The SDT270 system was used to collect ultrasound data at a methane plant for start-up and commissioning of newly installed equipment to get a baseline condition on the motor’s and gear reducer’s bearings and gears. We were immediately able to pinpoint a defect in the drive end motor bearing using the SDT270 as well as collect and record a sound file (play below) and present it to the customer.


After removal and inspection of the motor bearing it was found that the motor had been stored in a basement and got water into it through the electrical conduit during a monsoon storm that flooded the facility in late summer, causing the bearings to rust and corrode. It was also found that the electrical disconnect box for the motor had water in it as well.

Pump: Progressive Cavity Seepex Pump

Motor: WEG 75HP 1775 RPM

Not only were we able to identify a bearing fault but also the “smoking gun” root cause, thereby thwarting further damage in the electrical system.

After troubleshooting the entire facility we were able to save several other motors stored in the basement that had also had water exposure through the electrical conduits.

Thank you Brian Franks with JetTech Mechanical LLC for sharing this success story with us!

by Ana Maria Delgado, CRL

Detecting rolling element bearing defects can be difficult if a few things are not taken into consideration.  First, consider how the bearing fails.  Normally one of the races will begin to fail followed by the roller/ball and then the cage.  Initially these failures do not show up at the calculated bearing fault frequencies, but at multiples of around 5× to 7× or even greater.  At first you may only see a minor peak at say 7× the inner race defect frequency with side bands on one or both sides spaced at shaft rotational speed.  The time waveform will probably show the most changes initially with increases in the high frequency content of the data when viewed in terms of acceleration in G’s.  As the failure progresses you begin to see more harmonics of the defect frequency in the FFT spectrum such as 5×, 6×, or 7×, with changes in the time waveform data possibly showing what is often referred to as an angle fish pattern.  This is due to the rolling elements impacting the defect in the race, ringing it like a bell and then the energy decaying before the next impact. 
As the defect continues to worsen it will reach a point where the energy in the FFT will reduce to where the peaks are no longer present, but you see mounds of energy resembling broadband noise.  This is due to the clearances in the bearing opening up and the excessive clearance beginning to resemble more of a looseness pattern.  Also remember that the generated frequencies not always (and in fact ) rarely match the calculated frequencies of the bearing.  This is due to the calculated bearing fault frequencies coming from known bearing geometry.  When the bearing starts to deteriorate the geometry begins to change and the actual frequencies generated don’t match the calculated frequencies.  However, the patterns will match and be very close.

by Gary James CRL