Looking for help diagnosing laser consistency issues


Just got a new Laser, and started working with it using LaserWeb, and have had good success mostly, but I do have some issues, attached is the screenshot.

Red, should be the same, but always comes this way, thiner on one diagonal than the other, yet all 4 should be identical.

Blue, same red, having issues with consistency, and weirdly its inverted on the oposite side.

Yellow, all four cuts should be the same but again is not.

Finally green, not all the segments are cut complete. like a hair is left to cut.

Using DXF format from AutoCad.

Any advise would be great.

First, you will want to share whether this is a CO2 laser or an LED laser. Maybe screen shots of LightBurn set up for the cut to show the design intent, and/or even the original DXF file.

In general, this looks to me like problems that are often ultimately due to mechanical looseness of some sort. But without knowing what kind of laser machine you have, it will probably be difficult to give more targeted advice.

Thanks for the reply, the laser is a Mecpow X3 Pro, I guess one of the Chinese low budget ones.

I dont use LightBurn, I am using LaserWeb.

I have done them on a Full Spectrum CO2 with Retina3D and they come out right, sadly that laser is out of comision at the moment.

Let me know for any additional info needed.

The problem could be mechanical or controller/software compatibility
Did the manuf say the machine was LaserWeb compatible?
What controller is the machine using with what firmware?

I would start by eliminating one or more of the above causes:

  • Test print something simple, like circles and boxes with known integer dimensions.
  • Print on this machine with the software and controller it came with? Does that work?
  • Try with an alternate to LaserWeb.

It came with LaserGRBL, but it doesnt take DXF or have layer capabilities, it doesnt work for me, I can try it. But in the end it doesnt work for what I need.

The biggest issue is I use AutoCad, and it doesnt export compatible to LaserGRBL.

While @easytarget would know whether there is any potential software incompatibility to look at, this really looks like various forms of mechanical looseness to me, and this being a LaserWeb problem would definitely surprise me.

It could be anywhere in the mechanical system, from loose belts, loose head, loose eccentric wheels, loose belt pulley, and I’m sure other things I haven’t thought of.

There is consistency on the issues though. Even if I start the job at a different origin point I get the same results.

Wouldn’t mechanical issues move the error to different areas of the job if it starts at different places?

I have yet to test again today.

I would not expect the origin to make any difference. It would only make a difference if the problem were something like an inconsistent extrusion profile, which is not likely in any case.

Sorry late to the party; I assume you were also using LaserWeb when you were cutting on the CO2 machine?

Does this show in LaserWeb’s GCode preview? Are there any visual problems with gaps in the laser cuts there?

If you cannot see anything in the preview window it may help to export the gcode to a file and then load that in (say) UGS (Universal Gcode Sender) to give an independent view of the code.

Assuming the controller in this machine is a GRBL based it is hard to imagine any missmatch with the code LasweWeb produces, GRBL is extensively used and tested with it.

Do you have a manufacturers link to information about the controller? or at least to the sales listing for the machine?

1 Like

I would try this setup just to see if you see any problems with the base machine and its mechanics. Print some simple diagonal closed boxes similar to what is in the failed job.

1 Like

I was able to correct some of the problem by tightening the belts of the laser.

But I am running to another issue when working with a different file.
bend.dxf (211.8 KB)

It doesnt create the cut path for the lines in the middle, just the ones across the center and the outer edge.

NVM I got it to work

Great… how about letting us in on the secret?


Vertical lines were closed polylines, the ones that were showing were actually just a polyline. once I exploded the closed ones they started to show. I guess a line over a line cancels the line??

1 Like