Understanding Slow Roll Runout in Electric Motors

Papa Diouf, P.E. Baldor Electric Company

101 Reliance Road Kings Mountain, NC 28086 USA.

papa.diouf@baldor.abb.com

Bryan Oakes Baldor Electric Company

101 Reliance Road Kings Mountain, NC 28086 USA.

bryan.oakes@baldor.abb.com

Abstract – Vibration measurement in rotating components is a crucial

function in the monitoring and diagnostic testing of Electric Motors. This is performed with non-contacting proximity probes such as eddy current proximity probes. These probes continuously measure a varying gap voltage between the shaft and the probe tip. The variation is mostly due to vibration, but also includes mechanical and electro-magnetic defects in the shaft probe track called slow roll runout. This paper discusses the different types of runouts, the different measurement methods and instruments used, the acceptable levels according the American Petroleum Industry (API) standards, the various contributing factors to high levels of runout and how to mitigate it, and their impact on the vibration measurement.

Index Terms – Electric Motors, slow roll, runout, API 541, vibration, compensation.

I. INTRODUCTION

Vibration measurement of rotating components is an important function in the monitoring and diagnostic testing of Electric Motors. Vibration measurement (radial shaft movement) in rotating components is usually performed with non-contacting proximity probes such as an eddy current proximity probe. The non-contacting proximity probe is part of a transducer system that also includes an extension cable and a proximitor. This system measures the gap voltage variation between the probe tip and the probe track on the rotating element. This gap continuously changes mostly due to the shaft vibration, but can also include out of roundness of the probe track, concentricity between the probe track and the bearing journal, surface defects on the probe track area, shaft misalignment, shaft bending, and variations in the electro-magnetic properties of the shaft material near the circumference of the probe track area. All of these non-vibration dependent changes of the gap between the shaft and the probe tip are defined as Total Indicator Runout (TIR) or simply runout. This runout will appear in the vibration readings and can lead to measurement errors [1], therefore, understanding runout is crucial to the monitoring and diagnostics of rotating machineries.

II. WHAT IS SLOW ROLL RUNOUT

Slow roll as defined in API 541 5th Edition section 6.3.3.3 [2] is a condition in oil film bearing motors or generators in which the rotor is rotating between 200 to 300 rpm. At this speed the dynamic effects are minimal and the vibration is almost non-existent. In this condition, all the proximity probe readings should mostly include the mechanical defects at the probe track (out of roundness, surface finish, concentricity between

the bearing journal and track area, and non-straight shaft) and the electromagnetic defects in the shaft material.

Runout can be divided into two components whether it is measured during a slow roll condition in the assembled machine or measured on the rotating assembly positioned in v-blocks on the bearing half shells, or measured in a lathe: mechanical runout and electrical runout.

• Mechanical runout (MRO) is a measure of the shaft cylindrical surface deviation from a perfectly round surface and concentric with the bearing centers. These deviations include: out of roundness of the surface, mechanical defects on the surface (surface finish, scratches), lack of concentricity between the surface and the bearing journal centers resulting from manufacturing, assembly or changes during operation. The mechanical runout can be measured with a dial indicator or a contacting probe.

• Electrical runout (ERO) is a measure of the shaft surface electrical conductivity and magnetic permeability variation. Non-uniform shaft electro-magnetic properties interfere with the magnetic field of the proximity probe, thus causing a change in the processed signal as a gap voltage variation.

Additional potential errors:

• Ensuring the proper operation for a transducer

system requires the components to be matched sets; for example, a 5 meters probe is designed for use with a 5 meters proximitor system only. If these components are not matched properly the measured vibration amplitudes will not be accurate.

• Proximity probes are as a default calibrated to AISI 4140 steel. If the steel is significantly different, it could impact the accuracy of the measurements. The proximity probes can be calibrated to other materials if necessary.

III. HOW IS SLOW ROLL MEASURED

1. The inductive coil is excited with alternating current which creates an alternating magnetic field [3].

2. When a changing magnetic field interacts with a conductive material (such as the shaft), small currents called eddy currents are induced in the material.

Principle of Eddy Current Proximity Probe:

Figure 1: Eddy Current Principle

3. The eddy currents in return create an opposing magnetic field resisting to the original magnetic field.

4. The interaction between the two magnetic fields is dependent on the distance between the probe tip and the target material.

5. As the distance varies, the change in the interaction between the two magnetic fields is converted into voltage output.

6. The voltage output is then converted into vibration units of displacement in mils or microns. One common mounting configuration consists of two eddy current proximity probes mounted on the bearing housing and located at 90° apart and at 45° from the vertical shaft centerline.

Figure 2: Probe Orientation

These probes can be mounted inboard or outboard of the bearing journal depending on the motor design [4]. The probes are placed over the shaft in an area that has been specially machined and adjacent to the bearing journal. This shaft area, called the probe track zone, is specially machined in order to minimize the mechanical and electrical runout. The width of the track zone is dependent on the probe tip size. A minimum width of one and half times the diameter of the probe tip is recommended for the track zone, this will ensure

that the induced magnetic field from the probe tip fully penetrates the specially machined area. API 541 requires that the slow roll runout be measured during coast down when the rotor speed is between 200 to 300 rpm. At this speed range, all of the probe recorded displacement is almost purely runout without any vibration. On non API motors, the slow roll runout can be recorded at approximately between 10 to 15% of the operating speed [4]. The total runout is then recorded and must meet the required limit set forth by the motor specification.

IV. WHY AND WHEN IS IT NECESSARY

Slow roll runout is relevant only on motors with

hydrodynamic sleeve bearings. On hydrodynamic bearings electric motors, the shaft vibration measurement is the best tool for monitoring and diagnosing vibration issues. Because of the damping in the oil film, the shaft vibration transmissibility to the bearing housing is reduced, especially if the motor is flood lubricated.

On anti-friction bearing motors, nearly all of the shaft vibration is transmitted to the motor brackets trough the bearings, therefore the motor vibration can be monitored directly on the brackets using transducers or accelerometers.

High shaft runout can lead to inaccurate vibration readings, because slow roll runout is independent of the shaft vibration. The vibration measured during operation includes the shaft runout which could increase or reduce the recorded vibration.

If the vibration reading is higher than the true machine vibration, then unnecessary alarm or shut off conditions can be triggered. This could lead to a costly shut down for unnecessary troubleshooting and repairs.

On the other hand, if the vibration reading is lower than the true machine vibration, then premature failure can occur leading to costly downtime for repair.

Slow roll runout is a standard requirement on API motors when non-contacting probes are specified.

The scope of API 541 covers the minimum requirements for special purpose form wound squirrel cage induction motors 500 HP and larger for use in Petro-Chemical applications. Unless otherwise specified, oil film bearings are by default used in API motors. In this specification, all hydrodynamic bearing motors intended to operate at speed greater or equal to 1200 rpm, shall be equipped or have provision for non-contacting vibration and phase reference probes. When vibration probes are supplied or provision for probes are required, a probe track area must be supplied and treated so that the total combined mechanical and electrical runout does not exceed a certain limit.

V. WHAT LEVELS ARE ACCEPTABLE Electric motor manufacturers refer to customer

specifications to determine the acceptable levels of slow roll runout. API 541 limits the slow roll runout to 30% of the allowable unfiltered vibration peak to peak (1.5 mils), or 0.45 mils for induction motors. This limit applies to an assembled motor. The slow roll is measured on the shaft probe track of the assembled motor while the shaft is held at its axial

magnetic center and while coasting down between 200 and 300 rpm. If the runout limit is not met during the manufacturing process or initial testing, the motor will be dis-assembled and the shaft reworked. This process can be time consuming and costly and will impact both the motor manufacturer and the customer. Usually, in an effort to save manufacturing time, motor manufacturers partially assemble the motor (see Figure 3) and perform a quick test to check the slow roll runout, the bearing alignment and temperature. If the slow roll is within the limit, then the motor will be finish assembled before the complete testing is started. If the slow roll is not within the limit, then less time will be required to dis-assemble and rework the shaft than if it was fully assembled.

Figure 3: Partially Assembled Motor

For this reason, API 541 standards has set a runout limit on the rotating assembly (rotor and shaft assembled) while supported in V-blocks. With this method, the allowable combined mechanical and electrical runout limit is 25% of the unfiltered allowable vibration limit peak to peak (1.5 mils), or 0.375 mils. Keeping the runout within 0.375 mils increases the chances of achieving the desired limit with the motor assembled.

Figure 4: Rotating Assembly on V-blocks

Figure 5: Rotating Assembly on V-blocks Close Up

However, a rotating assembly can have a very low runout on V-block and yet still exceeds the limit after the motor is assembled. Some contributing factors include misalignment

caused by cocked bearings, non-concentric frame bracket fits, bent rotor during assembly, damaged probe track area, etc. Some motor manufacturers go further and self-impose a combined mechanical and electrical runout limit that is much lower (less than 0.25 mils) on the shaft bearing journal and probe areas during manufacturing in order to avoid finding out a runout issue late in the manufacturing process that leads to costly reworks and delays.

VI. PROXIMITY PROBES LOCATIONS

Proximity probes are rigidly mounted into the bearing housings and are positioned over the shaft probe area. This probe area can be inboard or outboard of the bearing journal, however, preferably adjacent to it. This area is specially machined to minimize surface finish, out of roundness, concentricity with the bearing journal, and variation in the electromagnetic properties of the shaft material. These probe track areas per API 541 must be free from stencil marks or any other surface defects, they must not also be metallized, sleeved, or plated.

VII. IMPACT ON VIBRATION

As stated previously, slow roll runout is noise in the vibration readings. It can be added or subtracted as a vector from the vibration readings, compensated vibration. In the past, simple arithmetic subtraction had been used to compensate vibration levels from slow roll runout. If the vibration amplitude was 1.6 mils peak to peak and the slow roll runout was known for example to be 0.45 mils, then (1.6 - 0.45) = 1.15 mils was considered the true vibration. This method is incorrect because both the vibration and the slow roll runout are waveforms and cannot simply be added or subtracted without filtering them. The unfiltered vibration contains all the frequency components that are in the incoming signal. When a vibration signal is filtered at a particular frequency for example at running speed, it is expressed in amplitude and phase angle which can be described as a vibration vector. As a vector, the filtered vibration at a given frequency such as 1X or 2X, etc. can be compensated with the filtered slow roll at the same frequency as a vector addition.

Per API 541, the compensated vibration displacement filtered at running speed frequency (1X) shall not exceed 80% of the unfiltered limit. Compensation is not used in general by motor manufacturers, but can be useful in certain situations. It can be beneficial to use the compensated vibration if it helps in meeting the limit and avoid costly rework and project delay. Compensation could also increase the vibration depending on the angular position of the vectors. See example below.

Example for calculating filtered compensated vibration: The filtered vibration limit at 1X is 80% of the unfiltered

vibration, which is equal to (1.5 x 0.8) = 1.2 mil. The probe readings for the filtered slow roll runout and

vibration at operating speed are respectively 0.35 mil at 65° phase angle (represented as vector V1) and 1.4 mil at 75° phase angle (represented as vector V2), see figure 5.

V1 = 1X slow roll runout

V2 = 1X vibration at running speed

Figure 6

Without compensation, 1.4 mils is above the API limit, therefore this motor would be considered not API compliant.

Now, let's calculate the compensated values: New Vibration Amplitude = V2 - V1 = [(1.4 cos (75°) -

0.35 cos (65°))2 + (1.4 sin (75°) - 0.35 sin (65°))2]1/2

= [(0.21)2 + (1.03)2]1/2 = 1.05 mil

Phase angle = Arctan (1.03/0.21) = 78.5° The compensated filtered vibration (1.05 mil) is less than

the limit of 1.2 mil, therefore the motor is compliant to API vibration limits.

What if the filtered vibration V2 was 1.1 mil at 265° phase angle?

Figure 7

Vibration Amplitude = V2 - V1 = [(1.1 cos (265°) - 0.35 cos (65°))2 + (1.1 sin (265°) - 0.35 sin (65°))2]1/2

= [(-0.24)2 + (-1.41)2]1/2 = 1.43 mil Phase angle = 258° In this case the compensated filtered vibration (1.43 mil) is

more than the limit of 1.2 mil, therefore not compliant to API vibration limits. The slow roll had the opposite effect in this case than the previous due to the position of the phase angles.

In vibration monitoring systems with compensation capability, the initial slow roll profile is saved in the system and used to calculate the true vibration because slow roll cannot be monitored at running speed. Therefore, the compensation

method is useful in machine diagnostics only and is not recommended for continuous monitoring system as runout changes overtime [1].

Example of a two pole motor with high vibration: This is a two pole motor in which the vibration failed on one

probe at filtered running speed. The overall vibration was 1.485 mil, met the API limit and the filtered vibration was 1.363 mil at 234° phase angle on 1 probe. The slow roll runout was 0.262 mil overall and 0.178 mil filtered at 295° phase angle on the same probe. See figures 9 and 10.

Figure 8

The compensated filtered vibration was 1.285 mil at 227° phase angle, did not meet the API limit, therefore the rotating assembly needed to be reworked. This required the motor disassembled and the shaft reground and reburnished.

The vibration after rework met the API limit both overall and filtered 1X, see figure 11.

If the compensated vibration met the API limit, that would have saved rework time.

Figure 9: Probe 1 Trend Plot

Figure 10: Probe 1 Slow Roll Trend Plot

Figure 51: Probe 1 Trend after Rework

VIII. WHAT IMPACTS THE RUNOUT LEVEL

Runout can be divided into two components, mechanical and electrical. Mechanical runout is the measure of the shaft deviation from a perfectly cylindrical surface. It is mainly impacted by the manufacturing process of the shaft, the assembly process of the various components of the rotating assembly and the changes overtime during the operation of the motor. The improper selection of the cutting tools and/or the machining parameters can lead to higher surface roughness. Mechanical damages (scratches, scoring, ding etc.) on the bearing journal or probe track will affect the mechanical runout. Since runout is measured in reference to the bearing journal, a non-concentric probe track to the bearing journal will result in high MRO. A straight shaft pressed into a bowed rotor or a bent shaft pressed into a straight rotor will increase the runout as well. A misaligned shaft caused by improper fit between the motor frame and the bearing cartridges will also affect the runout level. A sagged or bowed rotor due to thermal instability in the rotor will have an effect on the MRO overtime. A controlled manufacturing process from the machining of all the components to the final assembly is crucial to minimizing the MRO.

Electrical runout is the measure of the non-uniformity in the shaft material. When ERO is measured using non-contacting eddy current probes, the interaction between the emitted magnetic field and the induced magnetic field is converted into distance. Any phenomenon that can change the magnetic interaction between the probe tip and the shaft will affect the runout. This includes a non-uniform material grain structure, non-uniform electro-magnetic properties and a magnetized shaft. The manufacturing process of the shaft, whether it's a forging or a hot rolled steel process can affect the metallurgical properties of the material and consequently the ERO. How to minimize ERO: Case: k0316/0222 Shaft runout on V-block before assembling to the rotor: The probe track has been ground and burnished. The TIR is 0.13 mil on the DE and 0.22 mil on the ODE.

Figure 62: DE Shaft Runout

Figure 73: ODE Shaft Runout

After the shaft is assembled to the rotor, the runout was checked and the result is below figure 13. Notice the ERO has a high peak at 300°. The ERO is 0.45 mil higher than the limit of 0.375 mil. The ODE on figure 14 shows an ERO of 0.375 mil.

Figure 84: DE Rotor Runout

Figure 95: ODE Rotor Runout

After the rotor has been reburnished to smooth out the peaks at 300°, the final ODE total runout was 0.33 mil.

Figure 106: DE Rotor Runout after Reburnishing

IX. CONCLUSION

Slow roll runout for electric motors and generators is a condition in which the combined electrical and mechanical runout is measured on a rotating shaft at slow speed between 200 and 300 rpm, according to API. Since runout impacts the vibration readings and can lead to measurement errors, it is important to understand its various sources and how to mitigate it. This paper discussed the different contributing factors to the runout, how it is measured, how it can be used to calculate the compensated vibration and provided recommendations during the manufacturing process on how to check it before the machine is fully assembled. Being able to monitor the runout level during the manufacturing process can help avoid potentially disassembling the machine and returning the rotor to the lathe or grinder for rework. Not meeting the slow roll runout limit after the machine is

assembled can be costly to both manufacturers and customers.

X. REFERENCES [1] Nathan Littrell, Understanding and Mitigating Shaft Runout,

Orbit magazine, 2005, pp. 5-17. [2] API 541 Fifth Edition - Form-wound Squirrel-Cage

Induction Motors - 375 KW (500 Horsepower) and Larger, December 2014.

[3] Byron Knapp et Al, Professional Instruments Company [4] Mark J. Deblock, Barry Wood, J. W. McDonnell, Predicting

Shaft Proximity Probe track Runout on API Motors and generators, IEEE PCIC - 2007-32

XI. VITA

Papa M. Diouf (IEEE Member, 2013) graduated from

Purdue University in Indiana with a MSME in 2007. He has been with Baldor Electric since 2006. He has held several positions with Baldor including Manufacturing Engineer, Development Engineer and currently is a Senior Mechanical Design Engineer at the Large AC Motor plant in Kings Mountain, NC. He is a Registered Engineer in the State of South Carolina.

Bryan K. Oakes (Senior IEEE Member) graduated from the University of North Carolina-Charlotte in 1989 with a BSME degree. He has been with Baldor Electric in Kings Mountain, NC since 1989 and is currently the Manager of the Large AC Design & Mechanical Engineering Groups. He is member of the API 541 and API 547 subcommittees, and an author of multiple IEEE papers.