Skip to main content

Hi all

need some Advice on an Overhung fan arrangement.

110kw Motor 1480rpm fixed speed, belt driven. The pulley arrangement is 1:1, making the fan speed at 1480rpm. It has a filter system before it, so it is essentially sucking out clean air which it then is exhausting to atmosphere.  Historical data shows this fan has had high Horiz 1x going back to 2014. Cures to try and fix it back then included balancing the fan and putting anti-vibration mounts on the frame. This has taken the readings down from about 20mm/s to around 8-10mm/s.

so now I'm investigating this and want to know what could be possibly causing such a fault. My initially view is pulley eccentricity. I have used phase analysis, but this wasn't a conclusive find (i measured vertically underneath the bearing on the pedestal, and horiz and got 167 degrees) i then measured top of the bearing and horiz and got 45 degrees. I might be doing it totally wrong!

I Will dti gauge the pulleys in due time, but I haven't had a chance to switch off the fan due to production demands. i have ruled out it being balance issue due to the fairly low axial and vertical readings, am I wrong for doing that ?

The De fan vibration is reading as follow

H: 10.19mm/s at 1x

v: 5.7mm/s at 1x

A:3mm/s at 1x

NDE fan vibrations reading as follows:

H: 7.83mm/s at 1x

v: 3.3mm/s at 1x

A: 2.6mm/s at 1x


Ambient: 25C

De Fan bearing Housing: 45.6C

NDE Fan bearing Housing 40C

  Is there anything I am missing, any other suggestions for possible issues or ways to narrow down what it could be?


Images (3)
  • FFT DE Horiz
  • Waterfall Diagram
  • Waveform Autocorrelate DE fan
Tags: fan, Horiz, 1x, Overhung Fan High

Replies sorted oldest to newest


Have you any phase and amplitude data on the DE of the motor in the same direction as the fan data to compare the motor readings to the fan DE?

Have you any data or can your gather a very low Fmax, say 3x Fan Speed, on the DE of the fan, with very high resolution, in an effort to see if the motor is running the "exact" speed as the fan and confirm all the fan vibration is from the fan and not being increased by the motor's true speed?

Thanks and Have a Blessed Day,



Hi Ralph

I've attached a very basic document with the layout of the machine and the phase analysis for the DE motor and De fan bearing.  The phase data is almost completely reversed for both DE, not too sure of the significance of this though?

I have also attached the High res fft and waveform of the DE fan bearing. When taking accurate rpm data, I used a strobe, and the motor was running at 1486rpm, and the fan was 1483rpm. The FFT shows a peak at 1482.89 rpm, so i would draw the conclusion it is fan related.

The waveform looks like some modulation but then the vibration from the 1485 cpm is 2.29mm/s, surely that wouldn't be creating that much vibration when in comes into phase with each other?


Images (2)
  • DE Fan Waveform High Res
  • DE Fan High Res
Files (1)

Thanks James,

What is the resolution of the spectrum you posted?

I am not sure what your original Maximum Fmax scale was on the spectrum you show, that you "zoomed in" on, but if it (Fmax) were, say 4,500 to 6,000 cpm or 75 to 100 Hz long with 6,400 or more lines, it might show the differences in the motor and fan speeds around the 1x position when zoomed in at 1x. But being approx. 2 1/2 rpm difference, changing approx. every 24 seconds (based on the waveform data), one might miss seeing the change when taking a high Fmax, low resolution spectrum.

The waveform is an excellent pattern. It appears to be showing a modulation present in the data.

There appears to be a 2.5 rpm difference between the fan and the motor showing in the waveform you posted.

Watching a live, high resolution spectrum on the analyzer for a few minutes might would show an amplitude increase and decrease on the fan DE bearing.

This would a test I would perform.

Just my opinion.


@Registered Member posted:

If I were you, first measure runout and alignment of the pulleys, if everything seems ok balance with a lot of averaging.

Some place I read or heard that loosen the belts to induce slipping might help getting a stable phase Reading or speed separation of the fan

Yes i will hopefully be able to try this out at some point soon ! just got to wait for production demands to drop...whenever that may be. I didn't realise that was a think, might be very interesting to find out if that works.

@Registered Member posted:

watching the Live data shows a rise and fall in the amplitude at DE fan Bearing from around 13.5mm/s to 6.0mm/s.

The hi res FFT at 12800 and fmax 3000cpm. This shows two peaks, in slight better definition.

so, the most likely cause is modulation, how can I reduce the effects of this?


I would suggest, to supervision, separating the speed of the fan from the speed of the motor, from being a 1 to 1 ration, by changing the pulley size. If it were direct coupled to the motor, this problem would very possibly, not be there.

Surely, with it being an air exhaust fan, the fan could be slowed down or increased by 20-50 rpm's or more, unless it must rotate, by law or design, at a minimum of 1480 rpm. If the fan must run a minimum of 1480, I would suggest increasing the fan speed greater than the motor speed.

It is going to be difficult, in my opinion, to solve this problem with a belt drive system and a pulley ration of 1 to 1.There are too many variables involved with a belt drive 1:1 speed. One variable is, the motor speed being "uncontrollable" due to the "load" from the fan, with a 1:1 sheave ratio.

After separating the speeds, balancing the fan, if needed, would be quite simple.

Only my opinion and I could be wrong.

Thanks and Have a Blessed Day,



Just thought I'd add what I've tried so far. We managed to find a gap for an hour's downtime and I did as many tests as I could in that time.

Firstly no imbalance from the fan, shaft check with dti gauge and it's straight. Pulleys in good condition.

Motor checked and it has a stiff rotation before becoming loose and then tightening up. Think this may be where the problem lies.

Attached is a motion Amplification of the fan


Videos (1)
Fan motion Amplification

Add Reply

Link copied to your clipboard.