We are testing some new hard nozzle concepts at the University, it is part of a student project, and we used the XT-CF20 as a reference material.
After some initial testing with a normal model I went ahead and decided to design a "nozzle killer" model.
This model however had the opposite effect, it only caused about 10% of the wear that the normal model caused (!)

It appears like the wear the XT-CF20 causes comes from when the nozzle is changing direction rather than from the shoulder of the nozzle scraping along already printed surfaces.
The length of the carbon fibers might play a role here.
Here is how the fibers look when printed at 0.05 layers with a 0.4 nozzle at 50 mm/s
The fibers appears to be about 5 micrometer in diameter and 50-100 micrometers long.
Next up is to test print "nozzle killer V2"
