IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 31, NO.

2, MARCH,APRIL 1995

Comments on Harmonics: The Effects

on Power Quality and Transformers
Keith H. Sueker

In this paper, the two gentlemen have documented well a series

of real harmonic problems in an industrial plant. They made the
excellent point that defective components in SCR drives can produce
large increases in harmonic currents and that such drives require
maintenance. I suggest, however, that several points are subject to
misinterpretation by readers. Please note the following:
1) The article mixes references to voltage harmonics and current
harmonics. SCR drives produce current harmonics which, in
flowing through source reactances, produce voltage harmonics.
The current harmonics do not change in magnitude as they flow
toward the source but the voltage harmonic levels vary as the
reactance between the source and the point of measurement.
2) The statement is made that Recently, K-rated transformers
have emerged with a greater ability to withstand stress damage
and overheating. The K-rated transformers have no more
or less ability to withstand stress or overheating than any other
transformers. They are simply designed and constructed in such
a fashion that the eddy current and stray losses due to the
expected harmonic currents will not result in overheating.
3) A transformer derating factor is derived from the crest factor of
the (current) waveform. I am concerned that this procedure will
not necessarily result in the proper value since the K factor
of the current waveform is proportional to C(itt * h ) 2 where
h is the harmonic order and ih is the current at that harmonic.
The K factor is that factor by which the eddy current and
stray losses must be multiplied to amve at the total losses with
a nonlinear load. The procedure is described in (257.1 10.1986.
Unfortunately, it requires a separation of eddy current and stray
losses by measurements of 60-Hz copper losses and dc winding
resistances. I believe the procedure is considerably beyond the
capability of most users.
4) The attenuation of harmonics by an isolation transformer
is described. This attenuation occurs for the voltage THD
because the leakage reactance of the transformer increases
the commutation angle in the drive and thereby reduces the
higher frequency harmonics. There is no significant effect
on the lower frequency current harmonics or the current
THD. Transformers upstream from the isolation transformer
are subjected to roughly the same level of current harmonic
distortion. The statement that Isolation transformers attenuate
THD by an average of three times;. . . is misleading. Except
for eliminating ground faults and some high-frequency common
mode disturbances, isolation transformers do nothing more than
can be done with line reactors.
I am certain the authors understand well the distinctions I have
made. However, the whole subject of harmonics is so poorly understood by many engineers that considerable care should be taken with
details in a tutorial article of this sort.

Manuscript received July 15, 1994.

The author is a consultant at Pittsburgh, PA 15235 USA.
IEEE Log Number 9409115.
R. D.Henderson and P. J. Rose, IEEE Trans. Ind. Applicat., vol. 30, no.
3 , May/June 1994.

405

Authors Reply by Robert D. Henderson and Patrick J. Rose

We appreciate Mr. Suekers comments; however, we would like
to point out that it was not our intention to prepare a tutorial article
on the subject of harmonics, nor was it our intention to prepare a
comparison of isolation transformers versus reactors. Both subjects
are well beyond the scope of a paper of this type and length. The
purpose of this paper was to convey our findings of a study that
was initiated as the result of repeated failures of drive isolation
transformers and to present those findings to an audience of other
electrical engineers. This was stated in Section I1 (Scope).
1) In Mr. Suekers first concern, he indicates that we have
mixed references to voltage and current harmonics. In reading
other papers on this subject, we have also been concerned
by the confusion that those papers created by mixing the
same references and made a deliberate attempt not to do the
same. After reviewing the paper numerous times, the only
references that we have been able to identify that are not
clear are those which refer to IEEE Standard 519-1981 and
THD. While Section 1.C (Distortion Limits) of the paper clearly
identified that these distortion limits were applied against
the voltage waveform, Section 1.D. (Possible Solutions to
Harmonics Problems) states that Several possible solutions
exist for an area of a plant that exceeds the 5% distortion
limit established by IEEE. This certainly could have been
clarified by inserting the word voltage prior to the word
distortion. In the second paragraph of Section 1II.C (Do
Isolation Transformers Affect K Factor and THD?), we refer to
THD (total harmonic distortion) without indicating whether it
is voltage or current. Again, we should have clearly indicated
in the paper that THD is specifically a voltage distortion. The
same is true for Section 1V.E (Conclusions). We apologize for
any confusion that this may have caused.
We completely agree with Mr. Suekers statement that SCR
drives produce current harmonics which, in flowing through
source reactances, produce voltage harmonics. The current
harmonics do not change in magnitude as they flow toward the
source but the voltage harmonic levels vary as the reactance
between the source and the point of measurement. We believe
that we essentially stated that in the third paragraph of Section
1.A (Defining the Harmonics Problems).
2) Mr. Suekers second concern is that our statement recently,
K-rated transformers have emerged with greater ability to
withstand stress damage and overheating is incorrect. He states
that K-rated transformers have no more or less ability to
withstand stress or overheating than any other transformers.
They are simply designed and constructed in such a fashion
that the eddy current stray losses due to the expected harmonic
currents will not result in overheating. We believe that our
statement and Mr. Suekers latter statement are both correct.
Mr. Warren Lewis in an article, Is There a K-Factor In Your
Future, Power Quality magazine, dated Mar./Apr. 1991, states
the following with respect to K-rated transformers:
a)

b)

The primary winding has sufficient conductor capability to withstand high third harmonic circulating currents,
as well as the 9th and 15th harmonics.
The temperature rise of the primary delta winding
is limited within the rating of the insulation system

Manuscript received October 3, 1994.

The authors are with RDH Consultants, Inc., Charlotte, NC 28241 USA.
IEEE Log Number 9407682.

0093-9994/95$04 ..MI 0 1995 IEEE

406

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 31, NO. 2, MARCWAPRIL 1995

under all conditions resulting from 60 Hz (fundamental)

load current and magnetizing current as well as those
resulting from expected harmonic currents.
The primary winding and core have a lower normal
flux density to withstand the additional flux density produced by the harmonic currents from nonlinear loads.
The transformer can operate as high as +lo% of
the nominal line voltage without experiencing core
saturation and very high magnetizing currents.
The secondary windings can use multiple, smaller,
paralleled conductors instead of fat cross-section single ones. This reduces skin effect and associated ac
resistance that are considerable problems at harmonic
frequencies above 60 Hz.
The transformers neutral bus bar system can carry at
least twice the full rated line current on the secondary.

3) The formula that Mr. Sueker is referring to in his third concem

is:

This is the same as CBEMA (Computer and Business

Equipment Manufacturers Associations) Df as we discussed.
This crest factor method for derating transformers with
nonlinear loads was proposed in a CBEMA information letter
in 1988. This letter was published in an article titled, Double
the Neutral and Derate the Transformer-Or Else! by Arthur
Freund, Senior Editor, Electrical Construction & Maintenance,
Dec. 1988.
As Mr. Sueker pointed out, the procedures for derating
of transformers with nonlinear loads in (37.110-1986 are
complex, requiring information that wasnt available to us.
Because we were in a production environment, we werent
able to take the transformers out of service to Derform the

required tests. Considering these constraints, we decided to use

the CBEMA Df method to see if it would give a reasonable
approximation for predicting failure, and as pointed out in the
paper, it did predict the failure of two transformers.
We would like to add that we do not recommend using
any method for derating transformers. We feel that derating
is at best a band-aid approach. In addition to requiring
harmonic information about the loads, design and/or certified
test information is also required for derating. From a real world
perspective, the probability of someone in the future adding
load to a derated transformer certainly exists. We feel that the
real solution to harmonics problems has to be in the installation
of active or passive devices at the nonlinear load that minimize
or cancel the harmonics. Unfortunately, this is not always the
most economic solution.
4) To address Mr. Suekers last concem, it was not our intention
to imply that line reactors have no attenuating effect on
harmonics. This study took place on drive systems that utilized
isolation transformers that were failing. Isolation transformers
do have the advantage of trapping triplen harmonics in the
delta winding of the transformer, as well as reducing the
higher frequency harmonics as stated by Mr. Sueker. Significant
triplen harmonics were found on the secondary side of the
isolation transformers as a result of failing drives, and were
eliminated on the primary side. We do not agree that our
statement . . . isolation transformers attenuate THD (voltage)
by an average of three times;. . . is misleading. This statement
was made based on measured results which included high
triplen harmonics on the secondary side of the transformers.
Obviously this was not an ideal situation. We believe that the
circumstances that served as the basis for this paper are the
norm rather than the exception for most drive systems installed
for any length of time.
We appreciate Mr. Suekers comments and clarifications, and we
hoDe that we have satisfactorilv addressed his concerns.