I’m making some “round” boxes where the cover and base fit together and an running into issues because the pieces are not coming out round. The lid will bind as you try to rotate it on the base (I’m leaving 0.040" clearance too). I confirmed it with digital calipers that it’s coming out slightly elliptical shape. I originally thought it might be stress in the wood but have run enough attempts that I don’t believe it is.
I’m not sure what to look at on the machine or adjust to fix it. Suggestions?
I often cut round holes within 0.0002" (yes two ten thousands) on my Shapeoko, by making the program think it’s cutting an ellipse. I have tried adjusting the steps probably 20 times but while it got better in ONE spot, it got worse in others.
I finally mapped out my entire machine (inch my inch, in a grid) to better understand what was going on. The timing belts vary in their timing, as well as the timing bely pulley also has a tolerance…add all this up and you have an awesome +/- 0.003 machine. (Which it is sold as).
Number one cause (that you can fix to get it to a +/- 0.003" machine) is belt tension, and making sure that the pulleys are not spinning on their shafts. As a Shapeoko expert, I highly recommend that you tighten the belt, check the pulleys where you have the greatest error.
((and learn how to adjust the program to cut what you want.))
I’m going back to see if I can adjust this tomorrow. After checking the belt tensions; I plan on mounting my digital dial gauge on the X, Y and Z axis one at time. Zero it and then jog that axis a fixed distance close to the range of the gauge and let it tell me how much it actually moved. I should be able to calculate & adjust the values precisely this way and repeat until I nail it.
I finished the adjustment process. Did belt tension again first then adjusted X, Y & Z values. I started using mm but switched to inches as the math was easier. While I had the dial gauge on the Z axis I jogged it around the spoil board and found it to within 0.003" at it’s worst point. A least know how much to surface it.
Just attach the dial gauge and zero it. Then Zero Carbide Motion; jog it an inch then read the dial gauge to see how much it actually moved. It was easier than I thought. I had a chance to see how well it would repeatably return to zero as par of this process and found it most times at 0 and at most 0.003" off. Close enough!
I thought I’d give an update to this thread today as I’ve resolved my issue.
Today I took much of the machine apart and reassembled it down to the wasteboard. I then squared it, checked the eccentrics, and pretty much everything else. Calibrated $100, $101 & $102 again too.
What I did new was check the square between the X and Y axis. Playing with it I would measure that if I pushed the gantry in the Y direction from the left side and then right side manually it would go out of square (as it should). Then played around an tried to see if it had a point it wanted to return to if you pushed the gantry form the center. It did and that point the X extrusion was not square to the Y extrusions. I learned that on my machine if I push the gantry toward the read from the left side and before turning the machine on it would be perfectly square. The Y steppers hold it in place after the machine is on.
To test it I cut a 4 inch round pocket and measured the diameter every 45 degrees. They measured 3.9925, 3.9935, 3.9975 & 3.9980. Worst to best is 0.006" and some of that is probably noise in my measurements and the pine I was cutting.
That’s close enough for now that I’m happy. How the cut was made conventional changing to climb might improve it more. I’ll continue testing.