Reducing the Kerf

Being new to laser cutting. Is there a way to change what cuts inside of a line vs outside of a line to reduce the kerf? I have been using both Inkscape and Adobe Illustrator. My wine boxes go together well, but I think the fit could be tighter. Any help would be appreciated.

1 Like

Most CNC programs will let you decide to cut inside, outside, or centered on the cut line. Im assuming the glowforge will be no different. Dan has shown us some glowforge cutouts he did in acrylic that fit together perfectly.

check out this thread for some cut examples he did:

1 Like

You will have the option to choose.


At one point, it had been said that for things like boxes with interlocking tabs, the software was going to be able to figure it out for you and do it automatically, and I have seen and heard that you can designate what is inside and outside of the pieces too; presumably it would cut on the outside side of the lines.


In another thread on this forum it sounded like current laser software forces you to manually adjust for the kerf in your design software. In other words, if you don’t want your parts to be undersize by 0.008" because of your 0.008" kerf, you have to manually offset the lines of your design by 0.004". It sounds extremely primitive to me and if this is the case I hope the Glowforge software is closer to established CNC software than established laser software. As @takitus mentioned, cutting one one side of a line or the other has been a staple of CAM (“CNC”) software for a long time… I mean, even if you just want to cut a 2x4 with a skill saw, you still have to pay attention to which side of the line you aim for.

Regarding how laser software works, I should mention that I’m only trying to paraphrase what I’ve seen on the forum (as I understand it). I haven’t actually used any laser software myself yet.

1 Like

When cutting out different things so far, I’ve noticed depending on how they are created, some fit tighter than others. I was wondering if it goes by color of the lines? Black for etching and blue or red for cutting. I know u have to set the line width at <.03 or it tries to cut twice. One box I made you had to force it because it was nice and tight. The last 2 have had larger gaps. It may even be a speed vs power problem. I will chat with the owner and see what they say also.

It is my understanding the width of the kerf will vary with material type. I would not be surprised to learn that power and speed settings also play a role as well as material thickness. I would expect the cut to be in a slight hourglass shape and that may play a role in some geometries.


youre right, foams will have much larger kerf

1 Like

Surely the default should be to cut on the outside of the outline and on the inside of holes as a router would do automatically. Perhaps some designs you would want centre line if you wanted to use both parts either side of the line.

Sometimes, when a design gets really complicated, it can be hard for an algorithm to tell which is the “outside” and which is the “inside”. Might be nice just to be able to mark “this side waste” the way woodworkers do when sawing.

Good information here. We’re building the ability to indicate where the kerf should go. Kerf measurements vary with material, speed, power, and more. With most lasers you need to manage it manually e.g. by making all of the pieces oversize.


Is there any chance that the high-res camera could be used to measure the kerf? Do a little test cut (or two slightly separated cuts) in a waste section and report back?


That is an interesting question. As I try to imagine what is needed to control the Glowforge and different ways the camera can be used, the whole idea of what “computer vision” is becomes fairly complex. Right now, I just imagine the cameras are used to help position the material and the origin of the head. Then comes the question of how feedback is incorporated into the process to allow for positioning using the pass through slot, still a static alignment procedure. How would the procedure you envision work? Report back to whom? the head? the user interface? Just what are these cameras going to accomplish beyond alignment?

1 Like

That’s a very good idea!

If the user interface included a live view from the camera, a reticle, and if it displayed exact coordinates of the head, we could use that information to make our own judgement about how wide the kerf is. Some kind of automated system would be cool, but I’d be happy with doing it manually.

Actually, measuring/specifying the kerf manually could have some advantages. Surely we’ll sometimes want to leave a little bit more breathing room so the object can be finished using another process (sanding, finish-milling, scraping, whatever). In CAM software this is typically called something similar to “stock to leave” and is very helpful.

Barring the Glowforge making the job easy for us, a macro-capable camera and a calibration slide might be good enough to get an accurate measurement of the kerf. Ooh, feeler gauges could be useful too.

Here’s a random calibration slide on eBay…

Here’s a random set of feeler gauges on eBay…


Having nothing to do with the Glowforge specifically, if you have a decent set of calipers then figuring out the kerf is pretty easy.
Cut a 1" square out of your material and measure it and the hole it came out of.
The kerf is 1/2 the difference between them.
This has the added benefit of allowing you to make sure the kerf is the same in both directions, it also tells you which side of the line you are cutting on (i.e. is the hole bigger and the square 1" or vice versa), or if you are cutting down the middle.


Oh yeah, that’s a good method too!

I don’t want to ask you about your Glowforge, but can you help me understand how the kerf would be wider on one side of the line than the other? I could imagine the 1" square would be heated more than the rest of the material, given the fact that it has much less mass to absorb the heat, so might shrink/expand more than the rest. Are there other reasons?

Ooh, feeler gauges might be plausible.

One of the things I like about the camera version is the possibility of doing things in situ. Make one or more cuts at a given speed/power in a waste part of the sheet, measure, and then adjust based on that without pulling the stock out and messing with it. I don’t know whether you’d do better with the camera trying to tell you the width of the cut it sees, or making two cuts a small known distance apart and then having the camera measure the width of the stock between them.

One of the things I want to do is marquetry and marquetry-like operations (e.g. not on flat surfaces).


As I understand it, jkopel is not suggesting the kerf is wider on one side of the cut, but that you can use the size of the square and hole to determine if the laser software is adjusting one way or the other for the kerf. If your square is 1.000" and your hole is 1.010", then it is shifting the cut out. If your square is 0.090" and your hole is 1.000" it is shifting the cut in. And if the hole is 1.005" and your square is .995", the cut is going down the center. That is how I read what he is saying.

Even with a quality set of calipers the average person is as likely as not to introduce a measurement error greater than the kerf rendering the point practically moot if theoretically sound.


Bear in mind that the only laser cutter I have worked with directly is the beta Glowforge.
But… the kerf is the kerf, most internet resources quote CO2 laser cutter kerfs at approximately .008, but as @dan points out they also say it varies by material, cutting speed, and power level.

I know what I said was confusing, sorry. Let me try again. If you are cutting a square, and the kerf is OUTSIDE the line, then the square will be 1" and the hole will be bigger by 2X the kerf. On the other hand if the kerf is inside the line, then the square will be smaller and the hole will be 1". If you are going straight down the middle then theoretically the hole and the square should each be smaller than 1" by the full width of the kerf.


In situ would be good for sure!

I like the idea of just measuring a single cut. I’m thinking that adding a second cut introduces unnecessary variables, particularly the positional accuracy and parallelism of the second cut. The 1" square test is something that somebody might perform to make sure their machine is moving accurately. The trouble is: what if your outside-cut square comes out to 0.999" x 1.000"? Is the error more likely to be with an inaccurate kerf measurement or just a simple 1-thousandth error in the positioning of the laser head?

Actually measuring the void of the cut with the built in camera will be a feat, but it would totally be the best way to go.