TM 11-6625-1559-12
b. When a filter under test has deep rejection
(2) Connect the output of the transformer
notches within an octave band, the distortion of
or coupling device of the stage being tested to
the input of the auxiliary broad-band amplifier;
use of an auxiliary low-pass filter that is set to
again using a small capacitor such as that de-
cut off just above the octave band. Also, use the
scribed in (1) above. Be careful to prevent stray
lowest setting of the RF OUTPUT control that
will provide an adequate display of the response
(3) Connect the output of the auxiliary
curve.
broadband amplifier to one DETECTOR jack on
TECTOR jack with Adapter-Connector UG-
1640/U. If the output impedance of the broad-
back to back as shown. However, if the trans-
band amplifier is not 50 ohms, use a matching
formers are not identical, the response curve will
pad.
represent the composite response of both trans-
(4) Perform the procedure described in
formers rather than the curve of an individual
unit.
display on the oscilloscope screen.
(5) Check the effects of the feedback cir-
3-12. Measuring Unloaded Bandwidth and Q
cuits on the stage being tested. Vary the gain of
pattern changes, feedback originates in the ear-
lier stages. Vary the gain of the tube following
The sweep generator may be used to analyze ex-
the stage; if the oscilloscope pattern changes,
tremely narrow-band radiofrequency components,
feedback orginates in the later stages of the am-
such as resonant cavities and high-Q tuned cir-
plifier under test.
cuits. Make these measurements as follows:
a. Connect the input of the cavity to be tested
(6) Observe the pattern on the oscilloscope
to the RF OUTPUT jack through a test cable
screen; make adjustments and measurement as
Use a loose coupling method.
b. Connect the output of the cavity through a
(7) When the individual stage to be tested
test cable to one of the DETECTOR jacks on
is an input or output stage of an amplifier,
the sweep generator. Terminate the other DE-
neither of the above methods will give complete
TECTOR jack with Adapter-Connector UG-
satisfactory results. The use of a voltage compar-
1640/U. Again use a loose coupling method.
ator will provide a more accurate and convenient
c. If the device to be tested is a tuned circuit
method of marking such tests.
and does not have a self-contained coupling loop,
make up a small coupling loop for use as an
input circuit. Use a small capacitor to couple
3-11. Testing Frequency Response and
the test cable to the output of the tuned circuit.
Measuring Loss of Filters or Transformers
Note. It is important to use the least coupling that
will provide adequate observation of the response curve.
The general technique for testing radiofrequency
In general, when the coupling gives an insertion loss of
components is similar to that used for amplifiers.
20 db as compared to a direct connection, an error of 5
Components such as filters and transformers
percent can be expected in the bandwidth. With an inser-
have no self-contained amplification; therefore,
tion loss of 40 db, the bandwidth error is reduced to 1
more system gain is required. The available out-
percent.
put voltage and the sensitivity of the preampl-
d. Perform the procedure described in para-
ifier circuits of the sweep generator will handle
graph 3-6d through 1, to obtain the best display
insertion losses up to 50 decibels. For higher
on the oscilloscope screen.
losses, an auxiliary amplifier is required that has
e. Be careful, to prevent any stray coupling
sufficient gain to compensate for these losses
between the two test cables. No response curve
and sufficient bandwidth to cover the frequency
should be displayed on the oscilloscope with the
response range of the item to be tested.
cavity or tuned circuit completely detuned or
a. Make the same connections to the sweep
shorted out. If various physical placements of the
generator, and use the same procedure to meas-
test cables do not completely eliminate the stray
coupling, use inductive coupling from one cable
to the cavity and capacitive coupling to the other.