4' x 8' DIY CNC Router Build with high tolerance Z Axis
I have built a CNC router recently and I am starting a sign fab company. My router is 2' x 4' but I need a 4' x 8', for the material I am cutting I need a .001" tolerance on just the Z axis. The X and Y do not need as high of tolerance but I am cutting aluminum composite material and V grooving so it is extremely important that the Z axis is between .001" and .005". Does anyone have an idea on how to effectively do this?
Thanks,
Sam
Re: 4' x 8' DIY CNC Router Build with high tolerance Z Axis
I think some obvious points you already know are to use very high precision ball screws (they come in various grades), use no backlash ball nuts, run low pitch screws so your steppers or servos have the steps/counts spread over a smaller distance (like one Rev being 5mm movement instead of 10mm)...
Sent from my SM-G781U1 using Tapatalk
Re: 4' x 8' DIY CNC Router Build with high tolerance Z Axis
All you really need is for the axis to be repeatable. You're going to be setting your Z zero at the bottom of your sheet, so you only need the machine to move .001 from where you set your Z zero. You don't need .001" accuracy over say 12", you just need it to be able to return to .001" from your Z zero.
Now, actually getting to .001" is a bit trickier. If it's a stepper driven axis, I'd want the full step resolution to be .001" or better. At 200 steps/rev, that's a screw pitch of 0.2", or right around 5mm.
In reality, I doubt you'll be able to V groove within .001". I think .003"-.005" is more doable.
To get to .001", everything needs to be perfect. Spoilboard thickness can vary far more than that throughout the day.
Re: 4' x 8' DIY CNC Router Build with high tolerance Z Axis
For that precision, depends a lot on your Z-axis framing, slides & drive. The frame should be metal boxed in.
The slides can be linear or pillow block, but should be hefty and good quality.
The drive is ideally acme ballscrew, big diameter, or linear gear. Both have to contend with the axial (downward) load.
As ger21 says, everything has to be "perfect". Very tight tolerances on bearings and alignment.
A friend seeking such accuracy designed a probe to measure work depth variation on a 2d grid, then compensated by modifying the gcode depths to match distortion.