[Peter] has been having some positional repeatability problems with his CNC3020 Router recently. The problem was mostly in the Z axis and was measured to be up to 0.3mm off position after 10cm of travel. This may not seem like a lot but it was enough to break a few 1mm diameter end mills. The X and Y axes generally seemed OK. Surfing the ‘net reveled that the control board’s power rails did not have any filtering capacitors and that may have been the cause of the problems. Unfortunately, the positioning problem still persisted even after the cap’s were added. Frustrated, [Peter] then started a full-blown investigation to figure out why his Z axis wasn’t cutting the mustard.
In a CNC system there are 2 major components, the electronics and the physical machine. Since it was unknown which portion of the system contained the problem, [Peter] decided to quickly swap the X and Z channels, running the Z axis with the X axis electronics. The problem was still evident on the Z axis which means that there is something wrong in the mechanics of the machine. The Z electronics were put back on the Z axis and the testing continued by lowering the acceleration and the maximum speed. The positioning error was still there. Since it is possible that the Z motor could be the problem, it was decided to swap the X and Z motors but midway through the process the problem became evident. When trying to rotate the Z axis lead screw by hand there was a noticeable lack of smoothness and the axis seemed to jump around a bunch!
The Z axis had to be disassembled to access the lead nut and linear rails. This reveled the source of [Peter’s] problem; no lubricant, not on the linear rails and not on the lead nut or lead screw. Over time, the un-lubricated Z axis linear ball bearings had worn 4 gnarly looking equally spaced grooves into the linear rails. The bearings were flushed out which produced an alarming amount of metallic dust that had accumulated inside. Once cleaned, the bearings were lubricated and otherwise seemed undamaged. Since the grooves in the rails were only in 4 local areas, they were rotated 45 degrees so the bearings were now riding on undamaged areas. Even after the cleaning, lubricating and rotating the rods the bearings still did not feel as smooth as they should. Since replacement bearings were not available the Z axis was reassembled anyways. Grease was also applied to the lead screw and nut prior to reassembly.
[Peter] had low hopes for his ‘fix’ but after performing some more tests it seems that the Z axis is performing much better; 0.01mm error over 10cm of travel. Not bad! [Peter] is not afraid to get down and dirty with this CNC3020, he has added a coolant system, limit switches and a PWM spindle control.