Quantcast Insertion Loss Measurements of Electrical Filters

Custom Search

full rated current load, in a free space environment, or 40 deg., when
filters are mounted in a modified NEMA Type I enclosure.  It requires
insertion loss tests at 10, 50 and 100 percent of load with the use of
modified buffer networks to extend the lower test frequency to 14 kHz.  It
limits dc voltage drop to less than 0.5 V at full rated current, and ac
voltage drop to 1 percent of rated line voltage.  It limits filter impregnant
flash point to greater than 165 deg. C. It requires bleeder resistors to
drain filter capacitor stored charge in accordance with, NFPA 70, National
Electric Code, Article 460-6 requirements.  Significantly, it calls for
submission of certificates of conformance or compliance of equipment and
materials before their delivery.  Where customer requirements include power
filters for larger than 200 amps rated load, and/or greater than 100 dB
measured insertion loss, it may be advantageous to specify parallel
combinations of filters rated at 200 amps or less, or series combinations of
filters rated at greater than 60 dB insertion loss each.  Where feasible,
power systems should be designed so that individual filter rated loads do not
exceed 100 amps, the MIL-STD-220A type testing limit of most filter
Insertion Loss Measurements of Electrical Filters.  The available
test specification for the measurement of electrical filter insertion loss is
MIL-STD-220A.  The test methods in this standard are intended to provide data
for quality control during quantity production of power line filters.  The
test conditions specified with 50 ohm input and output terminations are
satisfactory for this control purpose, but do not represent conditions that
exist in actual circuits or installations.  The power source impedances in
actual installations are typically much lower than 50 ohms at frequencies in
the filter pass band, and up to 14 kHz where the filter is required to provide
greater than 100 or 120 dB of insertion loss.  The power filter load
impedances at actual installations also vary widely, depending on equipment
loading, are not constant as a function of frequency, typically have a leading
power factor, and are often nonlinear.  Currently, methods used for in situ
measurement of power filter attenuation using current injection and
measurement probes are being developed by the Naval Civil Engineering
Laboratory, Port Hueneme, California.  Some test results to date show the
power filters providing 20 to 30 dB less attenuation from 14 kHz to several
MHz than the greater than 100 dB of insertion loss required by NSA 65-6,
(Figure 10), as expected.  In specifying filter performance tests, the use of
extended buffer networks to allow for insertion loss measurements to a lower
frequency limit of 14 kHz should be required.  The buffer network assembly
consists of a series inductor, and feedthrough capacitor.  Two of these
networks isolate the load-current source from the receiver and output meter so
that greater isolation at the test frequencies is provided through the
networks than through the filter under test.  The original MIL-STD-220A test
specification called for a lowest frequency measurement of 100 kHz and a
maximum current loading of 100 amps.  The extended buffer network provides
additional isolation to a lower frequency of 14 kHz.  In order to measure the
rated load-currents higher than 100 amps, buffer network inductors with higher
current handling capability must be provided in the test circuit.  The use of
dc to load the filter during insertion-loss testing is not a representative
loading on filters designed for use with ac power circuits, since the dc
permeability of the filter-inductor cores is different than for ac.  The use
of ac current loading of ac power filters during insertion-loss testing should
be encouraged.  Where the normal buffer networks at higher current ratings are
not available, it is feasible to use an additional power


Privacy Statement - Copyright Information. - Contact Us

Integrated Publishing, Inc.