Hi everyone,

I am struggling with a strange problem when measuring a roundness profile of large rotor in a horizontal grinding machine. The rotor is a paper mill where both rotor ends are ground. The rotor is attached between centres (headstock/tailstock). Rotor alignment is supposed to be in tolerance. When I measure the roundness (not exactly the roundness, it is the total radial displacement though) with capacitive non-contact sensors I obtain the displacement waves. 

The problem arises when measuring with at the same time with two sensors. One is located at 0º and the other at 180º in radial direction, both in the headstock end. The measured displacement profile is not the same for both readings. When analysing the data, there is a clear difference in the 2x harmonic component.

While the 1x (runout related) remains the same, the 2x harmonic component disappears for the 180º reading. The 2x is dominant for the 0º reading.  I wonder how is possible to have different readings for 0º or 180º in the 2x component. I have tried with 90º and 2x appears too, but not with the same amplitude as in 0º.

I have tried this in two different machines and two different rotors and the same conclusion applies.

Any thoughts about this?




Original Post

I can't explain the 2xSS behavior, and I am not sure why it would be relevant.

There are several roll properties that could affect your measurements:

Eccentricity -- perfect cylinder but not rotating about centroid

Shape -- non-cylindrical shape rotating about its centroid; such as oval

Surface Condition -- distortions of smaller scale than Shape

Thickness and Material Variations -- anything that affects capacitance measurement

Shaft Bearings -- looseness and wear

Make sure your sensor brackets are not resonant within the frequency range of interest, and that sensor (with bracket) are not getting vibrations from the support structure instead of the relative motion between roll and sensor.



thanks all for your replies

I have used dial indicators as well. The total amount of displacement measured is higher in the 0º measurement than in the 180º. The results are coherent with the proximity probe's readings. The difference is on the 2x component which increases the overall value at 0º.

I agree with Walt it could be the oval shape of the roll. I am going to perform new tests to assure that. My idea is to run the measurement at 0º and 180º and track de 2x component amplitude. Then unclamp the rotor, rotate it 90º and measure again. Repeat the procedure each 90º and track the variaiton of the 2x. 

keep you updated


Are they located on the same spot on the rotating element? as if you were to wrap a string around the whole rotating element, would they be in line?

Its possible that there may be a defect in mounting or on the rotating element if they are not reading the same line on the shaft. 


Add Reply