Image quality "banding" problem with Epilog 32EX 120W

[Michael Perrault]

I am experiencing a random image quality problem I'll refer to as banding. As in the lower half of the attached picture, parts of my image will show a decrease in apparent laser power. I'm engraving 6 bottles at a time laid out side-by-side on an Epilog 32EX 120W system.
Despite the 9 years of employment at Epilog, the tech I've been speaking with cannot provide any solid information about this symptom. He's suggesting that it might be the laser tube, the RF module, or after I mentioned the possiblility of a degrading power supply, that too.
179648
I remember reading this article
http://www.optoiq.com/index/photonics-technologies-applications/lfw-display/lfw-article-display/278185/articles/industrial-laser-solutions/volume-21/issue-11/features/technology-report/controlled-cooling.html
from industrial laser solutions essentially describing this problem due to heat control issues, but I've never had this problem before with my air cooled unit and the environment of my laser is a constant 60F.
The unit has two power supplies. I've tried running tests using only one or the other but I can't seem to identify the problem based on that so they're both working at least to some degree.

People have talked about components breaking down, but what are the symptoms of a power supply, RF board, or laser tube that's going the way of the dingo?

[Mike Null]

Michael

Welcome to SMC.

There have been many threads on this topic so I recommend doing a search. There will be enough reading for an evening.

[Frank Corker]

Well it's definitely banding. Does it occur when you only do one? Does it occur when you use a similar type of graphic? There are quite a few variables as there always are with laser work. I have seen this happen when wet paper is being used and that type of banding appears when the paper has dried out, if that is what you are using, add dishwash liquid or shampoo over the paper. I always find that it keeps the paper wet and gives a nice frosty appearance, but that's the way I do it and there are a lot here who insist they see no difference (even though there is).

[Dan Hintz]

that's the way I do it and there are a lot here who insist they see no difference (even though there is).
;) No room for interpretation, right Frank? :p

[Michael Hunter]

;) No room for interpretation, right Frank? :p

Nah... He wrote it all in english

[Scott Shepherd]

Okay, that's not the type of banding I expected to see, which is a good thing. My first thought (and it could be wrong) is it is not a laser problem at all. I believe it may be the pattern you get when you do different width stuff. It happens on some materials. If you look at the width of the engraving at the top, and then where the banding seems to start, it seems to start where the moose gets narrower (shortly after that). I believe it's the machine switching from long strokes to short ones. I doesn't make that switch on the first line of the narrower part, so that's why it appears a little lower than the actually narrowing of the graphic.

My suggestion would be to run it again, but this time, but vertical lines, maybe .03" thick (a rectangle .030" wide and however tall the graphic is), just off the left and right side of the work. So you'll have a narrow vertical band, then the piece, then another band to the right. This will keep the stroke it makes the exact same and it will eliminate the problem, IF that is what's causing it. That's the first thing I would try.

Of course, I could be completely wrong and it could be a mechanical problem :) But that was the very first thing that hit me when I saw it. I've fought that issue on Epilog lasers as well as Universals. Universal has an option called "Frame Rasters" that does that automatically without having to put the lines down the side. It shows up in some materials far more than others.

Give it a try if you don't mind and report back.

[Frank Corker]

I like your thinking Steve. I agree you might be on the right track. It's made me think too on graphics, especially ai, not eps, the graphic itself is made up of gently altering bands of colour (shades of grey). I don't know how they would sort your problem, but if is the one I mentioned a solid fill and retrace might illiminate the problem. Not as clever as your theory, but then again, I'm not as clever as you!

[Scott Shepherd]

Not as clever as your theory, but then again, I'm not as clever as you!

I'll trade my clever for your artistic talent :)

It's not that I'm clever to come up with the theory, it was a real world problem we had on many jobs on our laser and somewhere along the line, someone suggested it, we tried it, and it resolved that issue.

Even today, from time to time, I have to use the "Frame Rasters" mode to keep it from happening. It's usually on what I would call "Delicate" materials. Clear acrylic, glass, etc. Things that show all the details from everything.

[Dee Gallo]

Okay, Steve, now I wonder: if your theory is correct, which I have seen myself in the past to be true, why didn't it produce yet another banding pattern where the plants started on the bottom left - they would have made a wider pass similar to making a vertical line on the side, no?

Could it have something to do with the shape of the container? It seems to change shape right there where the pattern starts. I wonder if the focus goes out a bit there?

Just asking, dee

[Scott Shepherd]

I don't know Dee, it might not be the issue at all. It's possible that the pattern shows up when the motion goes from wide to narrow, but it might not show up the same when it goes from narrow to wide. I have no idea if that's what it is, but it's where I would start. I'd draw that lines and try it and see if it resolves it. If it does, problem solved. If not, I have no idea :) I only have one idea a day, so this has pretty much tapped me out :)

[Richard Rumancik]

I have had similar problems when rastering larger text. I don't know if it was "banding" but the effect was similar to what you see with the moose image.

Since my text was a True-Type font I could easily generate the vector outline by converting to curves. If I recall correctly, I created a black outline a few pixels wide. Then filled with grey say 80%. Then I took this image and converted to greyscale (tif) bitmap, exported to PhotoPaint, then converted back to b/w using Stuki. You need to play with the grey setting and the intensity etc. By doing this, I had a solid black outline for the text, and a "stippled" black instead of solid black. The solid black outline is important or you will generate a fuzzy edge when converting the greyscale image to b/w.

My impression was that even if the grey was at a high level it still seemed to even out the raster lines. Whether it will get rid of "banding" I can't really say, but it tends to even out the look of the image in my opinion. Even if you don't have a banding problem, I think the technique can improve appearance of large lasered pockets or areas.

So if you have (or can generate) a vector image of the moose it might be worth a try on some scrap glass.

[Dee Gallo]

Steve, you crack me up!

Richard, I like the solution you propose - for general purposes whenever there is a large black space - it would certainly give a nicer look and less apt to look "machined". I will adopt this as standard procedure!

cheers, dee

[Richard Rumancik]

Dee, if you try it be sure to report back what you find. I haven't done exhaustive tests so I'd be interested if anyone else tries it out. It might not work on all lasers - I suppose it also really depends on what is causing the issue, too. Perhaps by forcing the laser to keep tuning on-and-off through the pass rather than turning off completely it affects the operation of the tube in some way. Don't know exactly but it helped in my situation. It does generate a more "random" and softer looking pattern in the background. I also experiemented with g/s pattern fills rather than solid fills and this might have some potential too.

[Michael Perrault]

I've used this image for a few years and have been engraving these same bottles with the image over this period without trouble until this last month. I'll engrave as many as 120 bottles with the same image in a day in lots of 6 at a time oriented horizontally across the engraving table. The banding will appear in different widths over different parts of the image during the same day. Some images are somewhat lighter altogether.
Excuse the blurry images from my attempt to take more pictures, but attached is another example of "banding" on a test bottle. Again, the bands seem to be random, and I have a hard time relating it to heat or strain on the laser from engraving larger solid areas.

179866179867

I've also spoken with the laser manufacturer, Coherent, and the Access Business solutions and feel I can rule out the Power Supplies. They feel the RF module would turn off completely if it did not receive enough power (not increase and decrease power which we may be seeing here).

It sounds like replacing the flying optics is the next cheapest thing to rule out but it doesn't seem like optics would change fast enough to go back and forth causing the distinct bands.

I am still trying to determine if these may be symptoms of a weakening tube or RF module.

[Dan Hintz]

Michael,

It's difficult to tell from the pics... does it look like you're getting a double/ghost image? It's noticeable on the antlers, particularly, but it could be an artifact of a reflection form the back of the glass... you'll need to verify.

[Scott Shepherd]

I agree with Dan on the ghost. Loose pulley, loose mirror or loose lens, or worn belt. If you're lasering a lot, it might be time to swap out the belt and see if that helps.

[Michael Perrault]

The additional pictures are to show another example of the same issue but illustrating the intermittent nature of the problem. Ghosting is not an issue and the quality of the outline shown in the first posted pic is typical. This and the other images I'm engraving on the bottles always have crisp outlines despite this banding issue. In fact, the outlines are particularly distinct as the first pulse after my 120W laser recovers between image elements is always much stronger. Epilog laser systems do not have "first pulse suppression" like more sophisticated power controlled machines. The first few discharges of the laser are particularly strong as the excess built-up energy is bled off.

I'm sorry to hear about your luck with the your system, Dan. If a laser tube typically ends it's life by a gradual reduction in power or by intermittent (all on/ all off) operation like yours, than I probably do not have that issue. Would the RF module be the next candidate to consider? At least my laser has not gone out completely but I need to figure out where to head given this symptom.

Lens, mirrors have been cleaned, alignment checked. Soap and wet paper towel used for engraving is consistent among good and bad resultant engravings...

[Richard Rumancik]

Michael, does this model have a separate RF board (outside the laser enclosure)? If so there should be some troubleshooting info available from the manufacturer. I assume it would be a lot cheaper and faster to repair the RF board than sending the whole tube back.

I have a hunch that the laser output has become unstable. This could be caused by bad RF board, or something that has gone intermittent inside the tube itself. And of course, if the power supply became intermittent that could make the laser output change. Easiest thing you is to check is to make sure that the power supply is stable - if you know a tech with a scope maybe they could watch the power supply voltage waveform while you are lasering something and see if the power supply shows anything unusual at the point the image goes bad. If the voltage is fluctuating that would not be good.

[Shawn Conklin]

I would try rastering a big solid black square (5"x5" +) onto a piece of regular acrylic. When we started having banding issues (on a 3 month old machine) it was painfully obvious and consistent on acrylic but other materials were on/off. It was the laser because Epilog replaced under warranty (in like 2 days, which was nice).

Something I also didn't see mentioned was to try running the machine on a UPS or generator to rule out power supply company issues ...

[Dee Gallo]

Dee, if you try it be sure to report back what you find. I haven't done exhaustive tests so I'd be interested if anyone else tries it out. It might not work on all lasers - I suppose it also really depends on what is causing the issue, too. Perhaps by forcing the laser to keep tuning on-and-off through the pass rather than turning off completely it affects the operation of the tube in some way. Don't know exactly but it helped in my situation. It does generate a more "random" and softer looking pattern in the background. I also experiemented with g/s pattern fills rather than solid fills and this might have some potential too.


This is an update- I tried the 80% fill in a large area on cast acrylic and it did engrave with a smoother surface. I did not even use an outline either. I don't have banding problems, but I can usually see the laser's path, however faintly, on acrylic if there is a big black area. With the 80% black fill, there was no such problem, just a smooth even burn, very similar to sand blasting. YaY and thanks Richard for prompting me to try that. It makes edgelit lights look so much better!

cheers, dee

[Richard Rumancik]

Glad it worked for you - wasn't sure if it was just an anomaly with my laser or had general application. Sometimes there isn't really enough time for "research" to do careful comparisons. The reason I was using the outline was because I was marking text, and noticed that although the nasty raster lines disappeared, the outline got a bit fuzzy due to missing pixels on the outline edge. So I suppose whether you need to outline or not will depend on the application.

I also see the regular parallel lasered lines you are talking about - and the more in-focus that you are, the more you see . . . I find that the random pattern looks better to the eye.

[Dee Gallo]

Rich,

I just got an order for 6 lights and did them with the new technique. I'm very happy with the smoothness of the engraved area as compared to solid black. They all came out the same, no lines anywhere. I added a photo to my other post... don't laugh at my shaky photography. I need to get a camera stand.

cheers, dee

[Robert Walters]

This is an update- I tried the 80% fill in a large area on cast acrylic and it did engrave with a smoother surface. I did not even use an outline either.

Happy New Year!

[Robert Walters]

Michael,

If you grabbed a flat pieces of glass instead of bottles, does the randomized banding occur?

Is the thickness of the bottles consistent?

You said you do 120/day, have you PM'ed you machine lately?
Clean, lube, check/tighten loose screws, belts, gears, pulleys, etc?

[Michael Perrault]

As a conclusion to the "banding" problem, it was determined to be the laser itself after both the laser and RF unit were tested on the bench at Epilog.
What I experienced did not appear to be a typical problem or symptom. Although tech support varied greatly between techs at Epilog, I thank Brian for finally seeing this through. The 2006 inspected Coherent 120W laser tested at 79W so was replaced because of this (Cost $4200). The first replacement came with a blemished mirror and scratched lens but the second replacement was refurbished as expected.
Without knowing the exact reason the output power seemed to be intermittently low and high, replacing the laser resolved the problem and I have not had the problem again since.
There was a disagreement as to whether a weak or intermittent power supply (two on this Legend 32EX) could have caused this issue but I did not have to continue down that path.
Although I've heard several times "The laser always stays sharp" unlike other machine tool bits, I take this as a reminder it is not free. You can buy a lot of bits for $4,200.