Precision Maintenance Missing: Fan Shaft Bearings

Home / Precision Maintenance Missing: Fan Shaft Bearings

This blog is to remind everyone that ‘this is the way we have allays done it’ doesn’t wash and also how important using the correct lubrication and lubricant cleanliness is!


This share has two questions;

  1. What the highest Acceleration levels you have recorded on a fan rotating at around 1498 RPM?
  2. Which bearing do you put as the fixed bearing on a fan shaft?

I appreciate the Accelerometer was only technically good for 50g’s but we had a reading of 116.28g’s Peak to Peak. Can you beat that?


We were requested to inspect a fan due to repeat failures of the fan bearings.

This fan process cold air, it is direct driven at 1498RPM and has two 22222 straight bore double row spherical roller bearings.

The fan NDE (Non-drive end / fan end) bearing was set as the fixed located bearing. The DE bearing at the coupling was set as the float bearing.

This fan had been in operation for 17 days.

Data was collected with a 100mV/g Accelerometer with a flat rare earth magnet.


The vibration data indicated that the motor to fan shaft alignment was good and there were no issues with the Velocity imbalance levels.

There was however extremely high Acceleration levels indicating excessive damage to the outer raceway together with an indication of poor lubricant condition.

Vibration Acceleration Data:

Figure 1 is the Acceleration Time Waveform from the Fan NDE (Fan end bearing). This shows the very high impacting levels with a -52.59g’s peak to +63.69g’s peak.

Fig 1: Acceleration Time Waveform

Figure 2 is the Autocorrelation of the Acceleration Time Waveform. This shows that all this activity is being generated mostly from the bearing outer raceway.

Fig 2: Autocorrelation of PeakVue Data

Figure 3 is the Acceleration Spectrum. This again shows that all this activity is being generated from the bearing outer raceway.

Fig 3: Acceleration Spectrum

Vibration PeakVue Data:

Figure 4 is the PeakVue Acceleration Time waveform. This shows very high general impacting up to 34.9g’s Peak.

Fig 4: PeakVue Time Waveform

Figure 5 is the Autocorrelation of the PeakVue Time Waveform. This shows that all this activity is being generated from the bearing outer raceway.

Fig 5: Autocorrelation of the PeakVue Time Waveform

Figure 6 is the PeakVue Spectrum. This shows that all this activity is being generated from the bearing outer raceway.

Fig 6: PeakVue Spectrum

Vibration Velocity Data:

Figure 7 is the Velocity Spectrum. This confirms that this is a late stage defect and that this energy is from the bearing outer raceway.

Fig 7: Velocity Spectrum


On visual inspection it was found as expected the grease looked oxidised in a poor state and there was a high area of damage to the bearing outer raceway – noticeably on one side of the rollers. Damage to this side of the raceway would have been caused by axial thrust from the fan shaft and motor.

Image 8: This is on removal of the bearing caps. This shows the oxidised poor condition lubricant.

Image 8

Images 9 to 12: These are further images of the grease condition.

Image 9: Grease extracted from the bearing housing
Image 10: Microscope image
Image 11 : Microscope image
Image 12 : Microscope image

Image 13: This shows the two tracks for the rolling elements on the outer raceway, and that it was highly loaded to one side.

Image 13

Images 14 and 15: These are close-ups of the defected area.

Image 14: Microscope image
Image 15 : Microscope image

Summary and Questions:

Researching the lubricant used we know this is not suitable for this application and that it will displace/separate and then oxidise.

But the question is what caused the high thrusting to the one side of the raceway, is it related to what is the fixed and free bearing? Is it true to say that due to the NDE (fan end) being the fixed bearing that expansion from the motor/ fan shaft would load up one side of the raceway?

A profitable plant is reliable, safe and a cost-effectively maintained plant

One thought on “Precision Maintenance Missing: Fan Shaft Bearings”

  1. Hi James, just read your interesting report on this fan NDE bearing failure. I have some thoughts on what may be causing these bearing failures, but can you please confirm this is an overhung fan. Also what type of fan is it? (does it have high axial load?). Your question regarding which bearing should be fixed. Well If it is a direct coupled overhung fan as I think it is, in my experience the NDE (fan end) bearing should be fixed and motor bearing floating. Any thermal expansion will be back towards the motor. This is because NDE bearing will have the highest radial load (from weight of impeller). The problem may be the type of bearing used and not that NDE bearing is fixed. Spherical roller bearings are not good in high axial loads (which you can have in overhung fan arrangements). Overhung fans are best suited to having tunnel housing with angular contact bearings (2 x back to back) at NDE, and ball bearing at DE. regards Andrew

Leave a Reply

Your e-mail address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: