Epilog kerf investigation

[Peter Lowe]

Hi everyone,

I am trying to get an Epilog cutting optimally for a project I'm working on and and am having some issues with the kerf width. It is a 60W helix from 2008, and I am hoping to get its kerf while cutting MDF very similar to that of a 40W helix from the early 2000's that I have used.
The kerf is fairly thin when the machine cuts at the lower right corner, but is quite noticeably wider at the upper left (closest to where the beam enters the table), and leaves a rounded charred edge on both sides of the cut.

Top left cut: https://www.dropbox.com/s/032ai62xww00e71/top%20left%20cutting.JPG
Bottom right cut: https://www.dropbox.com/s/h914vq9i8z17jf2/bottom%20right%20cutting.JPG

Examining the pattern of the red dot pointer on a piece of paper attached to the cutting head while moving it around gave me a theory about the problem. The pointer is fairly focused at the head when it is at the upper left of the table, and grows noticably as the head is moved toward the lower left, with two small red points migrating completely away from the cutting mirror (and so not going into the lens).

Top left pointer: https://www.dropbox.com/s/nmvk798aw0ixv72/top%20left%20pointer.JPG
Head slid 1/2 to right: https://www.dropbox.com/s/0ayzrgf6mtvsh5t/top%20left%20pointer%20slid%20.5%20right.JPG
Head slid fully to right: https://www.dropbox.com/s/6rlgiet6srobbi2/top%20left%20pointer%20slid%20full%20right.JPG
Bottom right pointer: https://www.dropbox.com/s/66rwpqw5mxljylb/bottom%20right%20pointer.JPG

I am thinking that if the cutting beam follows this same pattern (with the assumption that imperfections in the red dot pointer are due to the path through the machine, not the pointer itself), then some of the poorly collimated light escapes completely from the cutting path and doesn't end up increasing the spot size.
Additionally, the light that is poorly collimated, but not badly enough to completely escape gets focused tighter by striking the lens closer to its edge (the beam effectively being expanded by the increased distance to the cutting head).
My whiteboard sketch of wider beam having smaller spot size: https://www.dropbox.com/s/3zg3hnpzq071z9r/Cutting_theory.JPG

I have cleaned all the mirrors and the lens (and replaced the mirror in the cutting head), aligned the beam per the published procedure, and am trying to decide what to try next.
There are a few small spots on the mirrors that I did not replace and my only remaining thought is to replace them and see if that fixes the problem.

Does anyone else have thoughts?

I would also like to improve the bottom right performance if possible, as it is much better than the upper left but still noticeably wider than the 40W machine's cut.

Thanks!
Peter

[Dave Sheldrake]

Beds not level Peter, the bevel is a product of the short depth of field of the lens system in use.

Kerf is a product of incident beam diameter / focal length, the bigger the beam diameter going in the smaller the final spot size.

Scatter on the red dot isn't unusual, they are very low power diodes so tend to scatter much faster than the IR beam (much lower quality *resonators*)

Measure the distance between the material surface and the bearing rail, a 50mm / 2 inch lens only has an effective depth of field or roughly 2mm, the bed being out by 0.5mm over the entire surface will suck up 20% of that.


cheers

Dave

[Peter Lowe]

Hi Dave,

Thanks for the reply, I took a little time to investigate more.

I tested a piece of material by cutting ~.25" squares while incrementally raising and lowering the cutting bed at cut settings just barely sufficient to penetrate the material. I called "0" where the focus foot just touches the material, "+" numbers when I lowered the bed away from the lens, and "-" numbers when I raised the bed.
The most complete cut was at -3, which is consistent with my findings on the other machine- this puts the focal point about 1/3 through the sheet.

Front of sample: https://www.dropbox.com/s/aojqum6amb0mmrq/front_test.jpg
Back of sample: https://www.dropbox.com/s/asdw2l7umfbfb9g/back_test.jpg

This seems to suggest that focus isn't the problem, since the -3 focus on this sample is the only one to penetrate but has a wide kerf and rounded edge, whereas the sample cut at the bottom right was also at -3 and fully penetrated (same cut settings and presumably same focus, as determined by focus mechanism and degree of laser penetration), but looks much cleaner. I realize that focusing below the surface of the material could result in rounded edges due to the conical beam shape, but I don't believe that this would explain why samples cut at the same settings and focus on different areas of the table would look noticeably different.


The smaller kerfs at positive numbers like +5 are a little confusing to me. From the top they look properly focused (and similar to the sample cut at -3 in the bottom right of the table- https://www.dropbox.com/s/h914vq9i8z...%20cutting.JPG (https://www.dropbox.com/s/h914vq9i8z17jf2/bottom%20right%20cutting.JPG)), except that they come nowhere near penetrating the material.

Does any of that shed more light on the problem? :D

Thanks very much,
Peter