Way back when I first setup my home/limit switches, I too was concerned about the repeatability of the mechanical switches that I used.
On the Z axis, I have an optical interrupter type of switch that I salvaged out of a copier or printer. I made 10 passes at homing and moving to a fixed point to see what the repeatability was. In these 10 passes, I averaged a difference of .00029" (0.0073mm) so as far as I am concerned, there is no problem with that.
On the X and Y axis, I used some mechanical switches (AAP series miniature DIN limit switches) which are supposed to have a repeat accuracy of "0.01mm on the operating points at 1 million operations"
In my testing using the same type of operation as above, I got results on the order of .0018" (.046mm) on the X and .0009" (.023mm) on the Y axis.
These test were run using my finest .0005" DTI of dubious origin:})
On my router, I used Hall effect switches and remember getting excellent results but I cannot find the spreadsheet where I saved the data. These switches were based on the thread by RomanLini here on the zone http://www.cnczone.com/forums/open_s...made_easy.html
and used 1/8"dia x1/16" thick magnets of ebay.
Hope this helps
Art
AKA Country Bubba (Older Than Dirt)