[Misc] Eccentric circles when copying

[Guest]

Hi all. I am trying to copy a page that has a circle on it. However the circle does not come out the same every time. The circle was drawn with Photoshop and printed. Then that print was copied but the circles never line up.

I tried this on an HP 4325 and then on a few other non HP copiers all with the same problem

Does anyone know why? Or how to make perfect copies of circles?
Thanks
Shuelin

[blackcat4866]

Do you have a scan of the problem? =^..^=

[Shuelin]

Scan of proble,

These are drawn free hand to illistrate the problem. the solid lines indicate the orignal image. the dotted lines indicate where the copy does not line up the the orignal image.

samples.pdf

[blackcat4866]

Sorry for the delay answering. It took me a while to figure out that it's an HP Officejet 4325 all-in-one, and there are very few references in any English language literature, so it must be a non-US model. Your samples also leave much to be desired. A real scan of the image would be much more helpful.

But from what I've figured out, this is a $70 inkjet all-in-one. I suspect that it's black & white only. And I suspect that the problem has something to do with the type of media that the original is on, or the type of media that is being printed onto. It appears from your illustration that the image is stretched or distorted in length in the feed direction. This can happen when the original is being scanned, caused by wear on the document feeder rollers or excessively heavy media. This same distortion can happen in the printing process, again because of dirty rollers or excessively heavy media. The way you can determine weather its a scan issue or a print engine issue is to print an internal configuration page, and see if the image is distorted. If yes, it's happening in the print engine. If no, it's happening in the scanner.

In either case, I hope it is something that can be taken care of with cleaning or changing of media, since there are very few service parts available for inkjet models. =^..^=

[Shuelin]

I must apoligize for my typo. the make and model is an HP4345mfp.

the orignal media is plain white bond paper and is being copied onto the same. the samples are hand drawn becuase you cannot see the distorsion unless you hold up the orignal and copy to each other and up to the light.

the distorsion also changes its location depending on where the circle is on the page.

[blackcat4866]

Well, that changes my answer.
It can still be happening in scanning or printing, and you can still determine the difference by printing an internal report.
You've already ruled out media as an issue by trying 20lb bond as original and copy paper.

The 4345mfp scanner is constructed quite differently than the 4325. You'll have scanner rails to clean and lubricate with light oil. If the original was slide read, then you'll need to look at the document feeder drive.
The print engine is also considerably different. I'd look at the registration roller/clutch, and the laser cartridge (especially if its a compatible).

If you scan the original as a PDF, Adobe will allow the original to be enlarged 32x. Just circle the area you have identified as distorted. =^..^=

[Shuelin]

Attached is a scan of the copy. I have highlighted where the orignal does not match the copy. I have only highlighted the differences not where they line up.

over the last few days i have tired to do this on 2 canons, 2 xeroxs and 3 ricoh's and they all do the same thing. it is a general problem with copiers?

image1.jpg

[Hemlock]

It's your paper.

[Shuelin]

I have tried 3 different weights of paper. do you have a suggestion on what paper?

[CBMWilson]

One thing to consider that i haven't seen mentioned is a copy is not ever going to exactly reproduce the original. Try making a copy of a copy of a copy and you'll know what i mean. If this is very dimension specific i'd recommend only printing the original rather than printing then making copies. Your copy will not have the same precise dimensions as the original.

[OMD-227]

Try changing the auto exposure mode to something else (text, photo, map, copy of copy etc etc).

I vaguely remember a case like this years ago. A client had an adobe indesign document that copied with very shifted colors. It was almost like a registration fault, but it was only happening on this particular document, everything else was perfect.
A bit of research on various photoshop & indesign forums lead me to find that you can create an image within indesign at 16bit color. The copier in use had a 1bit scanner. Using auto exposure, the scanner was picking up various & extra colors within the image (as the black line of the drawing was actually a 16bit, multi-layered, multi-colored line, appearing as black to the human eye).

Changing AE to Text/Printed Photo enabled it to copy correctly.


A long shot maybe, but as soon as I read the part about photoshop and saw the sample, it rang all the bells. I bet it doesnt do that sort of copy with any other document.

[Shuelin]

CBMWilson: I dont disagree, and the copy of a copy is bad. but what is happening is random. the above is one sample If i was to copy it again, the skew is in a cmpletely different spot, without moving the orignal on the bed. It is strange.

[Shuelin]

WAZZA: all i have to say is holy crap. it worked. I just tried setting the resolution at the highest setting and then text only black and white. this was on an older officejet and it works, it is as close as I need it to be. I will try this on the laser mfp in the earlier post over the weekend.

thanks again
shawn

[OMD-227]

Nice fix bro!

I knew my tedious forum-scouring research would come in handy one day....

[blackcat4866]

Impressive! Great work wazza! I'll remember that.

I've dealt with the normal distortion you can get from any analog scanner. In my attempts to get the analog engine and scanner to perfectly register and with minimal skew, I discovered that there of areas of the image that tend to have more distortion than other areas. If you can envision a clock face on a letter portrait page, the areas at the center, radially out at 12, 3, 6, and 9, all can be registered perfectly every time. The areas radially out from the center at 1:30, 4;30, 7:30, and 10:30 experience the most distortion, but on any single machine remains consistent copy to copy.

The test pattern I created for centering, skew, voids, and registration, has a vertical line exactly centered 108mm from each edge, and a horizontal line exactly centered 139.5mm from each edge. If you can get these lines exactly centered, exactly registered, without skew, and perfectly aligned when viewed through a duplex copy, that is as close as you can get. The other areas will experience some unavoidable distortion, usually less than 1/2mm when using a fine grid original, and more exaggerated the further you get from dead center. My test patterns that had lines 10mm from each edge showed the most distortion, and were nearly impossible to get a "perfect" adjustment. Incidentally, most manufacturers write right into the specs +/- 1.5mm is normal variation of image registration. In most cases I've pulled it in within 1/4mm, many dead on.

I haven't made the same comparisons on digital scanners, but I suspect there would be similar results. Maybe one of you optics experts can explain the "why".

=^..^=