|
Post by lseguine on Nov 30, 2012 14:59:12 GMT -8
Hay Dave, got a strange thing happening and not sure what it is or if it is a problem or not. i ran calabration on my probe in a machined hole measured to be true x and y by less than 0.0005. the output said what the corection should be and i accepted it. probed the hole again and found the system said the hole was smaller that its actual size by about 0.015. that the X reading was very close while Y was off by about 0.013. so thinking back on things that i had done i thought that mabey my backlash might be a source for the differance. i'm having to run back lash at 0.016 for X and 0.022 for Y to get acurate circular cuts. i changet the Y back lash to .013 and got very close X Y values for the hole. big smiles...... then i cut a circle in a pice of wax and low and behold my old oblong hole returned put the backlash back, nice round hole. any thoughts on whats going on? i'm sure i can figure out a work around if necessary. puzzeled in alaska Larry
|
|
|
Post by DaveCVI on Dec 1, 2012 13:55:12 GMT -8
Larry, I'm going to have to think about this one (and probably talk with Brian and gang) - I'm uncertain exactly how mach handles backlash compensation internally.
My understanding is that using software compensation for backlash is generally considered a last resort - it's much, much better to take the backlash out of the machine mechanically (Probably not what you want to hear).
Dave
|
|
|
Post by lseguine on Dec 1, 2012 17:26:33 GMT -8
well not really .. i did what was sugested by hass in his site and changed balls in the ball nuts to a larger size but i still get some backlash not sure where i can tighten things up more any will keep an eye on this and see what you come up with larry
|
|
|
Post by DaveCVI on Dec 2, 2012 9:38:40 GMT -8
Larry, Two thoughts re some things to check:
1) are the axes steps per unit calibrated? I.e. measured for actual movement, not just using calculated values. the process for checking axis calibration is in the mach install and config manual. what you're looking for is if a (for example) 1" commanded movement actually moves 1" or not.
2) are you mounting the probe in the same orientation wrt to the spindle each time you use it? Some probes will have a different trigger latency in X vs Y. One way to check for this is to probe to a fixed block in x, get the trigger position. then rotate the spindle (and hence probe) 90 degrees and do it again - same position result each time?
Dave
|
|
|
Post by DaveCVI on Dec 3, 2012 10:55:57 GMT -8
Larry, I believe that backlash comp is handled in the low level motion device for mach (PP or SS etc), so would you please refresh my memory: Are you running on the mach Parallel port driver or a Smooth stepper or ?
Dave
|
|
|
Post by DaveCVI on Dec 11, 2012 11:38:03 GMT -8
Larry, Brian conformed what I thought: "Backlash is done in the motion device so we don't have to worry about it ... it tracks the direction and adds the steps... " MSM uses the reported positions given to it by Mach (and Mach depends on the motion device layer to handle backlash). I don't have an explanation for what you reported. Dave
|
|