The Delco Electronics Division of General Motors Corporation in Santa
Barbara, California, investigated the harmonic current problem as part of a
contract for the U.S. Army Mobility Equipment Research and Development Command
(MERADCOM), Fort Belvoir, to develop a 15 kW general-purpose power conditioner.
The approach they used is documented in Reference 21.
The investigators realized that no harmonic currents are generated when
single- phase power is rectified into a resistive load and that the simplest
model for a switching-mode power supply that has a near-unity power factor is
essentially a resistive load. The basic converter they were using was a resonant
converter. They divided 3-phase power into three single-phase rectifier
circuits, operated a separate isolated resonant converter from each single-phase
rectifier and recombined the outputs of the converters into a single DC output
With no additional filtering, and operating at one-third load, they found
that the 5th harmonic was reduced from 42% to 3.2%, the 7th harmonic was reduced
from 12% to 3.8%, all other harmonics were less than 3%, and the total harmonic
distortion was reduced from 48% to 6%. These figures normalize to one-third
these values at full rated load (18 kW) with measured total harmonic distortion
at rated load measured at 2.9% for a 18.9 kW load, 3.6% for a 15 kW load and 6%
for a 6 kW load. The approach has been shown to meet the Navy 3% requirement in
higher power levels with an insignificant increase in size, weight, or parts
The report concludes that the basic criterion for successful operation of the
converter configuration is that the DC-to-DC converter following the single-
phase full-wave bridge rectifier must be either a purely resistive load or have
such a low reactive component that essentially only triplen harmonic currents
are generated when operating in a single- phase mode. The triplen harmonics are
then canceled by the inherent characteristics of a balanced 3-phase
The concept of the converter approximating a resistive load is the starting
point for many electronic approaches to limiting harmonic currents or producing
a unity power factor load. The converter can then be used on single-phase power
or three converters can be combined as in the Delco configuration to limit
harmonic currents in three-phase systems.
In order to approximate a resistive load, the converter has to draw power at
near-zero input voltage, which eliminates the buck-derived configurations from
consideration. Configurations derived from boost, buck-boost and Cuk converters
are possible approaches. While isolation is not required in a single-phase
application, it is required in the Delco approach for 3-phase
A basic limitation of these approaches is the bandwidth of the control loop.
In concept, a single buck-boost converter could provide line and load regulation
as well as harmonic reduction. In practice, the control must be constant over
one- half the period of the line frequency in order to avoid introducing
harmonics. This greatly degrades the response to load and line transients. As a
result, a two-stage configuration is often used with one converter controlling
input harmonics and the other converter providing line and load
The starting point of many designs using this approach is Reference
Marco Venturini and co-authors have published several papers (References 23
and 24) on frequency changers that degenerate into a class of AC to DC
converters which draw no current harmonics from the line. These are based on
theoretical work (Reference 25) that predicts a class of converters which, given
no pulsating input power (i.e., 3-phase input) and no pulsating output power
(i.e., DC power). Power conversion and control can be accomplished by connecting
the input lines to the output lines through switches with no energy storage
devices in the circuit.
The circuit has unity power factor and no harmonics below the switching
frequency, which can be 20 kHz or higher. In concept, the total converter
consists of six switches, control logic for the switches and high-frequency
filters. Since all switching techniques are used, the efficiency is high. This
is an attractive topology even if it did not solve the harmonic problem.
The major restrictions are that the output voltage is limited to one-half the
peak AC input voltage and there is no isolation between input and output since
there are no transformers in the circuit.
The contractor that comes up with effective solutions to the 3% harmonic
current problem that do not significantly increase size, weight, or part count,
or degrade performance, will have an advantage over competition. For this reason
there is considerable industry internal R&D activity seeking solutions.
Several companies have a solution to the problem but are unwilling to publish
their results at this time.
The B revision of Military Handbook 241, "Design Guide For EMI Reduction In
Power Supplies" (Reference 26) has been expanded to discuss the reduction of
line harmonics caused by rectification.
The use of DC input power eliminates the harmonic current problem. Aircraft
power (MIL-STD-704) now specifies a 270 VDC power, which is the nominal voltage
obtained when 115/200 V aircraft power is off-line (no transformer) bridge-
rectified. There have been several studies recommending DC power for ships. The
DC voltage obtained by off-line bridge rectification of 115 VAC ship's power is
155 VDC and the voltage obtained by off-line bridge rectification of 440 VAC
ship's power is 591 VDC. Off-line techniques cannot be used to obtain 270 VDC
from ship's power; a transformer is required. However, 270 VDC has been studied
as a potential standard for submarines.