Pulsed RF sources
We've split off a full page on
building a breadboard RF modulator using a free Hittite amplifier,
with contributions from some people at NASA! Click
here to check it out!
Pulsed RF is important in radar
as well as communications applications. Read this tutorial to get
some basic knowledge about how pulsed signals can be obtained in
the laboratory, as well as in real systems. Build the breadboard
drain-pulsed amplifier described
on a separate page and you will have a useful piece of test
equipment for your microwave laboratory, for very little cash, you
This page came about through
some conversations with a Stanford University student who was is
need of a laboratory setup for pulsed RF measurements, and has grown
due to popular demand!
There are two easy ways to create
a pulsed signal from a CW signal. One is to use an RF switch, the
other is to use a pulsed RF amplifier. In our first example we will
show you how to create a pulsed signal using an RF amplifier.
For now we will assume that only
rectangular pulsed waves are desired, that no amplitude modulation
or pulse-width modulation are required. Maybe someday we will deal
with a more complicated modulation scheme...
If this topic interests you,
also see our page on oscilloscope measurements.
We also have a pretty good page on microwave
pulsed waveforms using a switch
pulsed waveforms using an amplifier
a breadboard pulsed source (separate page)
First, here are some definitions
that have to do with pulsed RF.
A continuous wave signal
is an RF signal that maintains full power, i.e. it is not switched
off periodically. A pulsed RF signal is periodic in that
is has two distinct and recurring states; the on state state provides
a stable medium or high power signal, the off state has much reduced
signal strength (ideally no signal is present in the off state).
The video input terminal
of a modulator is the signal that causes the RF waveform to switch.
The word "video" here was adopted from television, where
the video signal (the picture) is carried on a VHF or UHF signal.
The pulse envelope is
the shape of the RF pulse after it has been transformed through
a detector circuit.
Pulse rise time is the
time needed for the pulse envelope to pass two conditions, such
as 10% of final value to 90% of final value. Pulse fall time
likewise is the time from 90% to 10%.
The pulse repetition frequency
(PRF) is the same as the repetition rate (aka rep
rate) and is the frequency that the signal is turned on and off
through a full cycle, measured in Hertz. It is the reciprocal of
the pulse period.
The pulse width is the
length of time that the RF is switched on in one pulse period.
The duty factor is the
time that the pulse is on, divided by the period. A 100% duty factor
implies that you have a continuous wave signal. Duty factor can
be expressed in decibels as well as percent; a 10% duty factor is
A square-wave pulsed RF
signal is a special case in which the RF signal is on and off for
the same duration of the period. A square wave has 50% duty factor.
The peak power is the
signal power during the pulse. The average power is equal
to the peak power times the duty cycle.
The on/off isolation is
the ratio of the signal, usually expressed in dB. If you thought
"off" meant "off" in an analog world, you were
Whether you use an amplifier
or a switch to modulate your signal, note that if you amplify
after the pulse is created, but you may add some appreciable
noise power between the pulses.
pulsed RF waveforms using a switch
Check out our page on RF
switches for more general information on switches.
A SPST RF switch can be placed
in a signal path, and by operating the video signal at the desired
rep rate. Whether your SPST switch is terminated or not is not a
problem, so long as all active components in your chain are unconditionally
stable. An unterminated SPDT switch presents a horrific VSWR
to the RF signal in its off condition, while a terminated switch
tries to provide a fifty-ohm match on or off.
The rumor that FETs switch slowly
is exaggerated. The speed that you can switch a semiconductor at
is determined by simple RC time constants. RF switches can be designed
using FETs that switch in as little as one nanosecond. In a FET
switch the bias line to the gate is usually a high-value resistor
to choke off the RF, sometimes as high as 10,000 ohms. Accounting
for a switch-FET's gate input capacitance on the order of one pF,
provides an RC time constant of 10 nanoseconds. The trick to making
a one-nanosecond switch is to use a smaller resistor (250 ohms maybe),
and perhaps some series inductance to choke off the RF.
To drive a FET switch, we prefer
to use logic from the ACT family (advanced CMOS), because these
devices are relatively fast, and their outputs can pull almost from
rail to rail. In the figure below, a single-chip quad X-OR gate
is used to drive two switches (SW1 and SW2) with dual complimentary
logic, which is needed for FET SPDT switches. In this case the logic
voltages to the switch are 0 and +5 volts. An input-high condition
at SW1 drives SW1A low and SW1B high, while an input low at SW1
drives SW1A high and SW1B low. This driver circuit should provide
switching times of 10 nanoseconds or less. Check out our page on
digital logic for microwave circuits!
Microwave FETs require negative
switching voltages, while the circuit shown above provides positive
voltages... what do we do about that? There are three solutions.
You can design your switch so that the FETs are "floating"
at +5 volts (create an RF ground with a capacitor), or you could
float the entire switch at +5 volts on your hybrid or circuit card,
or you could run the 54ACT86 X-OR gate with pin 14 grounded, and
pin 7 at -5 volts. Note that the last solution will required negative
voltages for SW1 and SW2 commands.
PIN switches can easily provide
<1 nanosecond rise time if care is taken in the driver circuit
implementation. But PIN diode switches are a pain to use because
you have to drive them with DC current (as opposed to FETs which
require only a voltage).The best driver circuits for PIN diodes
are made by Whozzat.
Off isolation of a FET or PIN
switch depends on whose switch you are using, and over what frequency
band. If you can't find a switch with great isolation, just gang
two of them in series.
pulsed RF waveforms using an amplifier
Amplifiers can be rapidly turned
on and off by changing the supply voltage and current on the output
terminal (the drain, in the case of a FET). This can be effected
by applying a pulsed waveform to the drain or the gate terminal
of a FET-based amplifier.
Pulsing an amplifier can often
give you better isolation than a switch. This is because when an
amplifier is turned off, it can give you beaucoups isolation. You
might get greater than 50 dB difference between the on-state and
the off-state S21 for an amplifier. Unfortunately, amplifier vendors
are not in the habit of supplying S-parameters for their products
when they are turned off, so you might have to measure this yourself!
This topic is the subject of
our RF pulse modulator breadboard
circuit. Drain pulsing is simple in concept, you vary the drain
voltage periodically between two states: zero volts, and the voltage
VDS that your amplifier operates at. You need a circuit that is
capable of applying drain voltage, on command from a pulse generator.
We'll show you two ways to do do this, based on how much drain current
you need to supply.
For medium power amplifiers,
perhaps up to 100 mA drain current, you can use a MOSFET driver
such as TSC427 to directly
pulse the amplifier's drain bias (see figure below). These are commonly-available
from many suppliers, and you can order them directly from the Somethingkey
catalog. MOSFET drivers are usually sold in dual-driver ICs,
with a choice of inverting or non-inverting operation (TSC426 is
dual inverting, TSC427 is dual non-inverting, and TSC428 provides
one of of each). Be sure to terminate any unused inputs like a good
digital geek. Although no one else would tell you this little secret,
you can actually decrease the rise and fall time of the pulse by
operating both of the dual drivers in parallel, but you will probably
void the warrantee!
Note1: it is good practice
to place a terminating resistor at the input to the driver circuit
for high speed operation. Depending on your pulse generator, 50
to 200 ohms should take care of it, but remember to analyze how
much hear the termination will dissipate if the modulator is left
Note 2: we prefer the TSC427
dual non-inverting driver for this application, because once it
is terminated on the input, the modulator will remain in the "off"
state if the input pulse signal is lost.
The data sheet for a typical
MOSFET driver specifies 20 nanoseconds rise time. This is worst
case, when up to a 1000 pF load is being driven (a large MOSFET
has pretty high capacitance). When drain pulsing you will be driving
a capacitive load if a stability capacitor is required to make your
amp behave, but the amplifier also provides a low-impedance path
to ground, which helps discharge the capacitance. Even with 1000
pF stability cap, on and off times will be faster than 20 nanoseconds,
and you heard it here first! If it isn't, take a look at the off-chip
drain bypass (stability) capacitor(s) your circuit has. If there
is a 10,000 pF cap on the drain, maybe it will work OK with just
a 100 pF cap. A MOSFET driver should switch in less than 5 nanoseconds
under this condition.
For higher-power amplifiers,
you can use a P-channel MOSFET to supply several amps of pulsed
current, as shown on the figure below. Here you use the MOSFET driver
for what it was intended for (turning the MOSFET on and off). Depending
on the size MOSFET you choose, this may slow down the circuit to
20 nanoseconds. Unknown Rectifier
is a great choice for MOSFETs, they are so popular that their trademarked
name of "HEXFET" is often used by engineers when they
mean "MOSFET" just like "Xerox" is used by everyone
to mean "copy"!
Note 3: we prefer the TSC426
dual inverting driver for this application, because once it is
terminated on the input, the modulator will remain in the "off"
state if the input pulse signal is lost (the P-channel MOSFET
acts like a second inverter stage).
storage refers to the capacitance that actually powers the amplifier
when the pulse is in transition. The inductance of the power supply
leads prevents the DC current from instantaneously jumping from
zero to IDS. The amount of charge storage that you need can be simulated
or estimated, but in practice is must be verified experimentally.
You will know you have enough charge storage when your drain voltage
waveform droops just a few tenths of a volt during the pulse. Start
by using no less than 4.7 uF.
You can use surface mount tantalum
caps for charge storage, with voltage rating higher than the power
supply VDD (obviously!) If your drain supply is 6 volts, use at least
a 10 volt rated cap. Be sure to observe polarity. Axial lead caps
could be used, but the "enemy" here is electrical lengths
between everything. Don't rig the circuit up with three-foot test
leads and expect it to work!
Stability capacitors refer
to off-chip caps that are needed to prevent low-frequency (UHF)
oscillations. In the case of a pulsed circuit, you want to minimize
the stability capacitor value for high switching speed. Try using
100 pF at first, and if necessary, increase it to 1000 pF.
A FET or PHEMT amplifier can
be shut off by pinching off the drain current. Here you need to
switch the gate from its quiescent value (on state, probably about
halfway to pinch-off), to the off condition (below pinch off).
There is one obvious advantage
that is held by gate pulsing an amplifier versus drain pulsing:
you don't have to switch a ton of current, the gate terminal is
practically an open circuit. At first you might think that every
pulsed amplifier design would use gate pulsing. But you would be
very wrong, most use drain pulsing. This is because of the two disadvantages
of gate pulsing.
First, many power amplifiers
are operated close to their gate-drain voltage breakdown limit at
the quiescent point. When you drive the gate into the pinch-off
region, you are increasing the gate-drain voltage potential, and
perhaps entering the breakdown region. This increased stress can
present a reliability problem.
The second issue is that you
have to switch between two voltages that are not all that far apart,
and the gate voltage during operation is far more critical to performance
than the drain voltage. V1 and V2 might be -0.9 volts and -1.5 volts
for a PHEMT power amp. Any noise or settling issues on V1 and your
pulsed signal will suffer.
For a good gate pulsing circuit,
we recommend the Analog Device's AD8036
"clamping amp", shown below. It lets you set up on and
off gate bias voltages independently; for example, if you are using
a PHEMT power amp you can set VG(on) to -0.9 volts, and VG(off)
to -1.5 volts. You still need charge storage on the drain bias lines,
and stability caps on gate and drain biases, as near as possible
to the amp. Like all op-amps, the AD8036 can be configured as an
inverting or non-inverting amplifier. Yes, we need to add some resistive
feedback to the figure!