Text preview for : 5980-0798E.pdf part of HP 5980-0798E HP Publikacje 5980-0798E.pdf
Back to : 5980-0798E.pdf | Home
Agilent Technologies
Pulsed Measurements with
the Agilent 8720ES and 8753ES
Network Analyzers
Product Note
Introduction
High-performance vector network analyzers are used to
characterize the frequency responses of various RF and
microwave components. While most network analyzers
are used to measure continuous signals, some users have
a need to pulse the RF signal on and off. The need for
pulsed signals arises for various reasons. It might be
because the total power input to the device under test
must be limited to prevent overheating, or because the
component will only operate under pulsed conditions, or
because the component behaves differently in response
to pulses than it does to steady state inputs. Whatever
the reason, pulsed RF creates a unique set of conditions
that must be addressed in order to produce reliable
measurement results.
It is commonly presumed that only network analyzers
such as the Agilent 8510C or the 85108A that incorpo-
rate sophisticated timing circuits, dual reference chan-
nels, and wide IF bandwidths are suitable for
measurements of pulsed systems. However, the aim of
this product note is to show that under certain condi-
tions, network analyzers such as the 8720ES and 8753ES
can produce good results at a much lower cost. The
principles in this product note apply to either model,
and to some earlier models as well; for brevity we will
refer only to the 8720ES throughout the rest of this
product note.
In general, for a standard network analyzer to be able to
make pulsed measurements, there are three conditions
that must be met. The first condition is that the RF
source must be externally modulated, since a standard
network analyzer lacks any modulating circuitry. The
second condition is that if the pulse is a single or ran-
domly timed pulse, it must exceed a width of 400
microseconds. Also for this case, the user must supply
external triggering circuitry with variable delay between
the trigger signal and the RF pulse. The third condition
is that if the RF signal is a continuous pulse train, the
minimum pulse width must exceed 500 nanoseconds and
the pulse repetition frequency must be 20 Hz or greater.
The reasons for these constraints will be explained later.
(Note: the techniques described in this product note
cannot determine the impulse response of a device.)
2
Measurement theory
In general two types of pulse measurements may be Two types of low PRF measurements are possible. The
made: Those with a pulse width greater than the net- first type is commonly referred to as the pulse-profile
work analyzer's response time, and those with a pulse mode of operation. In this mode the source is operating
width less than the network analyzer's response time. at a CW frequency, and the analyzer is armed to trigger a
These will be designated as low PRF measurements and sweep. The RF signal is pulsed on after the sweep has
high PRF measurements, where PRF stands for pulse been triggered. Because the 8720ES does not contain
repetition frequency. Each of these conditions must be any circuits for controlling the delay between receipt of
treated as a separate case, since each has its own unique a trigger and the start of a sweep, the user must provide
requirements. this delay externally. If the pulse is wide enough that
several measurement points can be taken during a single
Low PRF measurements pulse, then we can see the shape of the pulse over time
as it is transmitted through the device under test (DUT).
Every network analyzer has a sample response time that Each point represents the same frequency over time.
depends upon fixed internal characteristics such as CPU This mode is useful for determining such things as pulse
speed, and variable characteristics such as the IF band- droop. The reason specialized network analyzers such as
width of the system. It is characteristic of the digital fil- the 8530A have an advantage over the 8720ES in this
tering used in the 8720E and 8753E that wide IF mode is that their minimum pulse width is much narrow-
bandwidths result in faster sample times, while narrow er, 1 microsecond or less.
IF bandwidths improve dynamic range.
The second type of low PRF measurement is commonly
If a given pulse is wider than the minimum analyzer referred to as point-on-pulse mode. In this mode the
response time, and the proper timing is established source is tuned, the RF pulsed on, and the analyzer is
between the external signal that triggers the analyzer armed to trigger on each individual pulse after a prede-
and the signal that triggers the pulse itself, we can be termined time delay. Each measurement point repre-
guaranteed that a valid measurement point will result. sents a different frequency. This mode is useful for
This is because during the entire time the analyzer is determining the frequency response of the DUT when it
sampling, the RF signal is on. We define this condition to cannot be powered up for long periods. Because the
be a low PRF measurement. Because the 8720ES trig- 8720ES does not begin to sample instantly when a trig-
gers on a falling edge, the external trigger signal must ger signal is received, the user must provide an
often be inverted to generate this edge before the RF adjustable delay externally. Specialized network analyz-
pulse appears. Also, it should be noted that the analyzer ers such as the 85108A have built in triggering circuits
does not begin sampling until some time has elapsed to adjust the delay between the time the RF pulses on
after the trigger signal has been received. Table 1 shows and the analyzer takes a measurement. The 85108A also
the relationship between the IF bandwidth and the mini- uses wideband detection instead of the narrowband
mum pulse width and delay for the 8720ES when operat- detection used on the 8720ES, giving it an effective IF
ing in the low PRF mode. Note that the trigger delay bandwidth of 1.5 MHz. This means the sample response
varies slightly between instruments. time is much faster, which allows the user to sample at
widely variable times within pulses as narrow as 5
Table 1. Using external triggering for low PRF microseconds.
IF Bandwidth Minimum pulse Trigger delay
width for a single (typical)1 There is a third type of measurement known as peak-
measurement point (milliseconds) response mode. This is when the pulse is narrower than
(milliseconds) the analyzer's response time, but not so narrow that the
6000 Hz 0.40 0.70 analyzer fails to register any signal. It may be considered
the transition between low PRF and high PRF measure-
3700 Hz 0.53 0.70
ments. Because of the critical timing requirements, this
3000 Hz 0.60 0.70
mode is not recommended for the 8720E or 8753E.
1000 Hz 1.10 0.70
300 Hz 3.20 0.70
100 Hz 8.90 0.70
30 Hz 32.0 0.70
10 Hz 120 0.70
1. The digital IF takes several ADC samples before it starts taking the "real" data for the point. It uses these samples to compute the IF gain setting; it does this even if IF gain is
controlled manually in the service menu. The most common value for this delay is 0.70 milliseconds, but some instruments exhibit delays of 0.81 or 1.00 milliseconds. There is no
way to tell which is which, except by experiment. Once the trigger delay is determined, it is constant within a given instrument (i.e. there is no dependence on settings). The user
is always safe by presuming a shorter delay than given in the table, however the minimum pulse width may have to be increased correspondingly to compensate for any
difference between the presumed delay time and the actual delay time. 3
High PRF measurements
If a given pulse is narrower than the minimum analyzer Now consider the spectrum of a rectangular pulse train,
response time, it becomes impossible to measure a sin- as shown in Figure 2. This is a MatLab