Hello,
I just finished tuning in the scan offset adjustment for my Boss laser but when I engrave a logo or words, I’m getting random spots that are being engraved outside of the desired area. It looks like the laser is firing a dot here and there before or after it’s supposed to as it rasters. It’s not a problem all over the image but does show up here and there, especially on tall thinner lines (in the y direction). Any advice on what is causing this or how to fix it?
Could you upload your original image of the LBRN so we can take a look?
This may be discouraging, but I haven’t seen anything to the contrary since pulling it together:
The comments cover many of the first thoughts you may have, so they’re worth plowing through.
There’s a link to the background info on my blog:
Bottom line: it’s how cheap HV laser power supplies work.
Thank you for the great info! Very detailed, and a bit advanced for me with the power supply inner workings but reading through a few times it makes sense.
What I had gathered from the links you sent as far as fixing the issue is concerned is that upping the power of the laser would help. For the work I’m doing engraving on coated metals, I’m able to increased the power basically as high as I want and it still looks good. It’s not like I’m engraving wood where upping the power will cause it to engrave deeper than desired or anything like that. I get a bit of a wider engraving line at high powers, but otherwise no visible change.
So with that, I tried several settings upping the power from the original 30% from my first photo up to 70% power. Attached is a photo of the results with the power increased to 70%. The top area with the thinner vertical line that showed the dots previously looks good, but the text at the bottom of the logo still shows quite a few dots. I wasn’t able to get rid of them at any power between 35% and 70%.
My basic settings other than power are:
Speed 400mm/s
Line interval 0.065mm
Single Pass
Air assist on, but very low pressure.
105 Watt tube
With upping the power not really making the dot issue go away (though a bit better) is there anything anyone has tried that has showed further improvement? Is that something that would be feasible to swap out?
If the power supply is the route cause of the issue, even on a fairly nice new machine, could replacing the power supply with a high end one fix this?
May me?
It’s nice when we know what’s the x and y orientation…
Do you need that small of an interval? I’d be surprised if you lasers kerf is anywhere near that small.
AFAICT, it’s inherent in the power supply design and probably related to tolerance stacking on one side or the other of nominal. Some supplies will be better, some will be worse, and there is no way to tell which you’ll get.
I have three HV supplies sporting three different names on their cases, with identical PCBs inside. IMO, all the commonly available supplies come from the same ZYE factory (or are counterfeits claiming that ancestry), so the branding you might rely on means nothing.
Although HV supplies are (comparatively) cheap & readily available, I’m reluctant to fund an investigation … 
Unrelated: the ragged edges on those letters suggest you should go through the Scanning Offset Adjustment measurements to crisp them up. That won’t fix the speckles, but it’ll make the good results look better.
At this power and speed, the laser’s kerf is 185 microns. From earlier testing, there was a distinct drop in quality when I increased the line interval to 0.100mm. I imagine the overlap of the lines may help smooth out the edges some, but I am new to this 
Thanks for the suggestion. I had actually gone through the scanning offset adjustment at 10 different speeds (100mm-1,000mm/s) in increments of 100mms and adjusted them based on microscope measurements. Once adjusted, the next set of test coupons looked great. I’ll post a pic. Oddly, I did note that while I used the left side to get the measurements and now the left side of the test coupons look good, the right side is not as good. So the lines while the laser moves in one direction are a bit longer than when it moves in the other direction.
The larger font letters that are about a centimeter tall look great to the naked eye on both right and left sides. but the smaller font letters make it look like the scan offset is off. Are you aware of any reason why the scanning offset may seemingly affect smaller objects differently than larger ones?
These 2 images are both from the same exact engraving, just different font size.
70% power (105 watt tube)
0.070 interval
Single pass
I may be able to fine tune the interval and scan offset a bit, but it seems like I can only optimize one font size while sacrificing the consistency of the other. After a dozen different tests with varying settings for power and interval, I’m finding I can get either the larger text, smaller text, or thin logo to look good, but not all at once.
The smaller letter does look like mumble.
I wonder if that’s an interaction between the slightly too wide line spacing, the image DPI, and whatever dither pattern you (may have) applied. Both edges vary by discrete increments, rather than continuously, as though a single pixel / dither dot is either black or white.
Do the offset measurements on simple rectangles set to Fill, rather than images, to eliminate the jank, then pick the offset for cleanest edges vs. maximum removal.
When you care about the edge detail, the image DPI should match the scan line interval to prevent LightBurn from resampling the image and introducing aliasing. If that isn’t the case, then I’d start from the original image and resample it manually to match the scan interval.
Now, if the original image is vector (SVG, AI, whatever) art, rather than a raster (JPG, PNG, BMP), then all that may not matter.
Perhaps this should become a separate thread to avoid confusing folks in the future.
Thanks for all of the info! I’ll make this a new thread, but before I do and lose the context I have just one quick clarification question.
You mentioned that it would be good to resample the original image to match the interval scan. The original image is a JPG, but was then vectorized with the trace image tool in lightburn. Would this get rid of the need to resample the original image, or is that still important prior to tracing the image in lightburn?
I assumed you used the original layout as an image, rather than tracing it to get vectors in a Fill layer. That conversion eliminates the resampling required for the laser scan lines, so you’re in good shape.
That said, if the original original original layout came from a program like Inkscape or Adobe Illustrator that works with vectors, rather than bitmaps, you should use the original SVG / AI file to eliminate the JPG export / import / trace process.
If you process an image through another product, where you have set it’s dither or other modes, you’d use the pass-through option on the layer to prevent Lightburn from re-dithering it before it’s sent to the laser. Note you lose some ability in settings, as the software assumes these are already correct.
Sites like https://www.imag-r.com/ pre-process an image and you should use pass-through with these.
Be aware if you change it’s size that changes things…
Good luck
I’m not currently using any other programs to vectorize logos, only lightburn. Do you find that the image trace in lightburn is sufficient, or does ImgR make vectorizing the images significantly easier or better quality than LB?
As far as I know, ImagR only works with images, not vector art. I have never seen anything in the docs about even using a vector as input type… I have to admit I’ve never tried it with an svg or other vector graphics.
Lightburn uses vectors for tool paths, as far as I know, it’s really not a graphics software… However it has excellent tools for manipulating an image to use on a laser.
Can’t compare it to Gimp as far as image manipulation ability – conversely, Gimp s*cks as laser control software…
Such is life…
