IEEE Transactions on Energy Conversion, Vol. 12, No.

4, December 1997

363

A COMPARISON OF TECHNIQUES FOR MEASUREMENT OF

SHAFT CURRENTS IN ROTATING MACHINES
Raymond Ong
Member IEEE
GE Industrial Systems
Peterborough, Ontario

J. H. Dymond
Senior Member IEEE
GE Industrial Systems
Peterborough, Ontario

R D. Findlay
Fellow IEEE
McMaster University
Hamilton, Ontario

ABSTRACT
Irregularities in the magnetic circuits of motors may result in
spurious voltages that lead to shaft currents through the shaft,
bearings, bearing supports and closing through the machine
framework. The IEEE Standard Test Procedure for Polyphase
Induction Motors and Generators discusses the shaft current
and presents a measurement method for recording either the
voltage across the ends of the shaft or the current. This paper
discusses an alternative measurement approach and its
application to the identification of shaft current in a large
induction machine.
Procedures were developed for measuring the shaft current.
The procedures include the shunt current method and the
measurement using the Rogowski coil. Only the Rogowski coil
measurement yields accurate measurements of shaft currents
whereas the other method either yields inaccurate measurement
or may result in other problems for the machines integrity. The
theory and justification for the superiority of the Rogowski
coil method is presented along with supporting test data.

JNTRODUCTION
Magnetic dissymetries may result in a voltage from end to end
of the shaft of a large induction machine. These dissymetries
may arise due to the jointing of segmented punchings around
the core of the machine, stacking of segments, rotor
eccentricity, rotor or stator sag, and other anomalies during
manufacture [ 1,2 1. However, there is evidence that shaft
voltage occurs on all rotating electrical machines to a greater or
lesser degree [ 3, 4 1. There is also the possibility that other
factors such as build-up of electrostatic charge may, in some
applications involving PWM inverters, give rise to bearing
current [ 5 1.

yields incorrect measurement for shaft current. The test

machine used for this work is a large, eight pole induction
motor. During initial conventional commercial testing, it was
found to have discolored oil and damaged bearings.
Examination of the bearing babbitt showed the characteristic
stitch marks with round shining bottom craters associated with
shaft currents. In this w e the oil was clean at the start of the
test, and precautions had been taken to insulate the bearings
against the possible development of shaft current. Bearing
insulation both before and after the tests was above 10,000 n;
hence there was no apparent cause for the bearing damage.
However, it is conceivable that a small voltage could be
established end-to-end across the shaft. On start-up, before the
establishment of an oil flm, this small voltage could lead to a
relatively high shaft current. On establishment of the oil film,
the interruption of the shaft current can then lead to a very
high component of induced voltage, resulting in breakdown of
the oil film with the consequent damage to the oil and bearing
[7, 81. The problem was then to develop a measurement
program to establish that shaft voltages did indeed occur and to
determine the magnitude: of such shaft voltage along with the
resulting bearing or shaft current.
The standard methodology as given in IEEE Standard 112
involved the use of a hi& resistance voltmeter to measure the
end-toend shaft voltage and a low resistance ammeter to
measure the shaft current. This paper will described the
reliability of the above method together with an alternative
method not given in the standard, using a Rogowski coil. It will
be concluded that the Rogowski coil method is the only reliable
method to measure shaft current, as will be shown below.

TEST METHODOLOCXES.
IEEE Standard 112-1991 [6] describes end-to-end shaft voltage
measurement techniques and states that shaft current can be
measured by substituting an ammeter in the circuit. As we will
show by means of theory and measurement, this procedure
PE-557-EC-0-01-1997 A paper recommended and approved by by
the IEEE Electric Machinery Committee of the IEEE Power
Engineering Society for publication in the IEEE Transactions on
Energy Conversion. Manuscript submitted August 2, 1996; made
available for printing January 8, 1997.

The first method used a low resistance moving coil type

ammeter placed in series with the shaft as illustrated in Figure
1.
, - ,
Drive End Shaft

I
\
W

AMMETER

SIZE AWG 2 CABLE

FIGURE 1: SCHEMATIC OF AMMETER METHOD

0885-8969/97/$10.00 0 1997 IEEE

364

Shaft current measurement varies inversely with size of cable.

The bigger the size of the cable and hence reduced cable
resistance, the larger is the shaft current. M e r a trial period,
cable of size AWG 2 was determined to be practical and
appropriate for stable shaft current measurement. The
connection of the cable to the two ends of the shaft can be done
through a brush assembly. However, due to space limitation as
a result of the installation of the Rogowski coils on both ends of
the motor and a collector ring on the opposite drive end of the
shaft, inch wide copper braided wires wound on the ends of
rectangular pieces of wood and connected to the AWG 2 cable,
were held in fm contact with the two ends of the shaft. A low
resistance ammeter was inserted in the circuit to measure the
shaft current. This method is considered unreliable as the
introduction of the connection from one end of the shaft to the
ammeter and then to the other end of the shaft results in a low
impedance path for the shaft current to flow. Without this
extemal connection there is no path for the shaft current to flow
as both bearings are insulated.
In the Rogowski coil method, two Rogowski coils were
installed around the shaft close to the rotor bore on both ends of
the shaft, Figure 2. The voltage outputs from the coils were
connected to the collector ring from which connection was
made to an isolator and then to a recording system. The isolator
is used to ensure that there is no grounding anywhere in the
connection from the Rogowski coil to the recording system. A
two inch diameter shaft hole was drilled from one end of the
shaft to the other end. The coaxial cable from the Rogowski
coil runs in this hole to the collector ring. In this manner the
influence of interference is further reduced. The collector ring
was installed on an insulated adaptor on the opposite drive end
of the shaft, Figure 3. All connection cables are coaxial type to
reduce noise interference.
ROGOWSKI COIL

FIGURE3: COLLECTOR RING INSTALLED ON THE

OPPOSlTE DRIVE END OF THE SHAFT OF
THE INDUCTION MACHINE UNDER TEST

THEORY OF ROGOWSKI GOEL

The Rogowski coil consists of a solenoid of fine wire
uniformly wound on a flexible nonmagnetic tube of which the
two ends are brought together to form a butt joint. The coil,
basically an air core transformer with a uniformly wound
secondary, is wound tightly around the conductor carrying the
current, in this case the shaft. There is no disturbance of the
conducting circuit unlike the case when the shunt/ammeter
circuit is used. The coil output voltage is proportional to the
time rate of change of current. An outline of the Rogowski coil
is shown in Figure 4.

Contour

FIGURE 4: OUTLINE OF ROGOWSKI COIL

The current- coil output voltage relationship, details of which
are described in [9] is given as:
FIGURE 2 LOCATION OF ROGOWSKI COIL IN THE
INDUCTION MACHINE UNDER TEST

365
where A is the area of the coil cross-section,
p, the permeability of free space
and n the number of turns per unit length of the coil
For sinusoidal current having a peak value I and a peak output
E of frequency o, equation ( 1 ) becomes

The bearing insulation resistance is at least of the order of

kilo-ohms, greater than slhaft resistance and oil film resistance
and hence equation (3) can be simplified as
vi

= I, Rb

(4)

Equation (4) indicates the shaft current to be of small

magnitude.

Area A is relatively small compared with the area of the loop

formed by the closed coil. It is important to calibrate the
Rogowski coil in respect of the conducting circuit being
measured.
JUSTIFICATION OF THE USE OF ROGOWSKI COIL
IEEE 112 indicates the use of a low resistance ammeter in the
measurement of shaft current. The ammeter method disturbs
the shaft current conducting circuit of the machine as indicated
in Figure 5 by providing a low resistance path through the
shunt resistance or the ammeter resistance

Using the ammeter method , shown in Figure 5 as the dashed

circuit to measure shaft current, voltage-current relationship
can be obtained as
vi

= 1, (Rsh+Rs + & )

(5)

where & is the shunt resistance or the ammeter resistance and

& the connection contact resistance.
The ammeter resistance is of the order of milli-ohms. With the
negligible magnitude of the shaft resistance and ammeter
resistance, the shaft current has been amplified. The amplified
shaft current does not represent the true value of the machine.
The amplification of this current through the ammeter method
can cause secondary effects on the shaft and the bearing.

The Rogowski Coil mlethod does not disturb the current

conducting circuit of the machine. The current measured is the
true current flowing in the shaft. The Rogowski coil is mounted
Opposite Drive End
tightly around the shaft on the inboard of the bearing bracket
_ _ - _ _ - _ . - _ _ - assembly.

R, (shuntlammeter resistace)

(order of m a )
Drive End

k h

(order of pL2 )

Oil Film Resistance Bearing Insulation

(Order of kilo-ohms)

- . - Ammeter Method
(

Rogowski Coil Measurement

i'
Figure 5 : Simplified Representation of Machine Under Test
This can be shown as follows:

Applying Kirchoff s law to the heavy lined circuit in Figure 5,

we obtain

where

Vi is the induced shaft voltage,

R a the resistance of the shaft,

and

Ro the resistance of the oil film

Rb the bearing insulation resistance

CALIBRATION OF ROGOWSKI COIL

Calibration of the Rogowski coil was done following
installation on the shaft. The calibration procedure involved the
use of a current generator ,replacing the ammeter in the circuit
as shown in Figure 1. The Rogowski was calibrated at a
number of frequencies. The drive end Rogowski coil had a
calibration value of 11.02 millivolts per ampere at all
frequencies from 8 Hi! to 4 KHz. The opposite drive end coil
had a calibration value of 13.72 millivolts per ampere at the
same range of frequencies.
TEST RESULTS
Both ammeter method and Rogowski coil method were
conducted at five different voltages when the machine is
running at steady state rated speed condition. Condition A as
indicated in Tables 1, artd 2 is the Rogowski coil Method with
the drive end (DE) shorting pin down and virtually shorting
the insulation while the opposite drive end (ODE) pin is up
and thus maintaining the bearing insulation. Condition B is the
ammeter method for the measurement of shaft current with
similar bearing insulation condition as Condition A. Condition
C is the ammeter method for the measurement of shaft current
when the shorting pins at both ends are up and hence
maintaining the bearing insulation. Under conditions B and C,
measurements from the Rogowski coil were taken at the same
time as the ammeter reaiding was taken and show in Tables 1
and 2. These measurements verify the accuracy of the usage of
the Rogowski coil in condition A. Figures 6 and 7 indicate
that the ammeter method introduces higher harmonics which
would not have been present in the machine.

366

SHAFT

120%

MOTOR LINE VOLTAGE

I 40% I 60% I 80%

I 100% I

2m

VP

TABLE 1 : ODE ROGOWSKI COIL MEASUREMENT

Hz

MOTOR LINE VOLTAGE

1K

X: 40.00 Hz Y:377.5pV
FIGURE 6b: SPECTRUMPLOT OF DE ROGOWSKI COIL
MEASUREMENT UNDERCONDITION A
I

TABLE 2 : DE ROGOWSKI COIL MEASUREMENT

SHAFT
CURRENT

MOTOR LINE VOLTAGE

120% 140% I 60% I 80% I 100%
Rated Rated Rated
Rated Rated

TABLE 3 : AMMETER MEASUREMENT

I END-TO-

FIGURE 7a: DE ROGOWSKI COIL MEASUREMENT

UNDER CONDITION B
5m

MOTOR LINE VOLTAGE

VP
TABLE 4 : SHAFT VOLTAGE MEASUREMENT

Hz

20m

1K

X: 180 Hz Y: 4.363 mV
FIGURE 7b: SPECTRUM PLOT OF DE ROGOWSKI COIL,
MEASUREMENT UNDER CONDITION B
CONCLUSION

-20m
0

Seconds

FIGURE 6a: DE ROGOWSKI COIL MEASUREMENT

UNDER CONDITIONA

800m

The ammeter method had been s h o m to amplify the shaft

current value. This value includes higher harmonic values in
addition to the fundamental values The introduction of
harmonic content m the shaft current in the ammeter method
can have adverse effects on the oil film, resulting in its
breakdown and arcing the bearing. The Rogowski method
provides accurate measurement of shaft current in the machine
under running condition.

The work described in this paper arose out of a research and

development program supported in part by GE Canada and
Natural Sciences and Engineering Research Council (NSERC)

367

of Canada. The authors wish to thank the test department of GE

Canada for providing the personnel and utilities to perform the
work and the other members of the research team, Dr. Haran
Karmaker, Dr. J. Dableh and Dr. Szabados for their valuable
insight to the project.
REFERENCES
[11 Paul L. Cochran , Polyphase Induction Motors, Analysis,
Design and Application, Marcel Dekker, New York
and Basel, 1989.

[2] C. Ammann, K. Reichert, R. Joho, 2. Posedel, Shaft

Voltages in Generators with Static Excitation Systems
Problems and Solution, IEEE Trans. on Energy
Conversion, Vol. 3 , No. 2 June 1988, pp. 409 419.

(31 Micheal J. Costello, Shaft Voltages and Rotating

Machinery , IEEE Trans. IAS, Vol. 29, No. 2 ,
March/April 1993, pp 419 426.

J. S. Sohre and P. I. Nippes , Electromagnetic Shaft

Currents and Demagnetisation of Rotors of Turbine
and Compressors , 7 th Turbomachinery Symposium,
December 1978 , Texas A & M University, College
Station ,Texas.
Doyle Busse, Jay Erdman , Russel K. Kerkman, Dave
Schlegel and G q Sibinski, System Electrical
Parameters and their Effects on Bearing Currents , IEEE
APEC Conference, San Jose, March 1996.

IEEE Standard 112 1991 , IEEE Standard Test

Procedure for Polyphase Induction Motors and
Generators, IEEE Standard, 1991.
S. P. Verma, R. S . Girgis and R. J. Fleming, The

Problems and Failures Caused by Shaft Potentials and

Bearing Currents in Turbogenerators International
Conference on Large High Voltage Electric System,
CIGRE Session, 1980, paper 11 10.

J. Boyd and H. N. Kaufman, The Conduction of Current

in Bearings ,ASLE Trans. October 1958, pages 67-77.
R. L. Stoll , Method of Measuring Alternating Currents
Without Disturbing the Conducting Circuit , Proc. IEE,
Vol. 122, No. 10, Oct 1975.

BIOGRAPHJES
Raymond Ong ( M 89) received
his B. Sc. degree in electrical
engineering from Southampton
University, U.K. in 1981 and
and M. Sc. ( Research ) degree
in electrical engineering from
Queens University, Kingston,
Ontario, Canada in 1985. From
1986

to

1089,

he

was with

Singapore Aerospace as Avionics

Engineer, where he was involved
in avionics integration and power
system analysis of aircraft. In

1989, he joined GE Caniada as electrical design engineer and

worked on induction and synchronous machine design teams,
shaft and bearing currenlt project and product structuring. His
interests are in the areas of electromagnetism, electrical
machines and electrical tlrives. Raymond is a member of IEEE
Power Engineering Society and a licensed professional
engineer in the province of Ontario
James H. Dymond ( S 67 - M 67
-SM89) received theB.Sc. degree
in mathematics from Memorial
University of Newfoundland, St.
Johns , Canada in 1965 , and
the B. Eng. degree in electrical
engineering ( power ), h m the
Technical University of Nova
Scotia, Canada, in 1967. In 1967,
he joined GE Canada on their
engineering training program in
induction motor design. Between
1968 and 1972, he worked on a
postgraduate project on lasers at
Memorial , returning to GE Canada in 1972 to work on
induction motor design and development. His project work
covers air-to-air coaller design, machine ventilation,
electromagnetic noise arid resonance, windage noise, transient
heating and cooling of induction machines, insulation systems,
harmonic heating, startbig calculations, induction machine test
methods and shaft and bearing currents. Jim holds a number of
patents and GE patent nwards and is contributor to a number of
IEEE ,AFT, Canadian and IEC standards working groups. Jim
is a senior member of IEEE Power Engineering and Indusby
Applications Societies and a registered professional engineer
in the Province of Ontario.
Raymond D. Findlay ( F 94 ) was
bom in Toronto, Canada, obtaining
degrees of B. A. Sc ., M.A. Sc. ,
and Ph.D. in 1963, 1965 and
1968 respectively, in electrical
engineering from the University
of Toronto. Dr. Findlay, a Fellow
of the Engineering Institute of
Canada , is a registered professional
engineer in the Province of Ontario,
and a member of ASEE. Dr. Findlay
is IEEE Vice President, Regional
Activities (1996), and has served
IEEE in many capacities over the past 25 years. His interests
include electromagnetic fields and losses in power apparatus,
an area in which he holds one US and one Canadian patent.
From 1967 to 1981 he was at the University of New
Brunswick. During 1972-3 he was a project leader at Canadian
General Electric Co. in Peterborough, Ontario. He has been a
Visiting Fellow at the University of Southampton (U.K.) in
1979-80, at the Katholieke Universiteit Leuven (Belgium) in
1988, and at the Commonwealth Scientific and Industrial
Research Organisation I[ Australia ), also in 1988. Dr. Findlay
is a Professor and a mamber of the Power Research Laboratory
of McMaster University, where he has held an appointment
since 1981. He was the Director of Engineering and
Management Programme at McMaster from 1991-4.