center point changes position after A-B rotation of head

[Paul Schuyler]

We have a machine that is failing a positional ring gauge test. The gauge stands upright on the granite with the cylinder aligned +Y / -Y. We measure the gauge as a cylinder with HH-AS8-H2.5 / HP-S-X1H probe head (5x50mm tip) rotated to +Y.
The software is aligned:
1) Level to face of ring gauge
2) No rotational reference
3) X and Z origin to center of cylinder
4) Y origin to face of ring gauge.
We then reverse the direction of the probe 180d to face -Y. We do not change the alignment. We remeasure the same points on the cylinder.
The diameter and form of the cylinder are within expected tolerances, but the X and Z positions change by up to .001”.
I’ve seen this before on other machines and was unable to fully resolve the issue.
Any ideas? Thank you.
Paul

[JEFMAN]

Did you calibrate both positions at the same time ?
What is the distance between the calibration sphere and the gauge ?

If the x and z deviations are constants whatever the position of the gauge, it should be a calibration defect.
If the deviations change, then it's could be a cmm calibration problem (dx = Y yaw, dz = Y pitch).
You can easily check it bi positionning the gauge center at the same location than the calibration sphere, after calibration.
remove the sphere and place the gauge at the right place...

[Cumofo]

Well, a long time ago that meant the night shift guy crashed the Ph9, and it rotated around in K about a .500 hardened shaft that attached the wrist to your mast. But now that has been replaced by a square block to minimise that.

I think think the CMM took a good hit when you went home early Thursday. I just replaced a sensor body that was quackered .120 in X minus.
here's how to diagnose that:

Set up 2 Heavy 2-4-6 blocks 8 inches apart in the failing axis.

Loop 2 points with a point info call out 100X, one hit in Y plus, followed by a hit in Y Minus.

You should repeat waaaayy less than 1 micron stddev. between each point in the same vector. B89 passing is 3 microns linear for most machines.

if it looks bad:

Measure the location of your Qual sphere with existing sensor,

remove tip
replace sensor with another sensor

replace tip
remeasure qual sphere

the delta is the deflection from the quill z axis in the XY plane of you old sensor.


I know theoretically the .120 offset in the sensor a CMM can easily compensate for, but in reality and practice, IT DOSE N'T DO IT WELL AT ALL!

Circles wont measure within .002 or so, and true position deviations appear double.

(If you think the Z axis is bad you can loop 100 points with point info VECTOR 0,0,1 to test for crash damage in Z too.)

[Paul Schuyler]

Thank you JEFMAN and sigpic.
Let us try some stuff.
And in your opinions, can you realistically expect to flip the probe 180d as described in the ring gauge test above and get repeatable numbers within, say, .0002" (.005mm) of the first position origins?
Much appreciated.
PaulS

[Mike Ruff]

If everything is set up correctly, and you are taking the hits in the exact same locations from both sides, I would expect less than .0002" between sides.

[JEFMAN]

I usually get less than 2µm with a LSPX1/L50D5 (A90B0 - A90B180).
I tested with a star probe (LSPX5 -Tip 1 :L90 D8- origin - Tip 2 L510 D8) and get less than 4µm at the best position (with very long tips, the cmm defects appear differently depending on the position on the cmm, and on the position from the calibration sphere)

[benj118]

Try leveling to the cylinder instead of the face, face could possibly be throwing this off.