A Simple Method For Determination

of Ratio Error and Phase Angle

in Current TransFormers
Correlation of Significant Constants of Current Transformers

E. C. WENTZ
ASSOCIATE AlEE

Synopsis: Up to the present time current

transformer performance has been described
by curves which show ratio error and phase
angle over a range of current at various
values of burden, (ohms) power factor, and
frequency. The number of curves that are
required to cover all possible conditions is so
great as to render them, by their very bulk,
practically useless. Another limitation is
that they do not describe performance under
transient conditions.
Performance of other types of transformers and certain other apparatus is invariably calculated from a few easily measured constants, according to simple standard
formulas. Up to the present time" available methods for calculation of current
transformer performance have had the disadvantage of either being difficult to use, or
of depending on constants difficult to measure, or else they have not been sufficiently
accurate for general use.
A method is described which depends on
the following constants:
1. Open-circuit saturation curve measured in the
usual way.
2.

Secondary coil resistance.

3. Equivalent leakage reactance calculated from

the voltage of an exploring coil enclosing the maximum value of leakage flux.

The method includes several itnportant

short cuts in the calculation which reduce
the numerical work to a practical minimum. Curves are plotted which allow direct reading of the exact components of exciting current which cause ratio error and
phase angle.
,
The objects of the paper are:
1.

To present a practical method of calculation.

2. To establish constants (1), (2), and (3) as the

fundamental quantities which determine real excellence of the transformer.

I.

The N~ed for a Simple Method

for Description of CurrentTransformer Performance'

H E conventional rati~ and phaseangle curves for a current transformer are often inadequate for solution
of a particular problem. Furthermore,
the tests from which such curves are
plotted are expensive, and the data are
not available for many of the older transformers now in service.
A simple method for determining current transformer performance from a few
easily measured constants (as power
OCTOBER

1941,

VOL.

60

transformer performance is determined

from copper and iron loss and impedance
measurements) has been earnestly sought.
This paper will present a method which
the author believes to be the best so far
proposed.

II.

Theory of Current
Transformation

The general principle of calculation of

current transformer performance is very
simple. The ampere turns in the secondary coil are less than those in the primary coil only because of the ampere
turns lost in circulating flux in the core.
Hence, we need only to calculate the ampere turns or magnetomotive force drop
around the magnetic circuit.
However, on account of the leakage flux
in the core, this is not as simple as it
sounds. This leakage flux is principally
the flux which passes between primary
and secondary windings, and completes
its path in the core outside the primary
winding (figure 1). The current transformer is different from other transformers in that this leakage flux is relatively
large, often larger than the working flux.
This flux has been investigated experimentally! and theoretically, 2,3 but no
method has been developed by which it
can be exactly calculated. Furthermore,
it is well known that it enters the core, not
at one point, but all over the whole surface, so that it is not the same at any two
points in the core
If anyone of a number of approximate
methods is used to calculate the leakage
flux, the ampere turns to circulate it
Paper 41-109, recommended b : the AlEE co nrnittee on protective devices and instru ments and measurements, and presented at the AlEE summer convention, Toronto, Ontario, Can., June 16-20, 1941.
Manuscript submitted February 20 1941; made
available for preprinting May 6, 1941.
E. C. WENTZ is with Westinghouse Electric and
Manufacturing Company, Sharon, Pa.
The author wishes to thank J. B. Gibbs of the Westinghouse transformer engineering department for
invaluable assistance in organizing and arranging
the material.
1. For all numbered references, see list at end of
paper.

Wentz-Current- Transformer Calculation

through the core can be calculated by

dividing the core into a number of parts,
assuming a constant equivalent flux in
each part. This is the basis of the method
described by Mr. Sinks," and this method,
although still not absolutely correct (its
limitations are fully described by Mr.
Sinks), is the most nearly correct method
now available. Its real limitation is the
great amount of numerical work involved.

III. Assumptions Which Simplify

Calculation
Certain methods have been developed
for simplifying this calculation. The
most simple method, which does very well
for ring type transformers, is to neglect
the leakage flux altogether. 4 Another
method assumes an experimentally determined flux, approximately related to
an average value of leakage flux, to flow
in the whole core." These methods are
not suitable for general use, because they
do not use the real magnitude and distribution of the leakage flux.

IV. Combination of Best Features

of All Methods
Giving first consideration to the fundamental soundness of the method, we can
say that the method should1. Consider division of the core into several
parts, as discussed in II with different leakage fluxes in the different parts.
2. Base the distribution of the leakage flux
and the division of the core, on values of
leakage flux actually existing in the core as
measured by exploring coils (figure 2). The
exploring coil can be simply ten turns of fine
wire.

Giving second consideration to the aspects of economics and convenience, we

can say that the method should not divide
the core into more parts than is necessary
for the required accuracy. Here a good
deal of experience is required. Any particular division of the core into any finite
number of parts is an empirical device,
and can be justified only by general experience which shows that it works. The
final result cannot be shown, in any rigorous way, to be the best possible result.
It will help to remember the primary
objects of the investigation, and to make
the division of the core accordingly. The
first object is to calculate an error that is
at least as large as the real error, to be on
the safe side. The second is to get consistent results, to be able to compare
transformers on a true basis. The first
object can be obtained if the average
flux values in each part of the core are
taken to be sufficiently large, even if the
TRANSACTIONS

949

core is not too well divided. The second

object will be obtained if the flux measurements are made in the same way on
all transformers of the same type.
Of course, we do not want to calculate
an error too much larger than the true
value. This is the disadvantage of not
dividing the core at all, but considering
the measured value of leakage flux to flow
in the whole core; the calculated error will
always be too large.
Division of the core into two parts,
with the real leakage flux measured by
exploring coils, is a compromise which the
test results to be cited, and other tests,
show to be suitable for general use with
dry type transformers. This compromise
is, however, definitely not fundamental;
the general method to be described can be
used with any desired division of core.
When the ratio error exceeds five per
cent, the iron is usually working near
saturation, and much of the leakage flux
which is in the iron at normal flux densities leaves the core at the higher density
and completes its path in the air leaving
less flux in the core than would be. expected. The result is that this method
of calculation will give errors which are
considerably too large when the error
exceeds five per cent. As the method is
conservative, it is still useful.
Actually less than the value of flux as
measured by the exploring coil shown in
figure 2 exists in practically all of the core.
It is obviously conservative to assume
that it exists in the half of the core which
. is outside the primary coil. Mr. Sinks
has shown that the leakage flux at the
center of the core inside the secondary is
effectively reversed to the leakage flux in
part of the core just outside the secondary. A reasonable assumption, confirmed
by tests, is that the average effective value
of leakage flux in the inside half of the
core is zero, and that the only flux to be
considered is the working flux.

v.

Simplification of Numerical
Work

The proposed method of determining

the leakage flux and its distribution does
not solve a real difficulty attending the
problem-the amount of numerical work
for a solution.
Several short cuts have been worked
out which do not detract from the accuracy of the calculation. These are:
1. Substitution of equivalent reactance for
the leakage flux. It can be shown (appendix I) that the transformer with magnetic
circuit divided into two parts, one with, one
without, leakage flux, can be represented by
the equivalent circuit of figure 4, where the

950

TRANSACTIONS

Figure 1. The leakage flux passing in

the space through
and between coils
completes its path
through the core

'109-'

The leakage Aux is

relatively large in a
current transformer

ASSUMED

APPROXIMATE

EQUIVALENT

REAL DISTRIBUTION

DiSTRIBUTION

OF FLUX

reactance XL corresponds to the leakage flux

and can be determined from the exploring
coil voltage at normal current by-(see
appendix II)

Is = secondary current

Phase angle =

R s = secondary coil resistance

N s, N x = num ber of turns in the secondary

and exploring coils, respectively
F x = induced voltage in the exploring
coil
(2). When the transformer is considered as
the equivalent circuit in figure 4, the problem of calculating primary (input) current
from the secondary (output) current is
simply vector algebra. The fact that we
have limited the method to five per cent
ratio error (approximately five per cent total
exciting current maximum) justifies the
assumption that the voltage drop in XL is
not (Is - Ie1)XL (see figure 4) but IsX L, and
the voltage on the number 2 exciting branch
is Is(Zb+Rs+X L) (adding vectorially).
The calculation of performance is then
exactly as if the two exciting branches were
two separate current transformers, one with
burden Z1J+Rs, the other with burden
Zb+Rs+XL, and the ratio and phase-angle
errors will be the sum of the errors of the
two.

3. I t is well known? that the component of

exciting current which produces ratio error
is:
F cos O+kf sin 0
F, M = watt reactive components of excit-

ing current, respectively

cos 8 = burden power factor

per cent turns rat io+

100 (F cos 8+ kf sin 8)

Is = secondary curren t

343R
~

(M cos 8 - F sm 0)

The phase-angle component is M cos (JF sin 8, and a similar set of curves can be
plotted for this value.
Figure 5 is an example of such a set
of curves for one transformer and figure
7 is a set of F cos 8+ M sin 8 curves for
a different transformer. The log-log
co-ordinates of figure 5 are more useful if
the lower range of the curves is important. This idea was originated by A. M.

Wiggins."
The data needed for calculating perforrnance then are:

(1). Saturation curves of the form FX

cos 8 M sin 8 and M cos 0 - F sin 0 plotted
against secondary voltage, for anumber of
values of cos 8 (see figure 5).

(2).

Secondary winding resistance.

(3). Leakage reactance XL measured according to appendix II.

The procedure is:

(1). Given burden and power factor, determine the total burden on the right hand exciting branch of figure 4. (Burden +secondary resistance = Zl.) Also determine the
sum of this burden X L, which is the burden
on the left hand exciting branch. Z, X L =
Z2. (Vectorial addition.)

This value, added to the secondary current,

gives the primary current reversed. The
formula is:
per cent ratio:

tween exciting current, whichisV F2+M2,

and secondary voltage. It is not much
more difficult to plot F cos 8+ MX
sin 8 for a number of values of cos (J, instead ofV F2+M2, and the desired value
can then be read directly.

(2). From these two values of burden in

ohms and the secondary current, calculate
the voltages on the two exciting impedances.

E 1 =lsZt

E 2 = I sZ 2

Is

8+

Calculating the value of F cos

ill sin 8 from the ordinary open-circuit secondary saturation and power factor,
curves is not difficult, but calculation of
a number of points is tedious. Now the
ordinary saturation curve is plotted be-

Wentz-Current- Transformer Calculation

(3). From the F cos 8+Jl sin () curves,

read values for E 1 and E 2 and find per cent
ratio by100
per cent ratio = per cent turns ratio+- X
,
2
[

( F cos 8+M sin 0)1+ (Fcos O+M sin

Is
18

8)2J

ELECTRICAl.. ENGINEERING

I
I

Figure 2. The most direct and

simple as well as the most
nearly correct method for determination of leakage Rux is
measurement by exploring
coils

PRIMARY

SE CONDARY
N S TURNS

I
I

I
I

I I

I
.1
I

The lengths L1 and L2 represent relative (not ebsol ute)

lengths of the part of the magnetic circuit which can be considered to carry no leakage flux
and all the Iakage flux, respectively, in a through-type
transformer

,~

((;1
~~
""EXPLORING COIL

NX TURNS

I
I

~
r
}~ 7'"-"'[

EXPLORtNG COIL
~ to-MUST- BE ON .BOTH
~===~

LE GS OF THE CORE, AS
ER RATIC MEASUREMENTS

r
i

1
;

I'

RE SULT FROM A COIL ON

ONE LEG ONLY, DUE TO
SLI GHT DISPLACE MENTS
OF THE PRIMARY
CONDUCTOR.

The subscripts 1, 2, refer to the two parts of

the core; the division by 2 takes account of
the fact that each component is for only half
of the core, whereas the saturation curves .
are for the whole core. It will be noted that
the exact turns ratio must be known. This
ratio is often difficult to measure, but the
manufacturer can always supply it. I t may
also be difficult to place the exploring coil on
a transformer. This also is easy for the
manufacturer, however. General use of
this method will probably depend on the
manufacturer's measurements of exact turn
ratio and leakage flux.
Phase-angle minutes =

3,438
-~

(M sin 8- F cos Oh +~M cos 8- F sin 8)2J

[
Is
Is

VI. ---COmparIson of Tests and

Calculations
The constants which determine the
performance of the transformer at a given
frequency are:
(1).

The secondary coil resistance.

(2). The secondary leakage reactance X

measured according to appendix II.

(3). The F cos 8+M sin 8 and M cos oF sin 8 curves for exciting current plotted
against secondary voltage.
(4).

The burden.

(5).

The secondary current:

The fact that this method gives correct

results; for ring type transformers, in
which there is no leakage flux, has been
amply demonstrated.
:The method has been demonstrated on
several wound type transformers, and the
OCTOBER

1941,

VOL.

60

constants for one wound type and one

through type transformer will be given,
together with calculated and measured
results.
(a)

The curves of figure 6 were calculated

in the same way, and one who desires to
may check them very rapidly by the use
of tabular calculations.
(b) THROUGH-TYPE (TWO-COIL)
TRANSFORMER

Good results are generally obtained by

dividing the core into two equal parts,
in the case of a wound type current
transformer, but the division should not
be made quite this arbitrarily for a
through type current transformer. If
the length of magnetic circuit not enclosed by the coils is taken as the part
carrying leakage flux, and the axial
length of the coil is taken as the part not
carrying leakage flux, better results are
generally obtained. These two lengths
are, more often than not, nearly equal,
so that the final division is usually not far
from division into two equal parts, but
some designs are made with unusually
long yokes which carry the leakage flux,
and account should be taken of their
length. The lengths of these parts of
the circuit are usually easily measured,
according to figure 2.
The particular design in question was
such that the core could be considered
to be divided into equal parts.

WOUND-TYPE TRANSFORMER

Secondary coil resistance :

0.3 ohm. (R s)

Secondary leakage reactance:

Exploring coil voltage Ex 0.148 in 10 turns
(Nx = 10)
At Is = secondary current = 5 amperes
N s = secondary coil turns = 239
From equation 7, appendix I

X=

~(

O . 148 X 2392) -031 2=0638 ohm

5XI0

The F cos 0+ M sin 0 and M cos 0- FX

sin 0 curves are given in figure 5.
The calculation of a given point is
as follows:
Secondary current = 5 amperes
Burden=0.6 ohm, 90 per
cent power factor =
o.54+j 0.262
Add R s (secondary coil =
0.31
(Zl = 0.89 ohm, 96 per cent \
power factor) Z 1 =
0.85+j 0.262
Burden voltage E 1 = 4.45
Add X L =
j 0.638
Z2 = 1.24 ohms 69 per cent
0.85 j 0.9
power factor
Burden voltage E 2 = 6 . 2
From figure 5, at E 1 =4.45, (F cos 0+
M sin 0)1 =0.0133
at El. = 6.2, (F cos 0+
M sinOh=0.0229
per cent turns ratio = 99.584+
~~~oc:nt 100 (0.0133+0.0229) = 0.362
2
5
99.946
Test value =
99.89

vv entz-Current- Transformer

Calculation

/109-] .

CE NTER ELEMENT

NO.1! NO.2,/

I
I

OF CORE

NO.3

-_.

NO.4

'

NO.6; NO.5',
SECONDARY COIL
RE~STANCERS

BURDEN

MAGNETIZING BRANCH
REPRESENTING CENTER
ELEMENT OF CORE

Figure 3. The Rux at the center of the core

inside the secondary coil is actually less than
enough to generate the total burden voltage,
because the "leakage" Aux passing through
the body of the secondary coil actually links
some of the secondary turns and induces a
part of the secondary voltage
This real reduction of Aux in the center element
of core can be represented exactly by a negative series reactance or a capacitor C
Each successive element of the core, startingfrom the center element, carries a greater part
of the real leakage flux, represented by added
series reactances, and the network is an dCCUrate equivalent

TRANSACTIONS

951

l ~.
[Qj

The curves for this transformer are

given in figure 7.
The secondary resistance was 0.345
ohm. (R s)
The secondary turns (Ns) were 280;
ratio 1,400/5 amperes.
Two exploring coils were used, one on
each yoke of the core, ten turns each, in
series as in figure 2 (Nx=20). Erratic

Is'"

0.1

aURDEN

.-..-------

NO.2

za

SIMPLIFIED WORKING DJVIStON

FOR PRACTICAL CALCULAT10NS

Figure 4. A practical working equivalent for

figure 3 is the two-branched network
(0.995

1000
800
700
600
500

1/0'1- SA

-L

Ir~~~
I I I
~V~
O'
.\\

oc,

400

~06.D'~ \

50 VOLTS

...J

t>

~ ~g
~

20
15

-c

5
ui
(/')

.:~

,~

v vlIVf

~V

/
/

~Ig

1.0 VOLT

//

30 40

e+M

200

60 80 100
SIN

300

50 VOLTS
II

VI

C'1

0
0

~/
o- -0'" qjf
~

0<:'>

150

o 100

o 80

./

,/

70
(/) 60
~ 50
<0

of::}

...J
U

.................

l..-------'---

a::
I 40

l.-----""
-I---

~ 30

-.........

"""

<{

20

>a::

15

10
8

500 700 1000

<,

7
6

5
4

1\

IIO'

0 - t-.-0
0
o

1.5
I
I

0',

II

f--0

87 GOJ!

V
3

~-Cb

0)0)

07 ~o'

/h

//-

//

//
//

'ff"

//'

V)~l?'

vllrJ'

//1

s.

Figure
F cos 8+ M sin 8 and M cos 8 F sin () curves for a typical transformer

t:.---O

0.1 VOLTS
F

TRANSIENT PERFORMANCE

The use of the secondary coil resistance

and saturation curve for determining per-

~0.5 VOLTS

~~~~
20

0 . 59
4.5
0 . 28
3.1
4.7
O. 75
5 .4

At larger currents, where errors are

much larger, the test values do not agree
with calculation.
This design is typical of two-coil
through type transformers, and the twoohm equivalent internal reactance is also
typical for this ratio. The burden may
often be less than the equivalent internal
impedance, so that the performance of
the transformer may sometimes not be
much affected by burden.
If the ratio is higher than 1,400/5 the
internal impedance will be higher, increasing approximately as the square
of the ratio, and if the core section is increased the internal impedance will increase almost in proportion. The result
is that a. two-coil through type current
transformer of conventional design will
not perform well at primary currents
much above 25,000 primary amperes,
regardless of the amount of iron in the
core.

(1)

//

lJLI

Per Cent Ratio Error

Test
Calculated

VII. Other Uses for Saturation

Curves and Secondary-Coil
Resistance Data

Vvv~~
~

vtf
r~

/~

/V/

/ I

aY/
.rI
II

/1/~~

//

/ V//

I/o

5 678 10

TRANSACTIONS

()

1/

!?

v<J

10 VOLTS

~ / /~V

).0 VOLTS!

r;;

l/

L0/,{.~~~

Cb

.<;1

~r--,(()
~~~
F!-.-r/{l lL

?;1
8

/
/

1/

II
.......

(()

O'

Secondary
Current
Amperes

0.1 VOLT

20

OhlTIS

{ 50 ....... 0 . 5
75
3.0
~ 25
0.1
2 . o. . . . . . 50
3 .3
55
4.2
4 0
{ 25
O. 7
. ..... 35
5.5

1I09-S.B

200

952

2=1.98=2.0

15

1 0
.

I//lf

5 678 10

300

CI)

vvVVVv ~

2.15)2
-0.345

0)

1000
800
700
600
500
400

8w

'l/

.~~

1/ /V/

F COS

/ V V/#

PJ')V /

OJ

(j

>

///

V:~
o:Ye J Vo: 0'!J
II 0'0'1;>/ ~
/ /
JV V1J ~~o~

Burden
Ohms

/ / / / I/
/ / / rff"

7/
//

1/

CI)

0'

LO

/ V Vv v~

//.

// /

1.0 VOLT

20

Q O'CbO)

/d0~0

/ I'~ ()

/V V~ /N
o

280

The details of calculation need not be

given, but the test and calculated data
are-

11// ~

10
8
7
6
5

~ 30

all

I 40

// /

60
50

CI)

'/

10 VOLTS

100

~(-

~~~~~\\
0~

~ 150

/ v I/V/

o~t;e;~~~~

XL=

V I//~~

LO

k-:VVv~ ~~

300
200

W
~
/ / V.#'

II

50 VOLTS

O' Q)
I - - - O'

results would be obtained from measurements on only one exploring coil, as only
slight displacement of the primary conductor causes considerable shifts in leakage flux, which has negligible effect on
performance, but which causes considerable difference in individual coil voltages.
The total voltage of two exploring coils
(Ex) was 2.15 volts at I s= 15 secondary
amperes, and thus

30 40

cos e -

60 80100
M SIN

200 300

500 700 1000

Wentz-Current- Transformer Calculation

These are calculated from the watt eF) and

reactance eM) components of the open-circuit
secondary saturation curve. Cos 0 is the burden power factor
ELECTRICAL ENGINEERING

formance at transient currents has been

fully described in previous papers.v?

(2)

PARALLEL CONNECTION

When current transformers are connected in parallel, the primary current of

one of the transformers may be very low,
or even zero. If it is very low, while the
burden voltage is relatively high because
of current supplied by the other transformers in parallel, the effective burden
impedance (:volts/current) is very high,
higher than any usual burden, and the
performance of the current transformer

cuit for calculations which includes an

internal reactance. This simplifies the
calculation, and isolates and defines a constant of the transformer: the equivalent
internal reactance of a current transformer is that reactance which as burden
would require an additional flux equal to
the maximum measurable value of leakage flux in the core. A formula for the
reactance in turns of exploring coil measurements is included.
The errors in the method are on the
conservative side. The real exciting current is less than the calculated value, and
at the higher values, corresponding to

101 r----r----,-~~-~--"'T"-..,--__r_-~__,

240

~
a:
~ 1001---+---+--BUROEN 0.1 OHMS 90

% P.F.

I
I

U)

ISI-----4-----t-+---+-+----+-+----t--t-----f

en 160 T I
/
II /

::E
a:
I

...

I
/

"..

,.-

/'

/'~

,.,

./

200

w
991---+---+-)( CALCULATED POJNTS
o TEST POINTS

--_-- -- ---- ------ ---- ---- -- -- -- .-

o.sss-cose

1.0
I

a:
w
Q.

the common figure of 10 times normal,

or 50 amperes secondary current,
F cos 0+ M sin 0 must be 2.5 to produce an error of 5 per cent. Then the
magnitude of the voltage, according to
the F cos 0+ M sin 0 curves, at which
F cos 0+ M sin 0 will be 2.5 is one general
measure of the excellence of the current
transformer (see figure 5).
The resistance of the secondary winding and the equivalent value of X L determine how much must be added to the
connected burden to determine how much
burden the transformer has to carry.
Hence these figures also are measures of

,",.-

........

" " ' l......

" ....

...

-c:

0.91

---

. -;:.

0.9

--

1--

--

l---- 1--::' ~

0.8
,.=:::

-- -:::::-

0.6
I-_~

..-

0.1=

cose

"

/77

,'1/ ~f/

~ 120 rr/lr;

2
>-

a::
c(

SECONDARY AMPERES

Figure 6.

Calculation versus test

Wound-type current transformer calculated

from figure 5, and secondary resistance = 0.31,
leakage reactance 0.638

at such a burden will seldom be given

by the standard performance curves.
The performance can then easily be determined by the method given.
If the primary current is zero (primary
circuit open) the exciting current drawn
by the dead current transformer can be
read from the same saturation curve, and
its effect on the current supplied to the
burden calculated in the same direct way.

VIII. Constants Determining the

Excellence of a Current
Transformer
The real excellence of a current transformer is determined by examination of
ratio and phase-angle data. However,
it is not economical, as has been pointed
out, to measure ratio and phase angle at
every possible condition, and comparison
of overcurrent data on two current transformers is usually awkward, and often
inconclusive.
When current transformers operate
relays, ratio errors of less than five per
cent are usually of no consequence. The
power factor of the total burden with the
usual relays will be in the neighborhood
of 60 per cent. If the l?rimary current is
OCTOBER

1941,

VOL.

60

o
z

I:/ill

I: ;'
80
I'
I'

II
II

M 40 I
V)

o
o

0.4

0.8

1.2
1.6
2.0
2.4
2.8
SECONDARY CURRENT - RMS AMPERES -

the excellence of the current transformer.

Thus two indices' of excellence which
can be readily compared are:
(1). Secondary
voltage
at
which
F cos 8+M sin 8 for 60 per cent power
factor is 2.5.
(2). Internal resistance and reactance.

IX. Conclusions
A simple method for calculation of current transformer performance has been
developed. It is not new in fundamental
principles, but it is an improvement and
simplification of the most accurate existing method. 3
It is more direct in its use of leakage
flux; it uses a real measured value in its
calculations; furthermore, the measurement of flux is easily done with exploring
coils without other special equipment.
The exploring coil can be threaded
through the space between core and coil
in many transformers, especially older
models.
The method is flexible. If further experience shows the necessity for further
division of the core, two sets of exploring
coils can be used with a division of the
core into three parts. It can be made to
be as accurate as desired.
It proposes a simple equivalent cir-

Wentz-Current- Transformer Calculation

Figure

3.2
FCOS

3.6

e +M

4.0
S'IN

4.4

4.8

7. F cos 8+M sin 8 curve For a

through-type current transformer

F cos 8+ M sin 8 is the component of exciting

current which must be added to the secondary

output current to get primary input current reversed (1/1 ratio basis). Example: burden
four ohms at 80 per cent power factor,
secondary current 50 amperes, exciting current
(from curve) at 4X50=200 volts=1.8 amperes.
Primary current reversed = 51.8,
ratio = 100 (51.8/50) = 1 03.6 per cent

saturation of the core, where the ratio

error usually exceeds five per cent, the
calculated ratio error is usually much
too high. Thus, the method is useful
even where it is not too accurate.
Curves are proposed for simplification
of the vector algebra. This expedient
is of great value when many calculations
are to be made. However, this part
of the method is not essential. Any correct method of vectorially adding exciting current to secondary current can
be used.
The performance under transient conditions, at unusual burdens or currents
(including performance of parallel transformers) can all be determined from the
characteristics listed above, using available methodsv? for transient performance,
and the method described above for performance at unusual burdens or currents.
TRANSACTIONS

953

The performance of a current transformer is shown to depend on a few simple

constants of the transformer:
(1).

Internal resistance

(2).

Internal reactance as defined

(3).

Open-circuit saturation curve

Comparison of (1), (2), and one point

on curve (3) allows formation of judgment about the relative excellence of
transformers.

Appendix I.

Equivalent Circuit

for Current Transformers

An inductance burden on a current transformer requires a voltage 90 degrees ahead
of the current, and as the flux to induce the
voltage is, in turn, 90 degrees from the voltage, the flux is in phase with the primary current. So we may say:
1. The flux required by an inductive burden is in
phase with the primary current.

Furthermore, it is well known that2.

The leakage flux is also in phase with the primary current.

We can combine statements (1) and (2)

and say-the flux required by an inductive
burden has the same phase position as the
leakage flux.
As both fluxes are directly proportional
to the current, they are equivalent fluxes,
and the leakage flux can be represented, if
desired, by an equivalent inductance in the
burden.
As shown by Mr. Sinks, a the leakage flux
at the center of the core inside the secondary
coil has, in effect, a negative value, but has
higher values in each succeeding element of
the core. For purposes of analysis, we can
separate the core elements and represent
each one by an equivalent exciting branch.
The fact that each element carries a certain
leakage flux can be represented by an equivalent inductance in series with the current
circuit, as in figure 3. The fact that the

954

TRANSACTIONS

leakage flux at the center corresponds to a

negative inductive reactance is represented
by a capacitance in series with the first
branch.
For practical work in which the core is
divided into only two parts, one of which is
assumed to carry no leakage flux, the circuit
takes the simpler form shown in figure 4,
in which the series equivalent inductance is
determined according to appendix II.

The leakage flux q, L can be expressed as an

equivalent reactance referred to the secondat'l winding of N, turns and peak current
2 Is

(6)

(7)

Appendix II. Exploring Coil for

Flux Determination
If the secondary winding of a current
transformer is short-circuited, the total real
linkages of flux with the secondary coil are
just enough to circulate the current against
the resistance of the winding. This flux
can be calculated as an equivalent flux cPs
which links all the secondary turns.
I sR s I 0 8

cPs = 4.44fN s

(1)

Is, R s, N, = secondary current, coil resist-

ance, and turns, respectively.

This amount of flux may be considered as
a working flux flowing all around the core.
If an exploring coil is placed on the outer
part of the core at a point where leakage flux
exists, it will develop a voltage corresponding to cPs plus (vectorially) the leakage flux

cPL.

The exploring coil voltage will be(2)

j expresses the fact that the vector cPL is at

right angles to the vector cPs. In absolute
values,
(3)

(4)

(5)

Wentz-Current-Transformer .Calculation

Bibliography
1. . PAPERS ON CURRENT TRANSFORMERS, H. W.
PrIce and C. Kent Duff. Bulletin 2, section 4,
University of Toronto school of engineering research, 1921.
2. CALCULATION OF RATIO ERRORS AND PHASE
ANGLE IN CURRENT TRANSFORMERS, WITH P ARTICULAR REFERENCE TO LEAKAGE FLUX, E. C.
Wentz. Thesis for the University of Pittsburgh,
1936.
3. COMPUTATION OF ACCURACY OF CURRENT
TRANSFORMERS, A. T. Sinks. AlEE TRANSACTIONS, volume 59, 1940 (December section), pages
663-8.
4. OVERCURRENT PERFORMANCE OF BUSHINGTYPE CURRENT TRANSFORMERS, C. A. Woods and
S. A. Bottonari. AlEE TRANSACTIONS, volume
59, 1940 (September section), pages 554-60.

5.

A PROPOSED METHOD FOR DETERMINATION OF

CURRENT- TRANSFORMER ERRORS, G. Camilli and
R. L. TenBroek. AlEE TRANSACTIONS, volume
59, 1940 (September section), pages 547-50.

6. RING-TYPE CURRENT TRANSFORMERS-LIMITATIONS IN DESIGN, A. M. Wiggins. Electric

Journal, December 1927, page 621.

7.

A STUDY OF THE CURRENT TRANSFORMER WITH

PARTICULAR REFERENCE TO IRON Loss, P. G. Agnew. Reprint 164, Bulletin of the Bureau of
Standards, volume 7, number 3.

8. CURRENT TRANSFORMERS AND RELAYS FOR

HIGH-SPEED DIFFERENTIAL PROTECTION, WITH
PARTICULAR REFERENCE TO OFFSET TRANSIENT
CURRENTS, E. C. Wentz and W. K. Sonnemann.
AlEE TRANSACTIONS, volume 59, 1940 (August
section), pages 481-8.
9. CURRENT TRANSFORMER-EXCITATION UNDER
TRANSIENT CONDITIONS, D. E. Marshall and P. O.
Langguth. AlEE TRANSACTIONS, volume 48, 1929,
pages 1464-74.

ELECTRICAL ENGINEERING