Shapeoko3 Z-Axis Carriage Stiffeners

Once I got comfortable with the basics of running a CNC (CAM, manual g-code, homing, zeroing, etc) I wanted to start to push the Shapeoko3 to reach the limits of material removal and surface finish. I started off with a block of aluminum, a 0.25″ 2 flute end-mill and just started entering g-code to jog the machine back and forth at set depths, widths of cut, router speeds, and feed rates. This quickly confirmed the difference in stiffness between the X and Y axis.

Despite being made from heavy stamped steel the Z-Axis carriage was flexing during cuts in both the X and Y directions resulting in loads of chatter at high material removal rates, and poor surface finish during finishing passes. To fix this I decided to build a pair of stiffening gussets that would tie the carriage plate directly into the spindle (while trying to keep things somewhat clean looking).

Initial sketches and dimensions taken from Shapeoko.
Initial sketches and dimensions taken from Shapeoko.

The stiffeners would run from the top of the carriage to just below the spindle mount. They’ll be notched to wrap around the spindle mount so that the attach fasteners will see lower loading. They’ll tie into the carriage plate at 3 locations with countersunk hardware.

Whipped up a model in Inventor and  got everything CAM’d and made a trip to my local material supplier of choice: Shapiro Supply. I cut out the right hand support first, tweaked the CAM, and cut out the left hand a day later. The first RH part required a some filing to fit it to the spindle mount the LH part with slightly revised G-Code (to repeat finishing passes and make up for machine flex) came out quite nicely.

STEP Files for anyone interested: Shapeoko3_ZCarriageStiffeners

Model for LH Stiffener after tweaks.
Model for LH Stiffener after tweaks.

Then I stripped the Z-Carriage off of the machine.


Centered the spindle mount and test fit the stiffeners.


Measured the offset of the stiffeners from the side and added half the thickness to the measurement (to give me the center line of the stiffener that I could transfer over to the carriage plate). Applied some layout fluid to the backside of the  and scribe a center line. Center punch holes and drill pilots.



Dykem the mating surface of the stiffeners and transfer holes. Drill pilots in the stiffeners and open with tap drill. Tap for the screws I’m using (#8-32).




Another sanity check after I finished tapping the holes in the stiffeners to make sure everything fit together before countersinking the holes in the carriage plate and laying out the holes that would tie it into the spindle.


Layout the fasteners that will attach the stiffeners to the spindle mount.


Drill, Tap, Countersink all fasteners.




Re-install the Z-Axis carriage. Done…almost.


After I installed the carriage I was still a little dismayed at how much movement  was present in the Y direction. Upon further examination I realized the movement was no longer the carriage plate flexing, but in the V-wheels…I thought I was stuck with it and the slop was just inherent to the design of the machine. Then while applying pressure to the spindle I noticed a slight popping coming from one of the wheels (that was associated with ~0.010″ of movement). I assumed something was loose, and tore back into the Z axis.

I tried to trace down where it was coming from: did I have a loose fastener somewhere…? I completely disassembled the Z Axis again and re-assembled, re-tension all of the v-wheels on their rails etc. The play was still there, the ‘popping’ was still there. It appeared to be isolated to my lower LH V-wheel…so I concentrated on that.


When I disassembled the suspect V-wheel,  I took some measurements and the shim between the two bearings mic’d at 0.042. Measuring the delrin V-wheel itself, it checked with a 0.030 wide internal ridge. The 0.042 shim was to thick and was allowing ~0.012 of float between the bearings.


I changed out the oversize spacers with thinner washers (which should result in a nominal fit to a  slight pre-load on the bearings) and the remaining slop was gone.

Chatter was greatly reduced allowing me to increase material removal rates, and finishing passes could easily take full depth cuts on 0.25″ thick stock while still maintaining great (for a desktop machine) surface finishes.

Video shows me machining a pocket in 6061-T6 at 20ipm, 0.10″ width of cut, 0.050″ depth of cut.

This was one of the first tests I did after the mod and have since settled in on 30ipm, 0.10″ depth, and 0.050″ width cutting 6061.

Now if only I could control the spindle speed through the software…