I noticed early on that vector cuts on my Epilog TT had wider kerf on the top side than on the bottom. The cuts were creating a shoulder, like the top was beveled. I always use the manual focus tool, but wondered if maybe I was getting too close so I checked the focal distance using a bit of inclined stock and it seems the tool is pretty much right on, getting the smallest kerf possible. Nobody else seemed to think it was a problem because the shoulder was pretty small, so I left well enough alone at the time.
Recently, I was asked to make small holes (about 0.033 inch, to accommodate a 20 gauge wire) in 3mm BB. The holes are perfect circles on the bottom/back side but on the top side they look like ellipses on something close to a 45 degree angle, and significantly larger due to the shoulder than on the back side. Now people aren't happy with that shoulder, especially since it is asymmetrical. I talked to Epilog about it and was told that's normal (?!). I asked if there was maybe a beam alignment issue that could be adjusted and was told it wasn't the beam but rather a problem with how the x and y motors advance. When pressed, the tech told me to try adjusting the ramp and acceleration compensation settings and tried to explain it as being a problem of the x and y motors not moving in precise synchronization to create a perfect circle. I told him that didn't seem right but I'd experiment more.
I've found that the ramp and acceleration settings had no visible effect on this problem, but the description in the manual doesn't make it sound like they should anyway. It describes these settings as controlling the reversals or turnaround of the head, not axis interaction. Besides, the holes were perfect circles on the back side. As a side test, I tried making squares instead of circles so that only one motor had to move at a time and I got the same oblique distortion... perfect squares on the back side and rhombus shapes on the front, stretched along the same back-left/front-right diagonal.
So, then I tried making a grid of holes, varying speed along one axis and power along the other. When the speed/power settings put low energy on the surface (such as high speed and high power, or low speed and low power), I got a perfect shape (whether using square or circle) lightly etched on the top surface. As the speed/power combination increased energy on the surface, the shape got increasingly distorted on the top surface but was perfect on the bottom. Even at the lowest possible speed, where acceleration seemingly couldn't be an issue, the cut was distorted on the top surface when enough power was applied to burn into the wood. Same thing on plexi. I tried making the holes at all four corners and the middle of the table and got similar top-side distortions everywhere.
The two pictures show the results using 0.03" squares vectored at one speed and 10 different power levels, one picture from the top side and the other showing the back. The "shoulder" and diagonal distortion on the top side are quite clear.
Seems to me like this has absolutely nothing to do with the motors or acceleration; rather it seems to indicate the beam is "hot" in the center but, instead of having a circular profile it is spread out in an elliptical shape on roughly a 45 degree angle, because it marks a perfect figure topside at low speed*power densities, and the bottom is perfect at higher speed*powers but the top becomes distorted as described above.
Any thoughts? Solutions/work arounds? Thanks!