To adjust for kerf or not to

I’m willing to ask Glowforge to make their software better than what people with two week’s experience will settle for.

re: my unwillingness to imput fake dimensions -
Kerf adjustment is STANDARD practice in CAM software. If people who have no experience with CAM software are confused by the word “kerf” for a couple minutes, so be it. Yeah, it’s a funny word, but the concept is as simple as can be. It’s basically “coloring inside the lines”, but with a laser instead of a crayon. I believe people are capable of learning. To those who don’t want to adjust for kerf, allow me to introduce you to the number “0”.

re: my unwillingness to watch the laser bounce around - It’s inefficient. Cutting line segments by randomly bouncing around the bed is, put simply, wrong. As I said, it’s not the end of the world, but I will not willfully allow it to happen, which leads to…

re: my unwillingness to spend hours tediously fixing disjoined files - I have Rhino, it joins lines perfectly, instantly, and doesn’t do anything stupid while doing it. I realize not everyone has Rhino, but (almost) everyone here has a Glowforge coming to them. I don’t think it’s unreasonable to ask Glowforge to save their customers hours by programming their cloud-based software to join lines intelligently.

re: my unwillingness to clutter my files with kludge-geometry - This isn’t entirely true, as this is basically what I do now. I draw something in CAD (whether it be in Rhino, SolidWorks, Fusion 360) or I obtain a drawing created by someone else, THEN I load the file into Rhino and basically use it as if Rhino is the CAM software for my Trotec. It’s not efficient and I wish it was different.

If I thought Trotec was interested in modenizing their software to get out of the late '90s, I’d probably be complaining on their forum (if they even have one). But I don’t think Trotec desires input from their customers, I think Glowforge does. Please tell me if I’m wrong about Glowforge.

So, in truth, I am willing to do it, but I won’t forget that doing it this way is… wrong. I’ve already enumerated why adjusting for kerf within your main design file is riddled with problems in a post above. Yes, it DOES work. It CAN be done that way. If you’re willing to live with the inherent problems that creates, there’s nothing stopping anyome from doing it that way. It goes without saying, but I’ll say it a second time anyway, having manual kerf adjustment in the Glowforge software won’t stop anyone either, as long as they remember the number “0” exists.

8 Likes

Agree. It makes the design file material and machine dependent. I hate the tedium of adjusting as I switch materials or power & machines. So I’ve got different files set for different machines (I keep them segregated by machine & material in separate folders). Makes for lots of redundant work.

4 Likes

The folks that did the Codex Silenda KickStarter laser-cut book generated the files with kerf-adjustments in 0.005-in increments for wood in thicknesses between 0.190 and 0.250-in - that’s a lot of files!

It would be so much easier to work with a nominal cut file and specify the material thickness at the time of cutting. That might be a good reason to use a cloud-based controller.

2 Likes

Kerf will be so wonderful to have in the app as a nondestructive mod to your art file. (inside outside or center, and by how much). yep +1

3 Likes

Did we get the proofgrade auto kerf compensation yet?

1 Like

No

1 Like

I know this is very late to the party and I read the first 1/3 of the replies, but since a lot of them predated shipping units, I think there is something very important to point out about the kerf and comparisons to CNC/CAM software.

A CNC mill uses a tool that cuts through a given material. Let’s set aside the wearing down of the tool surface for a minute and assume that the tool is a consistent size for cutting. It cuts/chews through a material to make a cut. That is a fairly consistent thing outside of some materials breaking down differently at the cut point and thus kerf adjustments (assuming software and mechanical movements aren’t also contributing).

A laser burns through a material. This adds a big variable to the kerf equation that CNC/CAM doesn’t have to deal with. When a material burns it behaves in a wide variety of ways and the burn/consumption rate is different for every material - if the material is made up of only one consistent fuel type. Most plywood contains wood from different trees with different levels of moisture, oils, cell structure and more. MDF contains a variety of materials that all burn at different rates. The laser (in simplest terms) burns completely through a given part of the material, but the fringes of that burn are where things get complicated and inconsistent.

A good simple example is when you light a piece of newspaper on fire and you blow it out. Very often you still see glowing orange embers on the edges of the area that burned that continue to “eat” away at the paper till there isn’t enough heat left to support combustion. That is a more extreme example of what’s happening when a laser burns through material and the edges left behind are cooling. Every single material and thickness is going to behave differently with this burn rate. Wood will burn differently than acrylic. Cast acrylic will burn differently than extruded acrylic.

So you have the following variables:

  1. Material composition and the burn rate of that specific material

  2. Speed of the moving laser (or fire let’s say) which affects how quickly the material around the cut area heats, consumes and cools.

  3. Power of the laser (or heat of the fire) and how that heat affects the area around it (again, heating, cooling, expansion, contraction, consumption).

So you could imagine how GlowForge could learn the kerf/burn characteristics of their Proofgrade materials that are consistent in material, thickness and presumably construction from batch to batch. But if you don’t want to use Proofgrade materials (or in my case I need thinner materials not available in Proofgrade) then you have to purchase material elsewhere and do a lot of testing to see where the limitations and compensations are.

The main point, is that this is NOT comparable to CNC/CAM software where there is a fairly consistent kerf variable (more or less). It is a bit more complicated when you burn through material. I could have 1/8 inch maple in the GlowForge and the kerf at 25 power and speed 1000 is going to be different than the kerf at 100 power and 50 speed. And that’s with the same thickness and material composition.

Now, all that said, manually being able to enter a kerf value would be a helpful addition for some people. That itself is a little complicated too from a GUI standpoint as you may only want to apply kerf to a given cut line in the interface and how you go about selecting just that cut line, face or side of a drawing. I can see situations where someone doesn’t want kerf applied to a given part but may want it applied to other parts.

Anyway, if you made it this far, thanks for reading and hopefully it helps explain why this kerf situation isn’t that simple.

11 Likes

Believe it or not, some people will read long posts as long as they are broken into paragraphs and reasonably well written. Yours is both.

I totally agree with all your points here and have had arguments with people here that have CNC experience but had not seen the :glowforge: .

One thing your essay leaves out though is that different joint types need different kerfs. The solution is methodological testing. For complex (lots of joints) projects I always make sample joints to figure out the amount of kerf compensation needed.

5 Likes

Yes kerf will vary a lot and may need to measured but once you have a value offsetting the design is just the same as any other CNC device. You dont specify where it should be done, outlines get expanded outwards and holes get offset inwards. So all it needs another input field.

The object should always be that the finished piece matches the dimensions in the design file.

1 Like

You are right here. If you are using Fusion 360 there is a plugin that gets it right 99% of the time (inside or outside) or even better once someone gets a bit of experience is to include these as parameters in their parametric design.

1 Like

Yeah, nice writeup. :grinning:

So I’m finally getting to the point where I’m using a lot of non-Proofgrade materials for projects and this has become an issue. I started to study in in the forums and this is one of the first threads I came to. I feel stupid for never realizing this was an issue, everything fit together so well in the past! :slight_smile:

One question I have for people who are CAD focused like myself - is it essentially just offsetting linework to a specific kerf? Meaning, if I’m using 3/16 plywood and I know my kerf is .005 because I’ve measured (found another post re: measuring kerf on the forums), offset the linework by .005 outward and I"m good? I think it’s that simple but I’m not 100%.

And I’d love to use other tools but frankly it seems to be a waste of time to me to draw it up in 3d when all I’m looking for is 2d linework to act as cut lines. I’ve done everything up to this point in CAD plotting 1:1 to PDF’s and have had zero problems.

Half kerf. But otherwise correct. (You basically want to adjust each piece by half the kerf - the laser beam always follows down the center of the path, so the amount of material on your shape nibbled away by the beam is half of the total kerf width. The other half gets nibbled away from the waste material.)

1 Like

Duh! 1/2 kerf. I won’t post pictures of the 1/2 dozen tests I cut yesterday when trying to figure out the kerf and what worked / didn’t. Or even admit I did it. Seems silly now. :slight_smile:

Thanks!

1 Like

Sort of. The typical CNC cuts are either on the outside of a line or inside so there’s no kerf - the software takes care of moving the mill out (or in) a qtr inch for example.

The GF cuts on the line so you get a cut that is the width of the beam. Instead of a qtr inch, it’s generally quite small - 0.007 inch is what my GF does. Since the beam straddles the line you’ll get half that width on the inside and half on the outside.

If you’re doing a hole, you make the hole half a kerf less diameter (0.0035") and it will be a whole kerf bigger when it makes it all the way around. Ditto for squares.

For tabs and slots you can adjust each one by half a kerf or just one component (tab or slot) by a whole kerf.

It’s more complicated because we don’t get to tell the GF which side of the line to cut. Since operations are by color it’s going to be a bit harder for them to implement kerf adjustment by specifying which side of the line you want - in VCarve for instance you pick the object and define the toolpath and the side of the line and can do that for any objects you grab & include in that toolpath. The GF doesn’t let you do that - if they keep the color orientation and they base kerf adjustment on colors we’re going to have to think about kerf and assign similar kerf actions to the same color vs just the operation & order.

One more complication - if you’re using Inkscape, it includes the line thickness in the width you define for the object. So you don’t want just a skinny line but a hairline for your lines.

1 Like

I believe you can change this behavior by selecting Geometric Bounding Box in settings…

ISettings

3 Likes

Good to know. Is it a V0.92 feature?

It might be, I didn’t really use the previous Inkscape versions that much. I was having some issues with this when I came across this tip elsewhere in the forum!

I used the old one for a long time and when I recently upgraded it I noticed they removed some of the doc sizes - used to be able to set yards and meters but not anymore. I hate change :slight_smile:

1 Like

There is a special plastic glue available from Micro-Mark.com that works by capillary action. Just swish it across kerf and it gets sucked in. Works in 10 seconds. It’s called Professional Plastic Welder.

2 Likes