Has anyone considered uncoupling the filament from the laser and having the laser read a constant metallic surface for example (much like an idler bearing maybe?)?
That's pretty much the solution used by jtx on his skelestruder and it would at least allow all filaments to be calibrated against a constant....