Fan Vibration


Your 208 Hz vib is likely from one of the 6220 fan bearings.  Not knowing what brand bearing is installed, a lookup using the SKF tools put a radially loaded (contact angle of zero degrees)  ball pass outer race defect at about 4.09x turning speed.  With the combination of bearing "tube" type mounting, overhung fan wheel unloading the coupling end bearing, and likely thrust load from the single inlet fan, the contact angle is likely not zero degrees.  It is very likely that the BPFO could be 4.18x turning speed, or even a bit higher in frequency.  Couple this with its origin coinciding with changing bearings and its a pretty sure bet.

The 1x objectionable vibration can come from many sources and interactions, as I am sure you already know.  Some things to consider are:

a.  The bearing/wheel arrangement may tend to under-load the drive end bearing creating an instability.  This will contribute to the family of turning speed harmonics in your spectra.  Have you examined the TWF?

b.  Your fan appears to be on an isolation system.  Have you ensured the duct work (inlet/outlet) and as well conduits, piping, etc. provide enough freedom of motion so that the isolation system is not short circuited?

c.  What is the design vertical natural frequency of the fan unit installed on the isolation system?  If you test this in the field with some form of impact/response testing, etc., does it seem to work correctly?

d.  I am sure you have already, but have you done checks to ensure the shafting is true with limited TIR and as well performed trim balance in the field?

Good luck!

Brian :-)


Please excuse my use of abbreviations.

TIR = Total Indicated Reading as when using a dial indicator measurement device to determine the concentricity of a surface of the shaft vs. its centerline of rotation in the bearings or other supports.  Usually measurements would be in 0.0010 inch units or microns.

TWF = As you said, the Time Waveform plot.  I belive ISO would refer to this as the "Time History", but to most, it is the amplitude vs. time plot of the measured vibration at sample rates as high or higher than the associated spectra collections and as well of a time duration long enough to capture multiple revolutions of the shaft or multiple events of interest. 


Hello Ismail,

First, if we want to try to really determine fault frequencies, we need the actual motor turning speed at the time of vib collection.  The 2985 RPM may be the actual turning speed, or the nominal full load turning speed.  Was it measured?  A small error here can result in some calculations not matching up.

Second, thank you for supplying the bearing table.   It's factors are calculated at a contact angle of 0 degrees.  As well the theoretical calculations are assuming no slip or other anomalies in the bearing; all uncertainties we accept.

Based on the number of balls, the ball diameter and the pitch diameter that you supplied, I calculated the BPFO factor and frequency (at 49.75 Hz) at different contact angles.  On a lightly loaded bearing, it is pretty easy to have contact angle which is not zero, but maybe 20-30 degrees due to thermal growth, axial forces (single suction fan), etc.

00 deg    4.093x    203.61 Hz
10 deg    4.106x    204.30 Hz
20 deg    4.147x    206.34 Hz
30 deg    4.214x    209.66 Hz

The difference of the 208 Hz from the zero degree contact angle freq of 203.61 Hz is not very large and easily within the error of contact angle and perhaps actual turning speed.  Coupled with the 208 Hz vib appearance coinciding with the change of the bearings, makes the bearing(s) even more suspect as the source cause.

Just my thoughts..


Brian :-)

4/ Based on your Spectrums, I would replace the Fan Bearings again. Continue to "Trend" the Fan Bearings until you reach an optimum time to schedule replacement.

Be certain to pay particular attention to the Outside diameter of the bearing and the

inside diameter of the Bearing Housing. If the Bearing is loose in the housing, the

movement can produce similar Spectrums. (3X appears more often 2X, but I've seen similar Spectrums on Monoblocks (fan bearing housing)

5/ Check fan bearing Housing (monoblock) End Thrust with a Dial Indicator.

     Typically .001 -   .003.   But you need to check with specific manufacturer, 

     And get correct end thrust tolerance for your unit.

6/. Measure vibration in 3 directions at all 4 Bearings. Shut fan off, lock out, Mark coupling with paint marker, unbolt coupling and rotate 90 degrees. (We usually start with keyways 180 degree apart, or locate original match marks if they exist) Get the vibration as low as you can using the coupling and than balance the fan wheel as low as you can time permitting.

7/  Record final set of readings and begin New Trend. Let us know the outcome

Thank You



Add Reply

Likes (0)