How close is close enough with MPCNC calibration?

Working with a pen, I have been using 2D plots to check the MPCNC calibration. There are two important tests. First, are the X-Y axis square, or perpendicular to each other. Second, is the Z or vertical axis perpendicular to the work table.

There are several structural adjustments which can be made to the Mostly Printed CNC. First is the spacing of the corner posts, which can be checked by measuring the diagonal across the machine. Both diagonals should be equal. Second, is the table level, and are the horizontal guide rails parallel to the table?

Using a bubble level, all of the horizontal surfaces – table top and guide rails – look pretty horizontal. Small adjustments were made to the corner brackets and leveling table leg. Measuring the diagonals looks close – within 1/16 inch.

MPCNC Calibration – Test Patterns

The first test pattern was some circles and lettering, shown above. I think I still have a little slop in the center mount, and I know there is some slop in the felt pen tip. This might explain why the spacing between the concentric circles varies a bit. But I also noticed that the printed letters became lighter towards the right hand side of the print. This tells me that the level of the machine is slightly out from the table – vertical space is increasing as the machine moves to the right.

The second test pattern was a simple 200.0 mm square, printed with a fine tip pen. Again, I noticed that the pen line was lighter to the right and darker to the left. When I measured the sides of the square, they turned out to be 199.5 mm. Also, the diagonals were slightly out, telling me that the square was not quite square.

So, this quick MPCNC calibration test tells me that I still have some work to do. I think the main issue is leveling the horizontal planes to get the tool truly perpendicular across the full work surface. The squaring is close enough for now, within 0.25% of perfect.