Work with probe

I am using probe screen, by verser
github.com/verser-git/probe_screen_v2

I will use the probe to locate things, but I would expect the measurements to be accurate.
Taking into consideration that the probe is within 0.01 accuracy, how does the final measurement ends up being +-0.1 off.
Can this error be due to some delay between the moment when the probe trips, till the moment linuxcnc processes that input?
I would like to understand the root cause of this.
I have calibrated the runout of the probe, reduced the latch speed to 5mm/min..so I don't understand what could be the cause of the error.

If you are not calibrating it that is the source of discrepancy. There is a delay.
sin you want to measure,
You can cheat the ball diameter or modify the macros to take that into account.

Also you should perform repeat touches untill you find a final probe feedrate that gives the repeatability you want.

vmihalca wrote: I am using probe screen, by verser
github.com/verser-git/probe_screen_v2

That does appear to use the #5061 (etc) probe results then.

Can this error be due to some delay between the moment when the probe trips, till the moment linuxcnc processes that input

It doesn't seem likely, as the probe trip should be seen within 1mS of probe contact. 0.1mm in 1mS would be a probe feed of 6000 mm/min.

Can you try running the probe-hole.ngc sample file? Just to see if that (slightly different) sequence gives a different result.

If you model the probe as a ball on a springy stick, it is behaving as I would expect.

On the inside, it doesn't trip until it has overshot the distance by about .05 mm in each direction.

Personally, I do all my probing with routines that detect the edge as the probe moves away from the edge. Others use routines that approach the edge at high speed, then back off, then measure as they approach the edge at low speed. I've always considered that a waste of time. I just approach at (relatively) high speed, then back off at low speed.

You might give that a try. At least try measuring in both directions to see what the hysteresis might be. (Of course, it you have a lot of backlash, that could explain the whole -- or hole -- problem.)

Ken

Another test, center of bearing with dial gauge, set x0y0. Do 1 point measuring x+ y+ x- y- to find the target-actual difference in 4 directions. This is the calibration, compensation or whatever. X+ inside should be the same as x+ outside. If you align (spindle orientation) the probe always the same for future measuring and use the compensation values.

Hello,

i get a Renishaw probe soon and want to mount it beside my spindle with a mount where i can move the probe up and down.

Is there a setting in the probe screen config to set a offset to the spindle?

Best regards,
Jan