Inspired by one of my readers I had a look at the dust shoe for the Kress again. I redesigned it completely and got rid of the additional tubes and clamps. I thought that it would be a good idea to have a highly flexible part between vaccum and CNC. If I had ordered the 20mm silicon tubes it may have worked as well but reducing parts is allways priority.
So in the end we got a pretty classic dust shoe here. It doesn´t have a brush at the bottom as it seems to work pretty good at the moment and I really like to see the endmill. However I included three holes at the bottom for a later brush design.
It works pretty good with lighter materials but GFK, aluminum or similar are not fully sucked up.
It has been printed in 0,3mm draft quality for testing purposes.
In order to find a suitable endmill I had to understand the principles involved first. That led me to feeds and speeds. The difficult part was to determine the limits of my cnc. I already had a look on that topic when I chose my Spindle. It has a minimum RPM of 10000 and a maimum of 29000 RPM so I decided to go for a 2 flute endmill with a diameter of 3 mm from cnc-plus.de. More on the calculation later..
After testing the gcode and finding out if the steps/mm of the steppers are dialed in correctly I used a simple ruler to check. Josef Prusas stepcalculator gave me 160 steps/mm and that resulted in the correct distance of movement with my 20 tooth pulleys. As I couldn´t set the parameters from the graphic display I had to reflash the Marlin firmware.
I used a M8 threaded rod like suggested in the original plans. It is not meant for machine operations and so It loves to do what it is designed for. It binds pretty often, resetting the z axis position and making sudden, unexpected deep cuts as shown in the video below. It is the straightest rod I could find and it is well greased but it was necessary to lower the acceleration and speed. I only had this one piece of hardwood and I wanted this to suceed. I´m trying to get the speed back up again because the milling took ages (1h15min) because of the adaptive clearing and the many retractions of the tool head.
The 3D printed rigid couplers didn´t worked to well for me so I ordered a jaw coupler also from http://cnc-plus.de to make it more reliable.
In conclusion : I liked the results of the cuts and they were somehow thrilling. The accidentially super deep cuts showed that I can go much deeper than 3mm in one pass with hardwood. The method used here was also a test for the ultimate goal for cutting alumium sheet and composites.