Below is the gcode for a missile badge. I started with a 1/8" bit and ran the 1-8 file to contour shapes and then changed to 1/16" bit and the 1-16 file (just contour on a separate location so you don’t have to run it). Last a 1/4" with the 1-4 file to pocket and cutout. Each bit had touch zeros at bit change. The 1/8" depth didn’t match the 1/4" and the 1/4" pockets were uneven. All bits were flat tip mill ends.
If you do test, can you run in mm to watch the z sled location compared to gcode.
Back from my Guam “deployment”, fired up the Maslow and noticed something with my z axis that I didn’t before.
In webcontrol, right click to move sled to far corner. The z axis position on webcontrol stays at 5mm my default. Right click return to center and it moves to center but the z axis position number on the screen changes by .08+/- mm.
The z axis position on the screen only changes when I right click center but not when I right click move to location. Note that the z does not physically move only the position on the screen changes.
Could this be related to my other z axis depth issues?
Bad encoder? Bad motor controller?
Should I buy/try a new motor/shield?
Is there a better z axis motor/encoder out there to provide better accuracy/tolerance?
The maslow only aspires to be accurate to ~0.4mm, so a difference of 0.08mm is
below the resolution of the machine. It’s probably that the movement of the sled
jostled the router just enough to trigger the encoder to move enough to be
detected and reported, but not enough for you to notice.
So I was thinking to myself and my theory is probably off as I know little to nothing with encoders. But here we go:
I’m running a 60:20 ratio on my z axis so if I want to move 1 mm the motor doesn’t need to turn much. If I run a 20:20 ratio on my z axis then the motor will have to turn more and provide more movement against the hall sensors.
Margin of error for 60:20 was .06-.12 mm over each of the 4 passes.
Margin of error for 20:20 was .03-.06 mm over each of the 4 passes.
The gcode plunge rate (8 in per min) was the same between setups. But the 20:20 gearing is incredibly slow.
If I ran a 3 in per min gcode with the 20:20 gearing it was .01 mm error over each of the 4 passes. But it was like watching paint dry.
I’m curious with a faster motor if you can keep the ratio 20:20 how accurate would it be, or even at 60:20 combined with a slow gcode plunge rate. @Metalmaslow looking at your motor, do you have some input?
You are correct that the 60:20 ratio is giving us more encoder resolution to work with, but the way the faster motor works is by adding less gears after the motor before the belt drive which sort of has the same effect. Basically the faster motor will reduce the encoder precision in the same way that a slower belt setup increases the encoder precision.
in the same way that a slower belt setup increases the encoder precision.