MCA Concept and Principle

Motor Circuit Analysis Concept and Principle

Howard W Penrose, Ph.D.

President, SUCCESS by DESIGN
Old Saybrook, CT

Abstract will then support the concepts by presenting a

series of experiments performed using MCA
Low voltage Motor Circuit Analysis (MCA) techniques and comparing to existing
techniques involve the collection and analysis of technologies.
resistance, impedance, inductance, phase angle,
current/frequency response and insulation to
ground faults. Output voltage of the test II. THE MOTOR CIRCUIT
instruments are less than 9Vac, sinusoidal output.
The resulting low level alternating magnetic fields Figure 1: Single Phase Equivalent Circuit
excite the dielectric dipoles and surrounding
magnetic steel dipoles, in both the stator and
rotor. Winding defects, including developing
shorts, cause changes to the dielectric and
resulting dipolar action (spin), which changes the
capacitance of the winding circuit at the point of
defect. The resulting change effects the phase
angle and current/frequency response in the
corresponding phase, causing a difference when The three phases of a three phase induction
comparing phase, or coil grouping, to phase. As electric motor are separated by 120o electrical.
the defect progresses, the changes to the insulation
The supply voltage phases are also, optimally,
system continue, allowing trending of the defect
over time. The purpose of this paper is to cover
separated by 120o electrical. Within each phase,
the concepts and principles in the physical as the voltage increases, current increases due to
changes to the windings in winding short and the impedance of the motor circuit (See Figure
winding contamination faults. The concepts will 1). As the current increases, the two magnetic
be supported with a series of experiments poles increase (or sets of poles), then decrease as
performed with MCA and traditional test current decreases. The stator back iron acts to
methods. strengthen and direct the magnetic fields within
the air gap between the stator and rotor. As the
I. INTRODUCTION fields pass through the rotor bars (conductors) of
the rotor, a second current develops in the rotor
Motor Circuit Analysis (MCA) techniques which interacts with the rotating fields in the air
utilizing Resistance (R), Impedance (Z), gap. The rotor follows the rotating fields,
Inductance (L), Phase Angle (Fi), current/ although lags behind the synchronous speed of
frequency response (I/F) and insulation the stator (slip) in order to maintain a rotor
resistance have been in practice since 1985. The current, and resulting rotor magnetic fields.
technique has been successfully applied for the
detection of winding defects (shorts, resistive As this is occurring, changes also occur to the
unbalances and insulation to ground), cable insulation system and back iron steel. As the
defects and rotor defects. It has also been found current increases in each phase:
to be able to trend and estimate winding failures
with a high degree of accuracy.i,ii 9 There is a skin effect within the copper
conductors that forces more current towards
In this paper, we will discuss the physical the surface of the conductor.
properties that allow for the detection of these
motor electrical circuit faults using MCA. We

©2008, SUCCESS by DESIGN Page 1 of 15 $25

 MCA Concept and Principle

9 Insulation dipoles polarize between The capacitance of the electrical insulation is a

conductors as the phase voltage and current direct function of the generation of dipoles
increases and decreases, causing constantly within the insulation system. As a field is
changing capacitance within the circuit generated across an atom, or molecule, of a
between conductors. dielectric, it will polarize, meaning that the
9 Insulation dipoles polarize between electron orbit of an atom will shift slightly,
conductors and ground as the phase voltage making one side of the atom more positive and
and current increases and decreases, causing one more negative.
constantly changing capacitance between the
winding circuit and ground. Figure 3: Generation of a Dipole
9 Magnetic dipoles polarize in the area of
effect of each pole within the stator core - +
steel. The reluctance to realign, with a
change in magnetic field, is termed as - -
hysterisis. -
- + - +
-
Operating voltages force the changes to occur - -
fairly rapidly. Changes to the circuit, or to the
dielectric or magnetic properties of the motor
effect its operation and the force of the operating Atom Dipole
voltage causes the defective areas of the
insulation or steel to heat. Continued As current passes through conductors near
breakdown of the dielectric occurs based upon electrical insulation, the insulation reacts by
the severity of the fault. polarizing the atoms (dipoles) within the
insulation. As the dipoles polarize, there is less
III. INSULATION AND MAGNETIC FIELD leakage (capacitance) between the conductors
EFFECTS and ground. This also occurs in the insulation
system between conductors when there is a
Figure 2: Insulation Model of Motor Winding System difference in potential. In a good insulation
system, the polarization of the insulation system
Phase A
occurs in a larger number of atoms. Once the
potential is removed, the atoms return to their
Ground
original state (dipoles randomize).

Figure 4: Dipolar Effect of Insulation

Phase B

Phase C

The electrical insulation circuit is modeled as a

series of parallel RC circuits between conductors GND
and conductors and ground. As changes occur
to the insulation system, the values of R and C
+
change. The values of the insulation in each -
phase are the sum of the turn to turn and coil to
coil RC values of each phase. Insulation to Conductors
ground values are the sum of the insulation Insulation
between conductors and ground for the complete
circuit.

©2008, SUCCESS by DESIGN Page 2 of 15 $25

 MCA Concept and Principle

Figure 5: Magnetic Dipoles the end of the coil, which causes a potential
between conductors to exist.
9 As the voltage begins to decrease, the
insulating and magnetic dipoles begin to
Magnetic randomize (move to neutral) at the
Flux beginning of the coil and release energy
back into the system as the fields collapse.
The fields at the end of the coil hit their peak
then start to decrease.
9 The voltage approaches zero, then passes
into the negative sequence of the sine wave.
The dipoles and fields continue to react, but
Magnetic Steel align in the opposite direction (as in a piston
action). We will define this action as
The same effect occurs in a magnetic field. The ‘dipolar spin’ of both the electrical
magnetic dipoles of the backiron and teeth of the insulation and magnetic steel dipoles.
stator core line up in the direction of the
magnetic field. This helps direct the magnetic The high potential of most electric motors force
flux and adds to the strength of the fields within the changes to the fields and dipoles to happen
the airgap. The reluctance of the steel to change quickly. As a result, work is performed and heat
polarity shows up as hysterisis losses from the is generated.
field. Once the field is removed, the magnetic
dipoles of the steel quickly randomize. The Capacitance of each portion of the circuit is
given, at any time, as:
The above descriptions for the polarization of
electrical insulation and core steel represent the
Q ε oS
steady-state application of an applied voltage [1] C=
potential. In an operating three phase system, Q−q l
the effects get far more exciting. As each
sinusoidal phase of voltage is impressed across Where an insulator exists between the
the windings: conductors and conductors and ground. The
induced charge, q, increases the capacitance by
9 As the voltage starts from zero, the the ratio Q/(Q-q). The dimensionless ratio q/(Q-
beginning of the coil energizes, the q) is a property of the polarizability of the
insulating dipoles between the insulation to material and is referred to as the electric
ground and the conductors within the coil susceptibility, Xe.iii At the boundary of each
are forced to polarize. insulation system (conductors, slot, phase, etc.),
9 As the voltage continues to rise, the the boundary conditions are such that:
potential at the beginning of the coil is
higher than the end of the coil, insulating [2] tan θ 2 = εr tan θ1
dipoles continue to polarize and the
magnetic dipoles begin to polarize in the
Where εr represents the relative permittivity of
direction of the magnetic flux generated by
the boundary of the insulation surface.
the coils.
9 As the voltage hits its peak at the beginning
By dividing each phase into tubes and slices,iv
of the coil, a majority of the magnetic and
the total capacitance for m slices and n tubes
insulating dipoles associated with the start of
through the system would be:
the coil have polarized and the ones at the
end of the coil continue to polarize. There is n m
δl −1
a lag in the fields between the beginning and [3] C = ∑ (∑ )
1 1 εδS

©2008, SUCCESS by DESIGN Page 3 of 15 $25

 MCA Concept and Principle

The inductance of the circuit can be figured as The circuit impedance would appear:
the flux linkage per unit of current, and is
represented by the unit Henry (H): VAB
[10] Z AB =
I AB
Nφ
[4] L=
i For example:
650.5∠120o V
For a motor with n coils, the inductance may be 32.9∠45o Ω =
defined: 19.8∠75o A

N p ( K pqφq ) Armed with this information, we can now

[5] L pq = review the effects of winding related faults on
iq the operation of the motor.

Where Kpq is referred to as the coupling IV. WINDING FAULTS

coefficient between two coils (p and q). When
p and q are equal, the inductance is termed as When a defect occurs in a winding due to a
self-inductance, when unequal, it is termed developing short, winding contamination or
mutual inductance.v severely damaged core steel, it effects the
electrical properties of the insulation system. In
The total impedance per phase as viewed from the case of a winding defect, changes to either
the stator input terminals is given as [6] where X capacitance or resistance within the insulation
refers to the leakage reactance (capacitive).vi system will cause a reactive problem due to
changes to the makeup of the insulation system.
R '2 For instance, in a developing short, the changes
jX M ( + jX 'l 2 ) to the insulation system cause changes to the
[6] Z t = R1 + jX l1 + s
capacitance due to changes in how the dipoles
R '2
+ j ( X M + X 'l 2 ) are excited (dipole spin). As a result, there are
s changes to how the insulation reacts in that area,
causing a leakage reactance variance and heating
In simpler form, impedance can also be viewed due to forcing the insulation to polarize with
as: high applied potential (operating voltage).
Winding contamination causes changes to the
[7] X L = 2πfL = Inductive Reactance resistive and capacitive reactance between
1 insulating surfaces, as well.
[8] Xc = = Capacitive Reactance
2πfC At design voltage, most defects do not become
[9] Z = R + (X L − XC )
2 2 apparent until a distinct change occurs, which
may be represented by a severe current
unbalance, nuisance tripping or a direct short
When looking at a balanced system, a wye circuit. In the case of winding contamination,
circuit should appear as in Figure 6. the end result is the same as a winding short:
Either a short between conductors or across the
Figure 6: Balanced Wye System
insulation system to ground.

As a result, as faults occur due to thermal

deterioration, contamination, moisture
absorption or other reactive faults, the circuit
impedance will change, slightly, at first, then
more dramatic as the fault progresses.

©2008, SUCCESS by DESIGN Page 4 of 15 $25

 MCA Concept and Principle

V. TRADITIONAL TEST METHODS D. Surge Comparison Testing

Most of the traditional test methods require a An older method of evaluating windings for
significant voltage application in order to work. shorts. A series of steep-fronted higher voltage
The purpose is to stress the insulation system by pulses are sent from the instrument to the stator.
forcing a reaction of the insulation dipoles, or The higher voltages occur too fast to properly
ionization of air and insulation medium defects polarize the insulation system, instead relying
in order to force a potential across a resistive or upon higher voltage to ionize gasses leaving the
capacitive fault. In this section, we will review a ability to detect a reactive fault as creating
few of the test methods in brief, including: enough potential to cross the barrier (Paschen’s
Insulation to ground testing; Polarization Index; Law) with the test ending prior to twice the
Resistance Testing; and, Surge Comparison nameplate voltage plus 1000 Volts or once an
Testing. arc is drawn. This method of testing causes a
change to the properties of the insulation at the
A. Insulation to Ground Testing (Meg-Ohm point of defect either accelerating the fault or
meters) completing the fault. In order to force slight
defects, a greater potential must be applied in
As described in Figure 4, a DC potential is order to stress the complete system. Due to the
placed across the motor winding conductors and steep fronted surges, however, the applied
ground. The applied potential is set and a value voltage is normally applied on the first 2-3 turns
of current (leakage) crosses the insulation in the first coil of each phase.
boundary. This value is converted to resistance,
usually in meg-ohms. It is, in effect, a method “The situation is quite different for detecting the
of measuring leakage across the insulation breakdown of the turn insulation in a winding
boundary, but only between the surfaces of the (parallel or phase) having many coils. The
conductors and ground. As the insulation breakdown of the turn insulation in a single coil
dipoles are only excited with DC, some time is in a winding of many coils produces a very
required for them to polarize. Standards small relative change in the characteristics (L, C,
normally indicate a winding charging time of R) of total load impedance seen by the surge
one minute and, as insulation resistance is generator. Hence, the change in the VWF
directly affected by temperature and moisture, [voltage wave form] shape produced by the
normalization for temperature. breakdown of the turn insulation somewhere in a
winding of many coils is relatively very small.
B. Polarization Index Hence the surge tests may not reliably verify the
presence of one shorted turn in a single phase
The Polarization Index (PI) test is a and may probably lead to wrong conclusions.
measurement of leakage at one minute then at Perfectly intact windings may appear to have a
ten minutes. The results are shown as a ratio of turn short. More importantly, a turn short
the ten minute to one minute readings. It is induced by the surge test by breaking down the
assumed that a fault will polarize slowly (high weakened turn insulation may not be detected.
ratio) or rapidly (low ratio) due to contamination In such a case, the stator winding would likely
and changes to the ground wall circuit fail after the machine is put back into service.
capacitance.
“In view of the above facts, caution is advised in
C. Resistance Testing surge testing of the turn insulation in complete
windings. These tests carry very significant
Resistance tests use a low voltage DC output risks, which should be carefully considered.
and bridge. The primary purpose is to detect Such caution is more important for diagnostic
high resistance joints, loose connections, broken tests on machines in service as such tests are
connections and direct shorts. carried out quite infrequently in contrast to

©2008, SUCCESS by DESIGN Page 5 of 15 $25

 MCA Concept and Principle

frequent tests on new, or refurbished, or repaired m1 k w1 N 1 2

machines in a manufacturer’s plant.”vii [11] a2 = ( )
m2 k w1 N 2
As shown, traditional testing has specific flaws
in the ability to detect faults, and the ability to Where 1 represents the stator winding factors
detect these faults in a non-destructive manner. and phases and 2 represents the rotor bar factors
and bars per phase. The result is a ratio, the
VI. MODERN LOW-VOLTAGE TESTING: same as a transformer ratio. When a rotor is
MOTOR CIRCUIT ANALYSIS stationary in an electric motor, the ratios are
different for each phase.
Modern MCA devices use a low voltage sine-
wave output designed to excite the insulation Winding contamination causes small changes to
system dipoles and surrounding magnetic steel the capacitance of the winding circuit. In most
dipoles with low current. There are several key cases, the capacitance increases within the
benefits to this approach: Size and voltage rating circuit. When referencing the simple impedance
of the machine being tested do not matter; formula, earlier in this paper, it identifies that an
Specific pass/fail criteria can be applied to phase increase in capacitance will have a negative
comparison; and, Degradation can be trended impact on impedance. Also, as the applied
over time without any adverse effects to the voltage is very low, capacitive reactance has a
existing condition. more significant impact on the impedance
(Ohm’s Law) as the capacitive value is more
“Based upon the physical and electrical dominant. The result, using a relatively low
properties of coil windings, insulation, systems, frequency and sinusoidal output, is a collapse of
transformer theory and electric motor theory, a impedance towards inductance in the phase
set of electronic measurements can provide the which has capacitive effects from the
necessary information to determine the condition contamination or water absorption. In cases of
of electrical equipment. The measurements high humidity, the insulation has to have fissures
must include circuit DC resistance, circuit or defects in order to cause the change.
inductance, circuit impedance, phase angle,
current/frequency response and insulation B. Overheated Windings
resistance readings. Resistance readings are
used for open or poor connections, inductance Overheated windings have a similar impact as
and impedance are used to evaluate winding winding contamination. The difference is that
condition in electric motors and phase balance in the insulation is thermally degrading causing
all other applications, phase angle and increased resistance to dipolar action. In this
current/frequency response tests evaluate case, the capacitance may decrease, causing an
windings for shorts and insulation resistance increase in impedance in one, or more, phases.
readings are used to detect winding to ground
shorts.”viii In both winding contamination and overheated
windings, the end result would be a winding
A. Detection of Winding Contamination short. Winding contamination can be corrected
if detected in its early stages. However, once
Due to the fact that one of the last measurements changes occur that allow for the detection of a
to change due to a turn short fault is inductance winding fault, the winding will have to be
(L), a test result of L can be used as a replaced.
comparative baseline. This is important as the
relative position of the rotor in an assembled C. Winding Shorts
machine will effect the reading due to mutual
inductance. One of the keys to proper MCA testing is that
inductance is not used as a primary method of
detection for developing shorts. Instead, two

©2008, SUCCESS by DESIGN Page 6 of 15 $25

 MCA Concept and Principle

specific measurements are used in combination Faults and defects can be detected in cables,
to determine the type and severity of the defect. coils, transformers, motors and rotor defects.
These measurements are: The circuit phase
angle; and, A Current/Frequency response E. Rotor Testing and Back Iron Effects on
method. MCA

When a defect occurs in the winding, it changes The effect of being able to evaluate the
the effective capacitance of the complete circuit. condition of the motor rotor is “based upon
The change in capacitance will directly effect Faraday’s law of electromagnetic induction,
how the low level current lags behind voltage according to which a time-varying flux linking a
with the usual result being an increase in coil induces an emf (voltage) in it.”ix
capacitance and a reduction of the phase angle in
the effected phase. Once the fault becomes [12] e1 = ωN 1φ m cos ωt for the primary emf
more severe, it will begin to effect the
surrounding phases. This normally occurs when
the defect exists in once coil or between coils in [13] e2 = ωN 2φ m cos ωt for the induced
the same phase. A very small change to secondary emf
capacitance within the circuit can be detected,
allowing the detection of single turn faults and e1 N 1
pinhole shorts when using very low frequencies. [14] = for the turns ratio
e2 N 2
A second method of fault detection uses a
current ratio, similar in method to the frequency Z1 N
[15] = ( 1 )2 = a2
response method used for transformer testing. Z2 N2
However, the low voltage current is measured,
then the frequency is exactly doubled and a Which is the ratio of the primary and secondary
percentage reduction in the low-level current is impedances of the circuit.
produced. When the frequency is doubled, small
changes to capacitance between turns or The motor circuit analyzer excites the core steel
between phases are amplified, causing a change based upon the amount of current available to
to the percentage reduction when compared the circuit and reacts across the airgap:
between phases.
Bliron Bl gap
The combination of phase angle and [16] nI = +
current/frequency response allow for the ur uo uo
detection of winding shorts and the type of short
being detected in any size machine. Also, due to The direct relationship to the ability to detect the
the use of low voltage and the result that only a rotor across the airgap depends upon the
small change to circuit capacitance is required to distance across the airgap, the area of the steel
detect the faults, early winding defects can be magnetized and the length of the stator core. In
detected quickly and trended to failure. longer cores, the effect will carry across the
airgap and excite the rotor core and induce the
D. Additional Tests instrument frequency into the rotor circuit. In
very short cores, the fringing effect the magnetic
In combination with the above tests, MCA field from the stator has a similar effect. In large
utilizes resistance readings and insulation to machines, the amount of energy available from
ground tests. This allows the technology to an MCA device allows for the detection of rotor
detect approximately half of the potential faults defects only above the area immediately
in the overall motor system and allows for the surrounding each coil side.
comparison of any two sets of insulated coils.
This produces multiple effects:

©2008, SUCCESS by DESIGN Page 7 of 15 $25

 MCA Concept and Principle

A. The mutual inductance changes as the rotor testing technologies allow for the detection of
position changes as a direct result of the incipient defects in the electric motor circuit
change to the transformer ratio between the including cable insulation, coils, transformers,
primary (stator) and secondary (rotor). A connections, motor and rotor windings,
good rotor will show as a repeating pattern, armatures, air gap issues and squirrel cage rotor
a bad rotor will change the transformer ratio defects, covering over 50% of all potential
and a defect will appear as a non-repeating motor system faults of any size or voltage
pattern. machine through the motor circuit and cabling or
B. Fractures will be readily detected as the directly at the machine. This is achieved by
induced energy is relatively low and the utilizing a low potential sinusoidal output from
oxides on the surface of the defect will the instrument which excites the insulation and
change the transformer ratio. Whereas, in magnetic dipoles of the circuit. The low
higher voltage rotor tests, the energy may be potential allows defects to become more readily
significant enough to pass through the apparent at early stages as it does not force, but
defect. excites, dipolar spin, causing changes to the
C. In rare instances, the airgap may be too circuit impedance, phase angle and current at
significant and very little to no variation of varying frequencies (current/frequency
the mutual inductance occurs. In this case, response), depending on the type of fault. These
larger defects, such as multiple fractures or a properties of the technology allow for long term
broken bar, will show as a variation in a trending of developing defects from insulation
straight line. breakdown and contamination without any
D. MCA technology has the ability to detect harmful testing effects due to insulation
wound rotor, synchronous rotor field and breakdown nor winding contamination.
other wound-rotor defects across the airgap.
Because of the impedance ratio between the VIII. DEVELOPMENT OF EXPERIMENTS
primary and secondary, rotor winding
defects will show as a change to phase angle In the first part of this paper, a number of fault
and current/frequency response and will detection capabilities and the method for how
vary based upon rotor position. faults are detected, using MCA methodology
were outlined. A series of controlled
F. Armature and Commutator Contamination experiments were developed, based upon this
Detection paper and previously developed tolerances,
using MCA and surge comparison testing. The
One of the unique abilities of MCA is the ability experiments were developed to provide the
to detect carbon buildup in DC motor armatures. following:
Due to the dielectric (capacitive) properties of
carbon, capacitance values of the circuit become 1. Experiment 1: A used 20 horsepower, 3600
unstable. This causes test results of impedance, RPM, 460 Volt stator only from a BJM
phase angle, current/frequency response and Pump, model KZN 150 (submersible) was
insulation to ground to become unstable and selected from the BJM Corp repair center
non-repeatable. As a result, armature circuit discard area. A visible winding fault,
contamination is detected by noting non- caused from what appeared to be mechanical
repeatable test results. This is important in that, damage, was identified. The original cause
if detected early, this type of defect may be of failure for the pump was seal failure,
corrected by blowing out the armature with low resulting in bearing failure with the bearing
pressure air. causing damage to the winding. The
winding was still operational when the pump
VII. SUMMARY OF MCA THEORY was sent for repair and was pulled due to
detection of the fault by MCA, then visual
Based upon the engineering principles of motor detection:
and transformer design, utilizing low voltage

©2008, SUCCESS by DESIGN Page 8 of 15 $25

 MCA Concept and Principle

1.1. Use MCA to detect fault using standard then re-tested using MCA. The
fault detection rules; purpose was to determine the ability of
1.2. Apply surge comparison voltage value MCA to detect the faults and impact of
until the visible fault was detected and surge comparison testing on winding
note voltage value. Determine if the contamination.
fault was detected in the same phase; 3. Experiment 3: Insert rotor in the stator from
1.3. Note effect of both test methods, during Experiment 1 in order to determine the
fault detection. impact of the rotor on the original test
2. Experiment 2: Performed in three parts: results and to identify severity of the faults.
2.1. A ¾ horsepower, 3-phase, 1750 RPM, Surge comparison testing was used to
460 Volt, used, complete Dayton motor degrade the fault condition and the results
was selected. The motor was monitored using MCA. The dielectric
disassembled and inspected, then strength was trended following each test.
reassembled. Testing was performed The differences between a running
before and after using both an ALL- equipment fault and progressive degradation
TEST IV PRO 2000 (AT4) and ALL- from the surge tester used are to be noted.
TEST PRO 31 (AT31). Surge
comparison testing was performed to The MCA devices were factory calibrated, as
1800 Volts for 3-5 seconds. The was the surge comparison tester, prior to test
purpose was to detect and determine evaluation. Testing was performed over three
field observed changes to impedance, separate days with variations to temperature and
inductance, and other test results humidity (dew point) in which the differences
following surge comparison testing in were noted between Experiment 1 and
the field. Experiment 3.
2.2. A Baldor, 1 horsepower, 1725 RPM,
460 Volt, new, assembled motor with a Through each experiment, the theories presented
rotor fault generated, involving the in the first part of this paper were supported and
removal of two rotor bars from the winding tolerances confirmed. Testing and
rotor circuit. The purpose was to detect results were recorded within instruments and via
the condition of the winding and a video log of critical test results. Variances in
determine the capability of detecting test results between the AT4 and AT31 involve
rotor bar faults, and the effect of rotor the use of different test frequencies ranging
position on test methods using the AT4, between 25 Hz and 800 Hz.
AT31 and surge comparison tester.
Confirmation of winding condition was IX. TEST TOLERANCES
determined by applying 12,000 Volts
using the surge comparison tester. A. MCA Tolerances
2.3. A 1 horsepower SPV style, 3600 RPM,
460 Volt, BJM Submersible Pump MCA Test tolerances for the technologies
stator with grey water contamination presented in this paper are well established:
was selected from the repair discard
pile. A coil-to-coil fault within the Table 1: Assembled Motor Tolerances
same phase, between individual turns, Test Tolerance
was observed during selection. The Resistance < 5%
stator was selected due to a low (0.98 Z and L Similar Pattern in Both Results
MegOhms) insulation test value, (phase comparison)
representing severe winding Fi and I/F +/- 1 Digit from Average Result
contamination conditions. The stator (phase comparison)
was to be evaluated using the AT4 and Insulation > 5MegOhm for <600V motor
AT31 for purposes of contamination Resistance >100 Mohm for >600V motor
detection, surge tested to 1800 Volts,

©2008, SUCCESS by DESIGN Page 9 of 15 $25

 MCA Concept and Principle

Table 2: Disassembled Motor Tolerances Figure 8: Surge Results T1-T3

Test Tolerance
Resistance < 5%
Z and L < 3% and <5% respectively
Fi and I/F +/- 0 From Average Result
(phase comparison)
Insulation > 5MegOhm for <600V motor
Resistance >100 Mohm for >600V motor

B. Surge Testing Tolerances

Figure 9: Surge Results T2-T3
Requires a visual tolerance between phases.
Separation of waveforms indicate a fault in the
windings, such as a short. Theoretically, surge
testing will provide a ‘shift’ in the right side of
the waveform prior to arcing.

X. EXPERIMENT TEST RESULTS

A. Experiment 1

The 20 horsepower stator was selected and The MCA test identified the phase to phase fault
placed on a wooden crate to ensure no in test leads T1-T3. To ensure detection across
interference from other surrounding materials. T1-T3, the stator was surge tested to failure
Testing was performed using MCA and (Figure 10).
following the rules for motors tested without
rotors, as shown in Table 3. Figure 10: Fault Detection (Night Vision Video)

Table 3: Pre-Test Results with MCA

1-2 1-3 2-3
Resistance 0.7209 0.7207 0.7206
Impedance 41 40 40
Inductance 8 8 8
Phase Angle 83 83 83
I/F -49 -50 -49
Ins Resistance >99

Surge testing was performed. All three phases The fault was detected on leads T1-T3 at 6,000
were, initially, limited to testing to 2,000 Volts Volts. In confirming tests, it was noted that the
(Figure 7, 8 and 9). voltage decreased to 4,500 Volts. No shift, or
other pre-arc conditions, were noted prior to
Figure 7: Surge Results T1-T2 distortion and concurrent arcing at the fault
point. Smoke was noted at the fault point, which
was found to be the arc burning through the
remaining insulation during surge testing (Figure
11). In the follow-up MCA, it was noted that
the impedance measurements across leads T2-T3
increased significantly (doubled) due to a
change in the applied frequency as a result of a
change to the condition of the circuit.

©2008, SUCCESS by DESIGN Page 10 of 15 $25

 MCA Concept and Principle

Table 4: Before and After AT4 Test Results with

Figure 11: Arc Damage (Night Vision Video) Differences
Before After Diff
Resistance 85.422 85.402 -0.02
Impedance 300 287 -13
Inductance 458 437 -21
Phase Angle 79 78 -1
I/F -41 -41 0
Ins Resistance >99

Table 5: Before and After AT31 Test Results and

Differences at 200 Hz
Before After Diff
Impedance 525 504 -21
Fi 72 72 0
The conclusion was that MCA testing was able
to detect the fault sooner, between phases, than I/F -44 -45 +1
the surge comparison test. Due to the location
The after tests were performed as quickly as
of the fault, this was expected as it was a ‘deep
possible following the surge test. Each result
winding’ fault in the center of a coil.
was performed twice to ensure repeatability (all
results were repeatable within 1%). The motor
B. Experiment 2a
was quickly disassembled, a metal piece placed
against the stator, which did not attract, then
One occasional observation, in the field, has
placed in front of a fan and reassembled and re-
been changes to MCA results following fast
tested. All tests returned to the original test
dV/dt high voltage testing, such as surge
findings.
comparison testing, including inductance,
impedance and other results. An assumption has
The original concept of dipolar action discounts
been the dipolar alignment of the magnetic steel
the following considerations: a) Paschen’s Law
and dielectric insulation following the motor
requires ionization of air and materials; b) As
being at rest for some length of time. If this
noted in the first part of this paper, the fast rise
were to be the case, then it would have to be
time does not allow for significant polarization
assumed that: a) Surge comparison testing
of materials; and, c) The assumption that there
would generate dipolar action and alignment;
would be remaining magnetic flux would require
and, b) The core steel would maintain the
steel with high magnetic retention properties,
dipolar action, assuming magnetic domains
which would result in extremely high hysterisis
would be generated, which would cause some
losses in the machine.
level of retained magnetism. Therefore, a used
motor with a good winding, that was estimated
The observations provide the preliminary
to show evidence of this effect, was selected and
conclusion that the cause of change has to do
surge tested to 1,800 Volts.
with ionization of the air within the stator, in
Figure 12: Disassembled Test Motor these instances. Ionized air and ozone (detected
as an ozone ‘odor’ when the motor was
disassembled) results in conductivity of the
atmosphere between the conductors of the
motor. Changes to the readings, during follow-
up testing with MCA, indicate that this condition
lasts a short period of time. This may also
identify the result of partial discharge in VFD
applications. The exchange of air supported the
conclusion as well as the lack of retained

©2008, SUCCESS by DESIGN Page 11 of 15 $25

 MCA Concept and Principle

magnetism following the test. Finally, none of As outlined in the first part of this paper, the
the tests noted in this paper, and other tests impedance and inductance of MCA followed the
performed for support of the findings of this same pattern, showing the insulation system was
paper, performed on stator-only conditions clean and dry. MCA identified a good winding
showed any of the same results, regardless of the in both instances. The AT31 identified bad rotor
amount of time not used. This particular area bars using a visual inductance rotor test. Each
has been noted for additional research for its of the surge waveforms followed the same
value in MCA. pattern, but required rotor position movement in
order to determine condition of the windings,
C. Experiment 2b but did not identify two missing rotor bars. The
surge test to 12,000 Volts identified the new
Figure 13: Rotor Position and Rotor Test winding was in excellent condition.

D. Experiment 2c

A stator was selected with winding

contamination from grey water in the windings.
The test results identified as 0.89 MegOhms to
ground. The motor was tested with the AT4 and
AT31, before and after the surge comparison
test. Surge comparison was limited to 1800
Volts, max. A visible coil to coil fault was
identified and detected using the MCA
instruments.

Table 8: AT4 Test Results on Stator

The motor was set up and isolated. The AT4 1-2 1-3 2-3
and AT31 were both used to analyze the Resistance 0.6400 0.5577 0.6189
condition of the windings, the AT31 was used to Impedance 27 26 27
analyze the condition of the rotor. The surge Inductance 5 6 5
comparison tester was used to evaluate the Phase Angle 79 80 79
condition of the windings and rotor. Following I/F -49 -49 -49
all tests, the surge comparison tester was
Ins Resistance 0.89
increased to 12,000 Volts to ensure condition of
each phase. Table 9: AT31 Test Results on Stator (60Hz)
1-2 1-3 2-3
Table 6: AT4 Test Results on Motor
Impedance 2.12 2.08 2.12
1-2 1-3 2-3
Fi 25 29 25
Resistance 20.368 20.336 20.334
I/F -20 -20 -20
Impedance 186 183 154
Inductance 147 144 122 Figure 14: Effect of Surge Comparison Testing on
Phase Angle 80 79 79 Winding Contamination
I/F -43 -44 -45
Ins Resistance > 99

Table 7: AT31 Test Results on Motor (60Hz)

1-2 1-3 2-3
Impedance 79.6 78.3 71.8
Fi 57 57 56
I/F -38 -38 -38

©2008, SUCCESS by DESIGN Page 12 of 15 $25

 MCA Concept and Principle

Table 10: AT4 Test Results on Stator Post Test Table 12: AT4 Test Results on Virtual Motor
1-2 1-3 2-3 1-2 1-3 2-3
Resistance 0.6400 0.5577 0.6189 Resistance 0.6821 0.6798 0.6865
Impedance 27 26 27 Impedance 21 21 39
Inductance 5 6 5 Inductance 8 8 7
Phase Angle 79 80 79 Phase Angle 77 76 76
I/F -49 -49 -49 I/F -47 -45 -47
Ins Resistance 0.0 Ins Resistance > 99

Table 11: AT31 Test Results on Stator Post Test Figure 16: Surge Values on Each Test
1-2 1-3 2-3
Impedance 2.12 2.08 2.12 4500
Fi 25 29 25 4000
3500
I/F -20 -20 -20 3000
2500
2000
The MCA tests identified the winding 1500
1000
contamination and coil to coil short. There were 500
no changes to the surge comparison waveform 0
1st 2nd 3rd 4th 5th 6th 7th 8th 9th
until the stator arced to ground at 1,600 Volts.
The stator was in a condition to clean dip and
bake prior to high voltage testing, but had a
Table 13: AT4 Test Results After Test 6 and Before
visible hole to ground following 3-5 seconds of
Test 7 on Virtual Motor
applied test voltage.
1-2 1-3 2-3
Resistance 0.6604 0.6592 0.6569
E. Experiment 3
Impedance 21 21 40
Figure 15: Virtual Motor with Spacers to Hold Inductance 8 8 7
Airgap Phase Angle 77 77 77
I/F -47 -45 -47
Ins Resistance > 99

Table 14: AT4 Test Results on Virtual Motor After

High Voltage Tests
1-2 1-3 2-3
Resistance 0.6786 0.6797 0.6819
Impedance 21 21 39
Inductance 8 8 7
Phase Angle 77 77 76
I/F -47 -43 -47
Ins Resistance > 99

Figure 17: Arc Damage to Winding and Insulation

Part of the purpose for the exposed ‘virtual’
motor is to reduce the influence of ionization on Carbonized
the test results. The stator was tested with
Fault from
MCA, then re-tested until the applied voltage
was less than 1,000 Volts before fault detection. Arcing

©2008, SUCCESS by DESIGN Page 13 of 15 $25

 MCA Concept and Principle

While this experiment proved that high voltage conducted to determine how best this effect can
testing will degrade defects in windings, MCA be used for fault detection, in particular in
will detect the fault and identify changes to the applications involving partial discharge.
condition of the winding, and leads to an
observation of the ability to detect and trend Experiment 2b Conclusions
degradation, it represents a fault different from
an operational fault. In an operational fault, the A number of important conclusions concerning
defect will overheat and a greater area of condition based monitoring resulted from this
insulation will be effected, causing a more experiment. It was determined that MCA is
significant change to the phase angle and/or capable of evaluating the condition of the stator
current/frequency test result as the insulation insulation system without having to move the
degrades. However, the test does provide motor shaft. This makes MCA an excellent
evidence that the prediction of the first part of method for Predictive and Condition Based
the paper, concerning changes to results, due to Maintenance of motors in-place. MCA was also
a progression of winding degradation, does found to be able to detect defective rotor bars
occur. It is observed that the final change to the using inductive-based rotor tests and movement
current/frequency reading came from a point of the shaft.
where additional conductors, other than the
original two, were effected and the area of High voltage surge comparison testing was
carbonization of the insulation increased. incapable, during this experiment, of detecting
rotor condition and required movement of the
It is noted, in Figure 18, that the separation of shaft in order to detect any winding defects.
the phase paper around the fault point is a direct This makes it less capable as a method for PdM
result of the surge comparison arc, evidence of and CBM programs. It was noted that new
Paschen’s Law in action. Low voltage test insulation systems should be capable of
methods had no impact on the existing condition withstanding high voltage surge comparison
of the insulation system. testing, including accidental over-potential.
However, it is noted that in all aged systems that
XI. CONCLUSIONS showed any defect, including minor, correctable
defects, the high voltage surge comparison test
The purpose of this paper, and the experiments degraded the insulation system significantly
contained within, has been to provide additional further than before testing. This presents the
supportive evidence for papers, research and conclusion that any defect that could provide an
field conclusions using MCA. estimated time to failure using MCA would
represent a near-immediate fault in high voltage
Experiment 1 Conclusions surge comparison testing. This further supports
the conclusion that it is less capable as a method
It has been determined that MCA is capable of for PdM and CBM programs.
detecting a phase to phase deep-winding defect
before traditional test methods as it does not Experiment 2c Conclusions
require a steep dV/dt. An arc is often formed
when high voltage surge comparison testing MCA was determined to have the capability to
detects a defect. The high energy involved detect winding contamination without a negative
degrades the remaining insulation life of the impact on insulation condition. This provides
insulation defect. for an excellent ability to evaluate system
condition during predictive maintenance or
Experiment 2a Conclusions condition based monitoring. High voltage surge
comparison testing was determined to destroy
Variations in MCA test results following high correctable insulation system conditions in
voltage testing appear to be due to ionization contaminated winding conditions. This supports
and not polarization. Additional work is being the EASA (Electrical Apparatus Service

©2008, SUCCESS by DESIGN Page 14 of 15 $25

 MCA Concept and Principle

Association) ANSI/EASA AR-100 Motor XII. REFERENCES

Repair Standard recommendation to clean
i
insulation systems prior to applying high voltage Penrose, Howard W Ph.D., “Estimating Motor Life
tests. Also, it provides evidence of risk in using Using Motor Circuit Analysis Predictive
high voltage test methods in predictive Measurements Part 1,” IEEE Electrical Insulation
maintenance applications. Conference Proceedings, 2003.
ii
Penrose, Howard W Ph.D., “Estimating Motor Life
Using Motor Circuit Analysis Predictive
Experiment 3 Conclusions Measurements Part 2,” IEEE ISEI Proceedings, 2004.
iii
P. Hammond and J.K.Sykulski, Engineering
It was determined that MCA has the ability to Electromagnetism: Physical Processes and
detect changes to the insulation system over Computation, Oxford Science Publications, 1994
iv
time, providing the ability to estimate time to The purpose of the tubes and slices approach, as
failure following fault detection. High voltage introduced by Hammond and Sykulski, is to provide
surge comparison testing was found to a manageable means to look at variances through a
significantly degrade the insulation system after system. This is done by taking the system in small
each successive test following fault detection, chunks referred to as tubes and slices.
v
Nasar, Electric Machines and Electromechanics,
rendering it less capable of estimating remaining
Schaums Outlines, 1981.
insulation life following fault detection. vi
Mulukutla Sarma, Electric Machines Second
Edition, PWS Publishing Company, 1996.
Summary vii
Bal Gupta, “Risk in Surge Testing of Turn
Insulation in Windings of Rotating Machines,” IEEE
Each of the experiments supported the theory Electrical Insulation Conference Proceedings, 2003.
viii
and conclusions of the first part of this paper. Dave Humphrey, Allison Transmission, “Which
However, it has been determined that additional Road Will You Take?,” IEEE Electrical Insulation
experiments are to be planned, with an Conference Proceedings, 2003.
ix
expansion to include Electrical Signature Syed A Nasar, Electric Machines and
Electromagnetics, Schaum’s Outline Series, 1981.
Analysis. The purpose of the additional work
will be to prove further the original intent of
Experiment 3 by generating an insulation defect, XIII: BIOGRAPHY
with the surge comparison tester, in a used
submersible pump motor. The pump will then Howard W Penrose, PhD, CMRP is the President
be run in normal operating conditions, with of SUCCESS by DESIGN of Old Saybrook, CT.
periodic MCA tests and periodic ESA tests in Prior to that he was a Senior Research Engineer
order to provide evidence of trending using (Energy Research) and Adjunct Professor of
MCA and to determine the point at which ESA Mechanical and Industrial Engineering with the
can detect a winding short. University of Illinois at Chicago following 12 years
in the electric motor industry as a Navy trained
Some of the conclusions of this study were Electric Motor Repair Journeyman and Field Service
Engineer. He is currently involved in motor
surprising to the investigators including the
diagnostic technology research and development and
repeated condition of the fault not showing until training.
an arc was drawn, although the arcs were Dr Penrose is the Editor in Chief of the IEEE
sometimes only visible with the lights out. In Dielectrics and Electrical Insulation Society Web and
each case, there were no ‘waveform shifts,’ nor eZine, a Past Chair of Chicago Section IEEE (1998-
other warnings, prior to waveform distortion and 1999), Vice Chair of Connecticut Section IEEE
arcing on the stators and assembled motors (2002-2004), Past Chair Chicago IEEE DEIS and
selected with real-life, versus laboratory PELS (1996-1999) and has published numerous
induced, faults. articles and books related to electric motor systems.

©2008, SUCCESS by DESIGN Page 15 of 15 $25