Case History Series

Small Motor - Soft Foot Frame & Bearing Problem

I. BACKGROUND

The production personnel of a South Carolina chemical company reported hearing a new noise emitting from a pump/motor unit. Maintenance was dispatched to investigate the noise and repair if necessary. The maintenance mechanic and electrician assigned to the job requested vibration diagnostics to determine the source of the noise and health of the machine.

II. CONCLUSIONS

A. A soft foot on the motor is warping the frame.
B. A bearing outer race defect exists at position 1.

III. RECOMMENDATIONS
A. While the motor is operating, loosen the foot hold-down bolt on the southwest foot. The foot will rise due to the warped frame. Temporarily (until scheduled shutdown) insert a shim the thickness of the gap below the foot then retighten the bolt. At scheduled shutdown, determine between sprung foot (as in Figure 1) and soft foot. Soft foot will require proper shimming; sprung foot may require milling.

B. During scheduled shutdown replace motor bearings.
C. During scheduled shutdown inspect the pump base for evidence of settling.

Figure 1. Illustration of a sprung foot

IV. DISCUSSION OF RESULTS
Manufactured by Reliance, the subject motor has an operating speed of 3580 rpm, 25 HP, and is outfitted with NTN6310 rolling element bearings. It is used to power the re-circulation pump for a 50% caustic soda storage tank. The driven pump is a Durco I horizontal centrifugal pump. Figure 2 is a sketch of the unit showing measurements positions.

Figure 2. Horizontal centrifugal pump & motor

A. Warped Motor Frame

An FFT analyzer was employed. The reading taken at position 2V showed a peak at 2X line frequency (2F_L). The peak clearly dominated the spectrum with an amplitude of .5 ips (peak vibration). Remember, "Concern should be given motors whose amplitude at twice the line frequency (2 X 3600 cpm in the USA) exceeds approximately .100 ips peak for in service motors… This would apply to motors ranging in size from approximately 50 HP to 1000 HP"₁. Our analyzer showed five times .100 ips on a motor smaller than 50 HP (25 HP). The survey three weeks prior to the diagnostics testing showed an amplitude of .305 ips or three times .100 ips (the level at which we should give concern). See Figure 3. The amplitude had increased 64% in just three weeks. A problem was not only present, but also developing.

Figure 3. Spectrum showing high amplitude at 2X line frequency

Such high amplitude at this frequency might lead one to quickly conclude that the motor is experiencing an electrical problem because many electrical problems are associated with 2F_L. However, mechanical problems should also be considered when motors show high amplitudes at 2F_L.

It is commonly agreed on that mechanical problems can cause high amplitude at 2F_L. A warped motor frame for example. The warp can cause an uneven distance radialy between the rotor and the stator’s magnetic field. At the location where the distance is reduced due to the warp, a strong pull may occur relative to the location where the distance is increased due to the warp. Both events (strong pull/weak pull) happen every rotation of the stator’s magnetic field. The stator’s magnetic field rotates at the line frequency of 3600 cpm (in the USA) X 2 events = 7200 cpm. The same frequency involved in many electrical problems, yet in this case caused by a mechanical problem.

When the mechanic began to insert the shim under the southwest foot (per recommendation 1 on page 1), he noticed a hairline crack propagating from the southwest foot and extending to the motor frame. The crack measured several inches in length. This discovery added much weight to the conclusion that the motor frame was warped and that the high amplitude 2F_L was the vibration result generated by the warp. A decision was then made to remove the motor, repair the crack, determine between sprung and soft foot, and replace the bearings. Figure 4 is a post repair picture of the hairline motor frame crack.

Figure 4. Frame crack discovered on motor (Photo taken after weld repair.)

The mechanic determined that a soft foot coupled with improper shimming twisted the frame and caused stress near the foot area. The stressed area was acted on by high amplitude vibration and resulted in the hairline crack.

B. Bearing Defect Outer race
To determine the source of the noise reported, previously taken velocity and acceleration envelope measurements were examined. Figure 5 is an acceleration envelope spectrum taken at position 1H. A band pass filter of 30k – 600k cpm was used to eliminate rotational vibration. The resulting spectrum clearly displayed multiples of BPFO indicating that a bearing problem existed.

Figure 5. Spectrum of Acceleration Envelope Measurement position 1H

To help determine the severity, the velocity spectrum (also previously) taken at the same position was considered. A bearing defect of moderate severity will usually display defect frequencies in the velocity spectrum. Figure 6 is the velocity spectrum, which displays many multiples of BPFO. A sideband cursor was placed over one of the defect frequencies. Note that the delta of 10962 is very near the BPFO fundamental defect frequency of 10920.

(BPFO= 3.07 X 3557 cpm running speed = the fundamental BPFO frequency of 10920 cpm).

Figure 6. Spectrum of Velocity Measurement position 1H, showing many multiples of BPFO

Having so many multiples of BPFO, this spectrum fits the description of a stage 3 bearing failure as described in the Technical Associates Illustrated Vibration Diagnostic Wall Chart. It also indicates that the source of the noise reported originates with the bearing defect.

As per recommendation 2 (on page 1), the motor bearings were replaced. To confirm that a defect actually existed, the outboard (position 1) bearing was cut open and visually inspected. Figure 7 is a picture of the outer race.

Figure 7. Outer race defect found at Position 1.

V. RESULTS OF CORRECTIVE ACTIONS

Motor repairs encompassed the welding of the frame crack, proper shimming/alignment of the unit, and the replacement of the motor bearings. A significant drop in amplitude at 2F_L resulted from proper shimming/alignment, and repair of the crack.

Compare the before and after repair spectra in Figure 8.

 

Figure 8. Waterfall plot displaying measurements taken before and after repair

The amplitude at 2F_L dropped from .305 ips to .026 ips (over 91%). In addition, bearing frequencies disappeared.

Figure 9. Acceleration envelope spectrum taken after motor repairs.

Note that no bearing defect frequencies are present.

It is understood that the motor continues to run with low to moderate amplitudes of 2F_L and no noises have been reported.

--------------------------------------------------------------------------------

1 - Technical Associates’ Concentrated Vibration Signature Analysis and Related Condition Monitoring Techniques (Level II Seminar Book) © Copyright 2006 Technical Associates Of Charlotte, P.C.