Banding on solid black fills.

[Brian R Cain]

I was told by a ULS technical guy who knew his stuff that banding on anodised aluminium is caused by the beam reflecting back and causing interfernce. Most noticeable when engraving large areas. The suggestion was made that to overcome this, slightly angle the workpiece. Probably a few sheets of paper under one side would be enough. I have to say I didn't follow this up at the time so can't speak with any authority about it, however in a different instance there was a machine we had that was exclusively being used for vector marking steel using HPDFO where it developed a random fault and occasionally during a job would stop marking. I eventually tracked the problem down to the lens coating being damaged by spatter off the material. A new lens cured the problem and I can only conclued that reflections from the material would occasionally be sufficiently out of phase to reduce the beam power.

As someone said earlier, this issue crops up from time to time without an adequate resolution, so it might be worth trying what the ULS guy suggested.

[Dan Hintz]

I was told by a ULS technical guy who knew his stuff that banding on anodised aluminium is caused by the beam reflecting back and causing interfernce. Most noticeable when engraving large areas. The suggestion was made that to overcome this, slightly angle the workpiece. Probably a few sheets of paper under one side would be enough. I have to say I didn't follow this up at the time so can't speak with any authority about it, however in a different instance there was a machine we had that was exclusively being used for vector marking steel using HPDFO where it developed a random fault and occasionally during a job would stop marking. I eventually tracked the problem down to the lens coating being damaged by spatter off the material. A new lens cured the problem and I can only conclued that reflections from the material would occasionally be sufficiently out of phase to reduce the beam power.

As someone said earlier, this issue crops up from time to time without an adequate resolution, so it might be worth trying what the ULS guy suggested.

While it sounds all "scientificy", there so much more going on that this would be lost in the noise (and I have lost my faith in anyone working for a laser manufacturer that calls themselves a technical person). Consider this... a few pieces of paper on one end would be a few hundred micrometers. I guarantee the tables on a 24x36 machine are not ground to micrometer-level accuracy to make those pieces of paper amount to a hill of beans. If it mattered, the interference would then vary many many times over a large piece (considering the wavelength of a CO2 laser is 10 micrometers and you're varying the height from 0 at one end to hundreds of micrometers at the other), so you get waves of interference over the entire item. Then try and keep the items moving all over the place within that same micrometer-level accuracy... mirrors, lenses, axes, etc. You can practically see them shudder around as the carriage zips back and forth.

No, I can't believe interference has anything whatsoever to do with problems like this...

[Scott Shepherd]

Any luck Dan? Just as a point of reference, the correction on rastering increases the power at the ends of engraving to compensate for the slow down and ramp up of the stroke. That would mean a value of 20 would put more power at the left and right edge, and leave less power in the middle, which is where you are having the issue (the middle not engraving as thoroughly as the ends). That's why I'm telling you to change that value to zero. It essentially over powers the edges a little to compensate for the ramp up and down in speed. It's not needed often and I'd guess it'd almost never be something you'd use on things like anodized aluminum, but I suppose some circumstance could show up that would need it, but in general, that's not needed on anodized aluminum.

Having a value in that field can make things like smaller text look fat on the ends of travel and the text in the center will look sharp and crisp.

It's a feature that's rarely used for me and it's one of those features you shouldn't use unless you're having issues that require it to correct the issue.

[Brian R Cain]

While it sounds all "scientificy", there so much more going on that this would be lost in the noise (and I have lost my faith in anyone working for a laser manufacturer that calls themselves a technical person). Consider this... a few pieces of paper on one end would be a few hundred micrometers. I guarantee the tables on a 24x36 machine are not ground to micrometer-level accuracy to make those pieces of paper amount to a hill of beans. If it mattered, the interference would then vary many many times over a large piece (considering the wavelength of a CO2 laser is 10 micrometers and you're varying the height from 0 at one end to hundreds of micrometers at the other), so you get waves of interference over the entire item. Then try and keep the items moving all over the place within that same micrometer-level accuracy... mirrors, lenses, axes, etc. You can practically see them shudder around as the carriage zips back and forth.

No, I can't believe interference has anything whatsoever to do with problems like this...

I hear where you're coming from, Dan, but something is physically causing it and it isn't unique to any particular laser. The regularity of the banding reminds me of the sound you get when plucking a pair of strings on a 12 string guitar when they're not quite in tune. I'm refering to the lower-pitched wow-wow-wow noise they make on top of the tone the stings make. If you looked at the signal from these on an osciloscope, what do you think you'd see? Something similar to the banding?

Whilst I appreciate that the technical guys laser manufacturers employ will sometimes be evasive of telling the truth about issues with the machines, and it isn't something unique to the world of lasers, the guy I spoke to was as straight as a die and would answer my questions honestly or say he frankly didn't know just then, but would find out. Never had him fail me.

His suggestion of angling the work to avoid the issue made sense. The bit about the paper was mine, hoping to avoid discussions about focus and so on. The bottom line though is there has to be a rational explanation to the matter and this was the best I've heard.

[Richard Rumancik]

In the initial post, Dan Kozakewycz called the problem he was having as "vertical banding". When we talk about banding it usually refers to horizontal streaks which appear lighter. I don't know if there is an official name for vertical streaks but I suppose "vertical banding" is as good as any for the moment - but people should not confuse the two phenomena.

Brian, was the ULS rep talking about vertical banding or horizontal banding when he made those comments? If someone says banding without qualification I would automatically think "horizontal".

Dan K, I would suggest that you look at mechanical reasons for the vertical streaks. That would include belt tension, bearings, debris on the rails, debris stuck in toothed pulleys or belts, a bad x motor bearing, loose pulley, loose optic . . . maybe others?

It is possible that you might be able to hide some of the problem (if you can't find a mechanical source) by changing your solid black fill to something slightly less e.g. 90% black or 95% black. This will cause a few black specs to appear in the solid white areas of the aluminum, but that probably would not be an issue if it reduces the vertical streaks. I can't say it is the ultimate solution as I suspect a mechanical issue, but it might alleviate the immediate problem.

[Dan Kozakewycz]

Hi Richard.

The machine is virtually brand new, it's probably got less than 10 hours use on it since the Trotec engineer was here setting it all up, and most of that time has been on anodised aluminium so very low output in terms of debris etc.

The banding still appears even on heavily dithered light grey fills!

At the moment my best fix when it is an issue is to increase DPI and defocus the lens to overlap the lines. Still some slight banding noticeable but only close up.

[Scott Shepherd]

The banding still appears even on heavily dithered light grey fills!

At the moment my best fix when it is an issue is to increase DPI and defocus the lens to overlap the lines. Still some slight banding noticeable but only close up.

Dan, have you tried it with the correction changed? Also, please post your settings. We can't help you without knowing that information.

[Dan Kozakewycz]

I've been to busy have a play with the correction and check the banding again the last few days.

At the moment, I've kept the correction the same at 20 as I know it produces nice results on my own products and being mostly line artwork, any banding is not an issue.

When I get some time I will try changing the correction and see if that fixes it.

I'll make sure I post settings on any further questions. I think (without being at the laser) the settings for this were in the region of 80-100P/90-100S and 1000 PPI. I was varying the power and speed and they didn't have much effect.

[Scott Shepherd]

Note- you already posted the settings, I forgot you did that.

I'd be in the 500 DPI, 70-80% power (for your machine), 100% speed, 500 PPI, Correction 0, High Quality On.

[Dan Kozakewycz]

OK just got these two, both 65P/100S, 1000PPI 600DPI and HQ, but the whiter one is 0 correction and the golder one with banding is with correction at 20.



I'm pleased with the difference, thanks for the tip on that one!

I was under the assumption that correction only affected the accelerating and decelerating parts of the motion. Perhaps it is changing it's rate of velocity over a longer distance than I am expecting?

[Dan Hintz]

I hear where you're coming from, Dan, but something is physically causing it and it isn't unique to any particular laser. The regularity of the banding reminds me of the sound you get when plucking a pair of strings on a 12 string guitar when they're not quite in tune. I'm refering to the lower-pitched wow-wow-wow noise they make on top of the tone the stings make. If you looked at the signal from these on an osciloscope, what do you think you'd see? Something similar to the banding?

Whilst I appreciate that the technical guys laser manufacturers employ will sometimes be evasive of telling the truth about issues with the machines, and it isn't something unique to the world of lasers, the guy I spoke to was as straight as a die and would answer my questions honestly or say he frankly didn't know just then, but would find out. Never had him fail me.

His suggestion of angling the work to avoid the issue made sense. The bit about the paper was mine, hoping to avoid discussions about focus and so on. The bottom line though is there has to be a rational explanation to the matter and this was the best I've heard.

Banding can be easily created... engrave a 501-pixel wide image at into a box 1" wide using 500dpi. That'll create one gradual band across the entire image (though you'd likely never see it). Now start increasing the image width pixel by pixel, engraving at 500dpi... every time you hit a specific resonant frequency, you'll see very distinct banding, increasing in number. the bands will fade in and out as you get closer/farther away from that resonance. There's no mystery behind creating banding, the pain is figuring out what is creating the resonance... usually it's a function of some software resonance (like mentioned above or hitting the "beat frequency" of the laser's pulse versus the pixel data) or hardware resonance (movement speed makes a mirror wobble, etc.).

The disappearance of the banding when Dan set the Correction parameter to 0 says it was a software resonance (or software/firmware, if one wants to get particular).

[Scott Shepherd]

OK just got these two, both 65P/100S, 1000PPI 600DPI and HQ, but the whiter one is 0 correction and the golder one with banding is with correction at 20.

I'm pleased with the difference, thanks for the tip on that one!

I was under the assumption that correction only affected the accelerating and decelerating parts of the motion. Perhaps it is changing it's rate of velocity over a longer distance than I am expecting?

Dan, I don't think they like my explanation of what it does, but in my opinion, the correction doesn't change the velocity, it just bumps up the power at the ends of the engraving. What you were seeing is a variation of the power in a small area, where the power is up on the ends, down in the middle, which causes the banding issue. It's really irritating that they have a default value of 10 in there. That default is the default when the correction only controlled corner speed adjustments and it's right for the default to be 10. However, in v 10.2, they change it to have two different functions. One function for vector, one for raster.

I took some time and dialed in a material I was engraving a lot of small text on. I think it was 150-200 of them on a sheet and the text was small. I got it all dialed in and the text was really sharp. Then I did a job like it but a slightly different material, so I wanted to use the same settings. I did it manually but didn't pick up on the fact that the correction was at 10. I engraved a whole sheet and noticed it looked decent, but not as sharp as the last job. It racked my brain for a while and then I noticed it was on 10. I changed it to zero and it got better.

It's just a really bad setting to have set at 10 as default.

Glad you got the issue resolved to your satisfaction.

[Michael Reilly]

I don't know about a change to correction, but my understanding of it was that it was a minimum power threshold. Whether 20 meant it could drop no more than 20% from the setting you are engraving at, or whether 20% was the lowest it could go, I never could determine. It exists because some materials have a threshold at which the material shows a mark. Glass is one, if it doesn't get a minimum amount of power, it won't register at all. But as the laser speeds up and slows down at the beginning and end of each pass, it varies the power to compensate for the change in speed. In those materials, this can lead to losing the edges of your image as the power drops below the threshold. But it can have the opposite effect too, if 20% shows more of a mark than 10%, then when it slows to a point where it would be 10%, it's overheating that area. It sounds like that's what was happening here.

You can also get patterns when using stochastic screening like Stucki, Floyd-steinberg, etc. I don't think those were options for Trotec though. But since they're mathematical random pattern generators, the patterns will repeat. On my Epilog I used to use Stucki all the time, but when I complained about the patterning to Epilog support, they explained that it shouldn't be used unless it's a photo or something with a lot of tonal variation, otherwise such patterns will appear. When I was trained at Vytek, they explained that even in photos, things like faces have similar tones and can show patterning and the solution is to add some noise to the image to break up the pattern (in PhotoGrav.)

Another thing is what Dan was talking about with dot overlap. While it seems like higher resolution is always better, some materials don't benefit from it and it can actually hurt it. If you're trying to do a photo that is composed of dots, the image is produced from the pattern of dots vs no dots. The average spot size for a laser with a 2-2.5" lens is 0.005". If we divide that into 1", that tells us that at 200 dpi, we'd have a row of dots that touch but don't overlap each other. In that case, we could tell if a dot was omitted. But if we double that to 400dpi, now we've packed twice as many dots in and that means they're overlapping 50% with their neighbor. Now if we omit a dot, it's really difficult to tell. So engraving fine text at high resolutions can be beneficial because those close overlapping dots can better fill out the solid letter forms, but if we're looking for detail in a photograph, higher resolution isn't necessarily better.

[Dave Sheldrake]

I was told by a ULS technical guy who knew his stuff that banding on anodised aluminium is caused by the beam reflecting back and causing interfernce.

What the guy is on about is Interposed Phase Stack Cancellation, *technically possible* but VERY hard to do even when working with lab equipment let alone a basically junky laser machine, to try and transpose phases in a lab has to take into account vibration from rooms several yards away when somebody shuts a door let alone when you have a moving gantry running over a workpiece that varies in thickness and distance from the focal point. You are dealing with wavelength distances that on a flat unsupported sheet would be affected by gravity let alone something like a laser machine.

My take? the guy has read the paper "Phase Cancellation Principles" by Prof R G Crooks and is spouting out what he has read in an effort to seem technically proficient to somebody that doesn't understand what they are being told. To anybody that does...it comes across as random rubbish.

Remember an "Expert" only has to know 15% more than the person they are talking to

cheers

Dave

[Dan Kozakewycz]

I think I've managed to get rid of this issue now, at least on the solid fills.

Solution was to turn down the X axis acceleration in the service settings from 9 to 4. This improved the mid section banding but increased it towards the outer edges, so a tweak the correction setting back to 15 saw that evened out. Interestingly, the correction value changed the hue of the entire area, making all of it slightly less white.

Downside is that it's slowed the engraving speed down considerably, but I guess that's a trade off that will need to be made when this level of uniformity is required, but at least solid fills now come out perfect. There does still seem to be some feint banding on photo engraving, but that can be tuned out with a slight defocus.