Polaris pixel shift problem

[adatsui]

We have 3 Polaris XTV-S CTP in our production. All 3 CTP have pixel shift problem, not too often but we just can not fix it.

Is this problem cause by NewsDrive ? Interface card ?

 

[gordo]

We have 3 Polaris XTV-S CTP in our production. All 3 CTP have pixel shift problem, not too often but we just can not fix it.

Is this problem cause by NewsDrive ? Interface card ?

View attachment 2839

It's hard to see from your photo, however, since it occurs in the same place and with fine lines that are inline with the rotation of the plate, what you might be seeing is what's called orthogonality correction.
When the plate rotates and is imaged by the laser, the laser moves horizontally across the width of the plate. Effectively the laser winds the image onto the plate just like a sewing thread is wound onto a spool. This means that lines that are inline with the rotation of the drum are actually at a very slight angle. Orthogonality correction causes a single pixel jump backwards once the line reaches a threshold length. This allows the device to image straight edges inline with plate rotation.
Usually this one pixel correction is masked by the Gaussian (softness) nature of the laser, and on press by the ink spreading onto and/or into the substrate making the correction invisible.

Best, gordo

 

[VladCanada]

Uuupsss, Gordon.
Your statement is correct... but not for this case.
Polaris is a flatbed (a-la capstan) CTP for newspaper needs. The spinner has a multifacet prizm and very tricky principle of compensations for non-ideal facets, based on oscillating cryctal with a mirror glued to it.
I was trained and had very good experience for the first Polaris generations (green laser) and never had a deal with violet extra-speed versions.
I do not think anybody at forum will be able to help, and the only one way is possible - call AGFA Tech support which becomes weaker and weaker nowadays
The last hope: may be Mr STSTech or 28-yr-AGFA-SE know more about XTV-S?

 

[maxon]

It doesn't look like pixel shift, is is single pixel size, have you put a microscope on it ? Can't see any jitter, doesn't seem to be spinner related. Lots of questions to be answered. Does the line replicate the image content at the beginning of plate ? Did it ever come up inside the image area, is there any image beyond that line, have you tried exposing anything to the edge ? You say it affects all three ctps (?), how many newsdrives do you have there ? Have you ruled out the plate ?

 

[gordo]

Uuupsss, Gordon.
Your statement is correct... but not for this case.

Hey Vlad, thanks for that! At least owners of drum CtP devices now know something about their imaging system that they may not have known before.

Best gordo

 

[STSTech]

Charlie,

I must agree with Maxon. The problem you are having does not appear to be pixel shifting. One question I have, is do you have a plate processor for eack plate setter, or do you have a line set up?


STSTech

 

[VladCanada]

1. Not the pixel shift but an extra-pixel generation caused by tricky optics.
2. Remedy - regular optics calibration performed by FSE or trained personell.

 

[adatsui]

Thans all for so many useful information.


I agree with VladCanada "Tech support which becomes weaker and weaker nowadays". All local support are my friends since we've computer typesetting/computer to film. I think they all do there best. They came many times and still can't fix this tricky problem.


>> have you put a microscope on it ?
Yes, Will take a photo and post it here.



>>have you put a microscope on it ?
Will take a photo and post it here.

>>Does the line replicate the image content at the beginning of plate ?
I'm going to check it now.

>> Did it ever come up inside the image area,
Yes.

>> is there any image beyond that line,
Yes.

>> have you tried exposing anything to the edge ?
Yes. control wedge, filename and some other information about that plate.

>> You say it affects all three ctps (?), how many newsdrives do you have there ?
3 NewsDrive for 3 Polaris + online processor VXP 85

>> Have you ruled out the plate ?
we just chuck it in the bin!

 

[STSTech]

Adatsui,

We got to ask ourself what is common to all three systems. There is only three things common to all.
1. Power: If it is the power, I think it would be more ramdom on the plate, and not always in the same place.
2. Plates: very possible.
3. A Work station: bad memory area or disk

When you send a job to the plate setter and you get the shifting, does the Rip have the ability to go bad to that job and do a preview of it? Does the pixel shifting show in prewiew?

Another thing we need to ask ourself, is when did this problem start. Did we install anything around the time we started to have this problem.

Hope this is some help to you.

STSTech

 

[adatsui]

STSTech ,
Thanks again. Let's ignore power issue. We've ask Afga , N94-V plate is good. The last thing to check is work station issue.

Our Rip can preview, we check the 1bit tif that send to NewsDrive 's hot folder . The 1 bit tiff always good.

We do install few other program on those workstations. May be our IT can help.

I've take some photo over microscope .
please view in full size.

 

[VladCanada]

What type of interface card you use? PelBox? PixelBurst? Agfa UPAC?

 

[28-yr-AGFA-SE]

If you are correct about laser direction in Sample_A.jpg, then there is a data shift causing the pixel shifting on the plate. Therefore, it is unlikely that all 3 NewsDrives and all 3 Polaris' have the same problem. Are you using a single RIP and sending it's output to the 3 NewsDrives? If this is the case, and the RIP's preview does not show any problem, then there is an output problem with the RIP or a communication problem between the RIP and the NewsDrives.

 

[28-yr-AGFA-SE]

If you are incorrect as to laser direction in sample_A, and it is instead rotated by 90 degrees, then Vlad is probably correct. Sample_AA seems to show a large shift of image that is directly lined up on the plate. There is an alignment issue with the spin optics in this case. You will need an SE to correctly align.

 

[28-yr-AGFA-SE]

If you are correct about laser direction in Sample_A.jpg, then there is a data shift causing the pixel shifting on the plate. Therefore, it is unlikely that all 3 NewsDrives and all 3 Polaris' have the same problem. Are you using a single RIP and sending it's output to the 3 NewsDrives? If this is the case, and the RIP's preview does not show any problem, then there is an output problem with the RIP or a communication problem between the RIP and the NewsDrives.

 

[adatsui]

may be i mislead about laser direction . please check attach picture.

we've 3 workstation , there is 1 rip + 1 NewsDrive on each workstation.

As we working on 1 bit tiff. one rip for diffent NewsDrive don't work for us , we need to transfer dot size to get a accurate dot on the plate.


each time we have pixel-shirt , we check 1 bit tiff produce by our rip. may be next time we also check NewsDrive's Export data , NewsDrive save output image here. I hope it never happan again


> What type of interface card you use? PelBox? PixelBurst? Agfa UPAC?
don't know.

 

[28-yr-AGFA-SE]

From your _sample_A_dir.jpg above the laser is really scanning vertically and the Polaris moves the plate horizontally. That means the "pixel offset" that you are seeing is in the direction of the spin optics rotation. From sample_AA above, I see that there is a corresponding clear area in the blue (ink) area where the laser is being offset causing the pixel shift. Doesn't always seem to be the case in sample_B and sample_BB. Suspect that you have 3 machines with similar, but not exactly the same, problem. As STSTech said above, there is possibly a single solution to all 3 problems. Must check anything that is common to all 3 units. This could also include temperature, humidity, AC power, etc. Less common cause could even be where the machines (Polaris') are located. They are supposed to be on a concrete floor rated at a minimum of 100 pounds per square inch. If they are on a lesser rated floor, then someone walking across the floor could, conceivably, cause physical shifts that would cause this type of error. Look really closely at all shared things between the machines and make sure that they are really having EXACTLY the same problem.

 

[maxon]

I agree with the above but I believe it must be electrical. It appears one single scanline is shifted at a time, we don't know if it contains the correct pixel data or is it garbage, the spinner itself seems stable there's no jitter or instability. VladCanada mentioned something interesting about the piezo facet correction but how can it happen on all three machines. In my humble opinion they have to look for single events, transients. Adatsui, collect several bad plates produced by each of your ctp's and compare them carefully, does the problem look exactly the same on all as 28-yr said ? Do your machines have proper grounding ? do you experience any power surges or brownouts ? Can you associate the bad plate exposure with starts/stops of offset machines, is there anything significant happening in the environment in the same time with pixel displacement ?

 

[Slammer]

If I remember correctly there was a issue with some older RIP's with the image size. It had to be dividable by eight, otherwise you got a pixel carry-over (Pixelübertrag?)