, LM, Y channels appear to get grainy anywhere from 65% to 100%. The curve I started with has a pretty high amount of ink in the mid tones.

K7-SEL-UHD-ConeType2.quad

[walkerblackwell]

It could be. Some P800s can pump out a lot more ink than others. We’ve had 7 of them here and 2 of them over-inked.

You can limit to 70% in all but the shadows and try that . . .

-W

[tgreen]

I’m thinking this has been a printhead installation issue. After a bit of trial and error, I found printing with bi-directional to be the culprit. I believe I have not successfully calibrated the bi-directional timing of the printhead. When printing my test page with ui-directional, the resolution is much better and the graininess appears to be far less. I’m at a loss of why the bi-directional calibration has not worked out, it is very straightforward.

[walkerblackwell]

Sorry, I was making an assumption that you were printing UniD. I never recommend BiD with Piezo at all on any printer.

-W

[tgreen]

Totally my fault about the BiD, problem solved!

New problems:

• I’m still seeing graininess on the T2 midtones.
• My shadows (0-8%) seem too dark, even after linearization.
• I’m seeing banding after linearization even with smoothing at 100% & 100%.
• I’m only getting down to 17.19 with UHD on the T2.
• I’m only getting down to 15.72 with UHD on HPR308.

Type2 Selenium Cone Quad Scanned

HPR308 Selenium Cone Quad Scanned

[tgreen]

I’ve fixed the graininess, it was indeed an overinking issue. Limiting the initial curve by 70-80% fixed this, I suspect this will also fix the shadows being too dark. Limiting also improved the dMax but I’m still not getting anywhere near what other folks are getting. T2 is coming in around 16.2L while HPR308 is around 15.4L, this is scanning a print of the ink separation in calibration mode.

[xander]

if you’re printing in cal mode, its not using the curve you’ve made. seems like you know your stuff, so maybe I’m missing what you’re doing here. what spectro are you using to read? how many patches? sorry, I know this is months old, maybe you figured this all out by now

[tgreen]

This is my understanding: I was printing the ink separation target in calibration mode to determine what the maximum density of each shade should be. This target prints 0% to 100% ink for each channel. When there is over inking, the density will become lighter as you near 100%. Knowing the best target density allowed me to limit the ink when creating a new curve. I’m reading 129 patches with a Colormunki. The colormunki appears to be reading accurately when tested against a colorchecker. While I’m not reading the dMax others are, I believe I’m making prints that rival what I’ve recently seen in the local gallery in terms of resolution and density.