Profilometer Surface Roughness Tester Issue - Ball Bearings

[Walleye]

We have a Mitutoyo SJ-400 Surface Roughness tester.
My issue is when i check a ball bearing it gives me shotty results like 10 ra when it should be like 3 ra. Our machine gets calibrated monthly with a surface finish master. (CAL 115.0/LIN 19/STYCK 12-20) Do i need to be using a step gage to calibrate this gage properly?
I turned on the curve so it does more than just a straight line.
we really dont have any issues with it outher than this, im just curious if i am missing something.

 

[Michael_M]

I am not familiar with the specific model. Can that model check a rounded surface? All the surface finish gages I have used require the surface to be essentially flat so the small probe tip can move up and down while the 'exterior' moves over the actual surface.

If this particular model can check rounded surfaces, I may have to put it on my 'wish list'

 

[Kales Veggie]

Not enough information, really.

Is it the correct tester for a sphered surface? (skidded?)

Did it work before? Are the results repeatable?

Correlation with other testers?

Is the tip damaged / chipped?

Is the tester calibrated in the same range as it is used?

 

[Walleye]

Yes, it is the correct detector.
It checks skidded and non skidded.
Never checked a ball bearing before, as far as repeatabilty its good.
we have a single tester so no correlation. never had an issue with this unit.
No the tip is not damaged or chipped.
Yes it is calibrated in the range it has been and is being used. I have worked with this profilometer for 12 years, i am confused as to why it wont do this task properly.

 

[Walleye]

Yes this model can check rounded surface, it has R-Surface compensation which alllows it.

 

[Michael_M]

These are the things I would do (you may have already done them):

1. Check the gage against a 'known' surface (round) and see if that comes out in spec.
2. Check the gage tip (see if the tip is damaged, bent, etc).
3. Measure the same surface without moving the gage or the surface being checked 5 times. See if it reads consistently.

Any of these may help pin point the issue. Please, let us know what the issue was if/when you discover it. It may help others later.

 

[crager34]

Maybe a setting needs changed?

 

[Ron Rompen]

Digging back a few years in my memory, I had the same problem (with a different make and model of surface tester). After contacting the manufacturer, we discovered that the tester was not properly setup to do measurements on a curved surface - once we made the necessary corrections, it was fine.

I do recall, however, that we were limited in the minimum diameter of the curved surface that we were measuring - below a certain value, there was just too much 'noise' from the curve of the surface to provide a reliable measurement. Also, check and make sure that you input the 'R' value correctly (and with the right sign, if that is appropriate).