Homing Accuracy

[Alex S.A]

Hi All,
What is your advice to improve the homing machine, making it more accuracy?
I am having problems with that coz the homing is not kept at the same position. I checked that by making a mark on the longitudinal rail and after a couple hours with machine working I do the homing again and the machine stopped in a different position.
I am using a Inductive sensor as homing switch, a incremetal encoder and it is a gantry machine using as plasma cutting tables.

What Can I do to improve it?

Thanks,
Alex S.A

[gar]

061201-1114 EST USA

Alex:

If your incremental encoder is on a lead screw, then you need a once per revolution mark on your incremental encoder or add a second encoder that is only a once per revolution unit.

If you have a linear incremental encoder, then get one with a second channel with widely spaced marks (equivalent of once per revolution on a rotary encoder). Otherwise your need to create a very fine resolutution sensor on your axis.

Once you have the once per revolution marker, then set your coarse limit switch to normally trigger about 1/2 way between the once per revolution occurances. Now correlate the two functions for your zero reference.

.

[Geof]

There is some discussion in this thread;

http://www.cnczone.com/forums/showth...ght=home+limit

[ViperTX]

Well....I guess the big question....you need to quantify the amount of variation from the original home position.

[Al_The_Man]

What you need to do is ascertain if the error happens at any other time, IOW is something mechanically changing when running, for e.g. a loose encoder/coupler/rack gear etc.
Anything like this will show up as parts not retaining true dimension over a period of time of running.
If the machine only exhibits this when homing, then it points more to something occuring only at homing time.
Does the homing routine stop the machine progressively in one particular direction?
Al.