Banding/striations/unwanted lines with rotary

A postcard from rotary Heck
My chuck rotary, Lightburn and my fiber galvo laser are conspiring to give me a huge headache. If anyone can get to the bottom of this I’ll send chocolates and possibly champagne, maybe even cake if it’s a good solution.
The photo should be enough to illustrate the issue. I’ve spend five days (five complete days with only my cat for company, and that’s wearing pretty thin) changing every setting there is in Lightburn, and no matter what, there are lines on the output to rotary.


Messing with split size and overlap just moves the lines around.
I’ve tried multiple diameters and materials.
The laser is set up correctly and calibrated accurately. I think I accidentally marked a test pattern onto my cat, that’s how many test patterns I’ve marked. I think the fur will grow back.
I’ve tried different file types, vector and raster, both imported and created or converted within Lightburn.
I’ve tried hatch/no hatch, using different scan angles including 0 and 90 degrees.
Line intervals? Ptchah, I eat them for breakfast.
Power, speed, frequency? Done them to death.
Reversing the rotary direction? Been there.
Rotary motor speed settings? Acceleration times? Tried’ em all.
Dithering? Yes, I have dithered and not dithered.
Non-rotary output (on flat plane) is perfect, absolutely tickety-boo.
And yes, my caliper is accurate and I have not entered the wrong diameter.
No matter WHAT, there are ugly lines across whatever I’m marking using the rotary.
Changing things makes them closer together or further apart, but they won’t go away.
What I want more than anything - even cake - is for somebody to pipe up and make me look like a fool for typing this. As the sweet nectar of enlightenment rains down and the sun peeps out from behind its rotary cloud, I will be happy to be mocked because the problem will be solved.
By the way, I can be serious and I promise to be serious from now on, just letting off some steam.
Happy to provide whatever info is asked for if it forwards the action towards solving this extremely persistent problem.
Please help.

1 Like

Hello, @DHD1
Rotary Heck is not a fun place to be. Probably look around you’ll see several others there as well. Nice that they have a shop to buy pens and postcards.

Some stuff to provide:
Diameter as accurately as possible
Material (Looks like stainless steel?)
Screenshots:
Rotary set up window
image

Rotary Marking window
image

Cut/layer window
image

Stand back and take a photo of the whole thing also.

1 Like

I’ve tried multiple diameters and materials.

Start with one and get it dialed in

The laser is set up correctly and calibrated accurately. I think I accidentally marked a test pattern onto my cat, that’s how many test patterns I’ve marked. I think the fur will grow back.

9 lives, probably a few left.

I’ve tried different file types, vector and raster, both imported and created or converted within Lightburn.

Start with SVG

I’ve tried hatch/no hatch, using different scan angles including 0 and 90 degrees.

Parallel with the rotary’s axis.

Line intervals? Ptchah, I eat them for breakfast.

I like to make the LI an even interval of the split. I also like to make the split divide into the steps, like divide the circumference into the steps and use a multiple of that for the split, then find a LI that divides more or less evenly into the split.

Power, speed, frequency? Done them to death.

Low power multiple passes change up the split and LI

Reversing the rotary direction? Been there.

the right direction works best.

Rotary motor speed settings? Acceleration times? Tried’ em all.

With direct drive chuck probably 500, 3000, 500, 3000 is best bet. Won’t affect your problem anyway unless you are way off the Rez.

Dithering? Yes, I have dithered and not dithered.

No dither

Non-rotary output (on flat plane) is perfect, absolutely tickety-boo.

Good start

And yes, my caliper is accurate and I have not entered the wrong diameter.

Good start

No matter WHAT, there are ugly lines across whatever I’m marking using the rotary.
Changing things makes them closer together or further apart, but they won’t go away.

Infinitely far apart or infinitely close together are the 2 best results.

What I want more than anything - even cake - is for somebody to pipe up and make me look like a fool for typing this.

Doing a good job on your own.

As the sweet nectar of enlightenment rains down and the sun peeps out from behind its rotary cloud, I will be happy to be mocked because the problem will be solved.
By the way, I can be serious and I promise to be serious from now on, just letting off some steam.

No problem, good to see someone else with a sense of humor.

1 Like

Great reply, and thanks for all your suggestions. I’ve tried all of them.
I’ve got a jewellery business to run and I need to produce rings. So for now I’ve given up on rotary because it’s just producing poor results no matter what I do.
Instead, I’m engraving onto rings without using the rotary - in small patches, obviously, because I can’t do wrap-arounds. The results are actually quite pleasing to my eye. The photos are before cleanup and polish.
The material is titanium/zirconium damascus. You can see the stripes in the material behind the engraving. I’m going to flame anodise it when I’ve polished and etched it.
Sorry if I’m not engaging with your response technically, it’s just that I’ve had it with rotary for now. I’ve bought a different rotary to see if it makes any difference whatsoever but I’m not holding out a ton of hope. I’m beat.


So if you break your artwork up like that, you can run it on the rotary as “Run Whole Shapes if possible” and set your max split size and your split to just a little larger then the largest object and the rotary will only move after completing all the passes on the current object, splits occur only where there is no engraving.

I think your big issue is to do it conventionally with the rotary it would take a “Heck” of a lot of passes to get that much depth with no lines at the splits. My guess is you are overpowering to get the depth with minimum amount of passes. I messed around with a 25mm dia piece of stainless last night didn’t get anywhere near that depth with one pass but it came out pretty much line free. I sped this up 5X, normal speed it was about 3 min per rotation.

https://firearm-videos.sixguns.com/channel/Albroswift/video/63/25mm-stainless-steel-on-the-rotary

On silver I got good depth no lines pretty easy. Still about 15-20 passes so talking around 1 hr.

1 Like

Looks like you might be re-melting the ablated material into the etched pocket - I would suggest an air assist to remove that material, and a stronger vacuum if possible.

1 Like

Thanks again. It’s a real Black Dog having this problem, but you have prevented a Communication Breakdown and you may still help me find my Stairway to Heaven.

Let’s have a look.

Run whole shapes if possible

In the case of the image I’ve used in my photo, I don’t know what constitutes a shape. Does the software try to work out what the shapes are, or does the user define them? How is it likely to treat the Zoso image/vector, for example? Are letters shapes, do they need to be closed shapes if it’s a vector, etc. I’ve had a look online but haven’t found a clear explanation of the Run whole shapes if possible feature. I’ll just try it today and see what happens.

Existential question: if Lightburn treats the whole Zoso image as a shape, then do I need the rotary at all?

Depth and power (like Led Zeppelin)

Yes, I probably am using too much power. This is titanium and it lases like it machines. It can’t conduct any heat and will move around and decide to melt, anneal and then harden seemingly as it pleases. If I dial back the power I might get it to behave better. In fact I know it does because I’ve tried it.

I drive a fast car. If I kept my foot down on the loud pedal too long I wouldn’t be here writing this. The same probably applies to engraving. Too much power equals bad. I’ll turn down the power and do more passes. Like driving to the shops in a Nissan Leaf. You get there in one piece.

I think that the grade/alloy of titanium really matters. I’m using 6Al4V (Grade 5), which is way harder than pure Ti and makes for more durable jewellery, but it’s a pig to machine. That personality seems to translate to laser engraving. But I specialise in Titanium so I really need to get this right. Another fun fact is that I use titanium damascus: different alloys pattern welded together. So I can expect varying results across alloy boundaries. With any luck I can make a feature out of that and make like it was intentional all along.

Number of passes

One Titanium-specific deep engraving settings page recommends 450 passes. That’s on stainless steel at 20 watts, 900mm/s, 180 kHz. The deep Zoso ring above was 60 passes (but without the rotary). I tend to do a few passes, observe, and run some more until I’m satisfied.

The deep engraving settings I’m using come from apparently credible sources* and they are specifically for titanium, but they may may need significant modification to work with my setup.

In my Zoso ring photo, the background inside the rectangular recess is pretty messy. I think I need to turn the power down as you suggest. Probably defocus somewhat as well. Higher frequency and/or speed will dump less energy per pulse into the material and might give a smoother overall effect. I’m also going to try reversing the relief so that the letters are recessed instead of standing out. That gets rid of the rectangle, but the bottom of the letters (on the Z axis) might be messy too. Either way, I’ve no way to sand or polish the deep parts to improve the appearance post-engrave. Clean up passes have limited effect on this titanium alloy in my experience so far.

Ramp feature

I’ve experimented with the Ramp feature with this type of engraving. Although it’s intended for cutting rubber stamps, it has given some interesting effects on metal. Coupled with the galvo’s changing angle of incidence when not using the rotary, I can get a kind of hand engraved, italic pen nib effect on raised letters. So that’s kind of fun.

Dot width

Also thinking about dot width. Lightburn has an option to adjust (or compensate for) the dot width. I have looked into it, but haven’t yet joined the dots (see what I did there) when it comes to this application. It’s hard to tell which engraved lines are due to line interval and which are due to beam dot width. A dot width of 0.08mm, the LB default with my machine, is definitely visible. The beam can’t be burning lines 0.08mm thick if it can do line intervals many times smaller than that. My thinking on this is still in the stone age.

I’ll be experimenting with some or all of the above today and reporting back here on my findings.

I’m going to go and play now. Maybe today there’ll be a breakthrough. Thanks to your reply, I’ve gone from blaming the rotary and believing something is fundamentally broken (though I reserve the right to return to that mindset) to hopefulness that I can tune those lines out with some judicious fiddling with settings and lateral thinking.

  • ‘Apparently credible’ is probably an oxymoron.

Quite possibly. I’ve got a powerful fan on the work at the moment.

So you treated the zoso as one shape, did the whole thing without moving the ring “One Whole Shape One Slice” If you left, say, a 0.05mm unengraved area and then did it again, the software would burn the first Zoso “Object”, rotate, And then burn the second “Object”
Example:
https://firearm-videos.sixguns.com/channel/Albroswift/video/43/grid-run-whole-shapes-20x20-1

Example with text objects:
https://firearm-videos.sixguns.com/channel/Albroswift/video/30/run-whole-shapes

Beam size and LI are unrelated. You could have a beam as fat as a sharpie and still move it a hair width each pass.
The Dot width adjustment is for the start and stop of the dot. Should really be called dot length.
All that is mostly for engraving photos not deep engraving like you are doing.

I would not have put it that way, but I understand what you mean.

I don’t know what lens he’s using, so I have no idea about what spot size he’s dealing with.

25:45 into this video, the developer explains the purpose of dot width.

:smiley_cat:

Yep, great video.
I do some images on business sized cards, what they call dot width = the length of the dot = how long the laser is turned on for, and dot width correction is correcting how long the laser turns on for to make the dot’s length appear roughly 50%. Actual width of the dot is the width of the laser beam, can’t change that unless you change lens or de-focus. That’s my story, anyway. Probably saying the same thing just differently.

Gotcha. Makes sense now.

So, do you also see lines if the power is too high?

The more you try to ablate with each pass, the harder it is to not have lines, let’s put it that way. Like when I color a SS mug you can’t see any lines unless you get out the USB microscope. On the silver ring, it didn’t take a lot of power so 15-20 passes no lines. Kryptonite-beryllium-titanium-zirconia alloys probably a little tougher. Like you said in an earlier post, 400+ passes.

Another member is doing it with a hotkey macro that automatically changes split sizes and LI every so many passes. Down by the bottom of this thread:

1 Like

I’d love to see your settings for that nice clean ring.

The AutoHotkey hack should be a feature in LB. But I’m sure the dev time would be costly.
I’m having better results with lower power now. I’ve also gone to an extreme to eliminate lines by setting the split size to the line interval - both 0.024mm. I figured that if it’s only scribing one line per split, then this would reduce the opportunities to introduce artefacts and/or for the stepper to misbehave. Obviously it’s slow as Heck, but the results are clean. Getting to any depth, though - that feels like water dripping on a cave floor.
Anyhow, here’s where I am with it. This is after 150 passes.

Cool idea. But it still begs the question in my mind, why use a motorised, software-driven rotary at all? I don’t mind using a dividing head or a rotary table and moving it manually a few times between cuts. Yes, your method will space the cuts out nice and evenly round the ring, and automatically too, so it’s great for productivity.
There’s also the question of doing deep engraving for quite large chords across the workpiece’s circumference. It creates big flat spots. That can look OK, but it’s nice to have the option to make rings that are circular rather than hexagonal, octagonal and so on. I also have continuous full wraparound designs to engrave onto rings.
Grumbling aside, I’m learning all the time here on the forum and I appreciate the help.

Deep engraving on steel and similar takes a lot of passes. Limit it to parallel with the axis and yes, dripping water on the cave floor. If you get your splits really dialed in, you can run bands of crosshatch that blend together but it is not the dominant paradigm. Here is a 0.135mm split with 45 degree scan and crosshatch.
https://firearm-videos.sixguns.com/channel/Albroswift/video/65/crosshatch-on-rotary

And why take the ferry when you can swim across the sound… :smiling_face_with_sunglasses:

nothing special. I changed the split size each pass, from memory around 0.09mm to 0.13mm, and used the “Divide circumference by steps” method to determine precise splits. Did the same with LI, so the LI divided equally into the split, and the split divided evenly into the C. I have a 1:6 belt drive rotary so my steps = 38400, that smooths things out a little as well.