as of 7/11/98
The following manuscript was originally written under contract for the Amplifier and Projects Chapter (13) of the 1995 ARRL Handbook. About a month before the due date on the contract, I received a phone call from the ARRL Handbook Editor. He told me to stop writing.
Amplification can be defined as: The process of increasing the
magnitude of a variable quantity--especially the magnitude of a
voltage and/or current, without substantially altering any other
quality.
The (pure) sine wave is the only periodic waveform that contains no
harmonic energy. All of the energy in a sine wave is contained in the
fundamental frequency. In other words, a pure sine wave is coherent.
Maintaining the quality of sine waves during amplification is a major
concern in the design and operation of amplifiers.
Even though the following discussion pertains to the operation of
gridded electron tubes, some of the concepts apply to power FETs
because both are voltage-driven amplifying devices.
In the discourse that follows: if a voltage is said to be
negative-going, that does not necessarily mean that the voltage is
negative. It could be a positive voltage that is moving in the
negative direction. If a voltage is positive-going, or moving in the
positive direction, it could be a negative voltage or a positive
voltage that is moving in the positive direction.
In an electron tube, the cathode emits a cloud of electrons. Since
electrons carry a negative charge, and unlike charges attract,
electrons are strongly attracted by the positive voltage that is
applied to the anode. Unless something is placed in the way, much
current flows between the cathode and the anode.
The term "grid" describes appearance--not function. A grid is made
from a number of closely spaced wires or bars--like a bird cage. The
grid is placed close to the cathode. As a result, the grid has more
influence over the cathode's electrons than does the more distant
anode. Thus, a small change in grid-voltage produces a large change
in the flow of electrons. Because the voltage applied to the grid
controls the flow of electrons from the cathode to the anode, in
function, the grid acts like a valve or gate. The grid requires
virtually zero current to perform its job. As a result, the power
gain of a gridded electron tube is theoretically high.
Since like charges repel, a sufficiently negative grid can stop
electrons from traveling to the anode. As the grid is made
less-negative, the flow of electrons to the anode steadily increases.
In other words, a positive-going, albeit negative polarity,
grid-voltage causes an increasing flow of electrons between the
cathode and the anode. The reverse is also true: a negative-going
grid-voltage causes a decreasing flow of electrons from the cathode
to the anode. As long as the grid remains negative with respect to
the cathode, the relationship between grid-to-cathode voltage and
anode-current is fairly linear.
When an appropriate load resistor is connected between the anode and
its positive voltage source, the changes in anode-current produced by
changes in grid-voltage create a proportional, typically much larger,
voltage change across the resistor. The ratio of changes in voltage
across the load resistor to changes in grid-voltage is the voltage
amplification factor. It is designated by the Greek letter Mu. Since
Mu is higher at high anode-voltages than it is at low anode-voltages,
average Mu is a more meaningful number than maximum Mu. Average Mu
ratings vary from about 2 to 240. Mu is partly determined by the
spacing between the grid wires and the distance between the grid and
the cathode..
The duration of anode-current conduction per cycle determines the
class of amplifier operation. A conduction angle of 360 degrees means
that the anode is conducting current during 100% of the input sine
wave cycle. A conduction angle of 90 degrees means that the anode is
conducting current during 25% of the input sine wave cycle. Long
conduction angles produce a more linear representation of the input
sine wave. Short conduction angles produce more efficiency--and less
linearity.
Class A is defined as a conduction angle of 360 degrees. Class B is
defined as a conduction angle of 180 degrees. Class C is defined as a
conduction angle of less than 180 degrees. The subscript 1 indicates
that no grid-current flows. The subscript 2 indicates that
grid-current flows--the result of driving the grid into the positive
voltage region.
When the conduction angle is less than 360 degrees, the missing part
of the sine wave must somehow be filled in. One way of filling in the
missing part of the sine wave is by utilizing the flywheel effect of
an output tank circuit. Another way to produce a smooth sine wave is
to use a push-pull configuration. If each device in a push-pull
circuit conducts for at least 180 degrees, a smooth sine wave can be
produced.
Class A is the most linear class of amplifier operation. Class A
amplifiers produce only about 1/100,000 part, or minus 50dB,
distortion. The theoretical efficiency of a Class A amplifier is 50%.
The practical efficiency is slightly lower. Class A is used mainly in
low level amplifiers--where efficiency is not much of an issue. Since
Class A operates with continuous (360 degrees) conduction, no tank
circuit is needed to complete the sine wave. Class A is ideal for
wide band amplification.
The zero-signal anode-current [ZSAC] in Class A is set to
roughly half of the electron tube's maximum anode-current rating.
Although the meter-indicated anode-current remains constant from zero
signal to maximum signal, the instantaneous anode-current typically
varies from just above zero to c. 6-times the meter-indicated
anode-current.
The maximum available power in Class A is roughly equal to the
anode-dissipation rating of the electron tube.
The Class A amplifier can be compared with a gas turbine engine. Both
have a smooth, continuous power stroke--and neither one is very
efficient.
Class AB1 amplifiers are roughly 60% efficient. The trade-off for
increased efficiency is slightly more distortion--roughly 1/10,000
part, or minus 40db. Since most transceivers produce more than minus
36db of IMD, such an amplifier would not add significant
distortion.
The anode-current in Class AB1 varies in proportion to the
grid-voltage for about 60% of the input sine wave cycle. Thus, the
anode-current is off for about 40% of each input sine wave cycle. The
missing 40% in the output sine wave is filled in by the flywheel
effect of the output tank circuit.
In Class A or Class AB1 a somewhat unusual relationship exists
between the work being performed by the tube and the grid voltage.
The grid is operated in the zero to negative voltage region. Maximum
instantaneous anode-current, maximum anode voltage swing and maximum
peak power output coincide with an instantaneous grid-voltage of
zero--i.e., maximum stoke equals zero grid volts.
The grid must not be allowed to become positive. If the grid became
positive, electrons from the cathode would begin flowing into the
grid. Whenever grid-current flows, the linear relationship between
grid-voltage and anode-current deteriorates.
Since there is zero grid-current in Class A or Class AB1
operation--and any voltage multiplied by zero amperes is zero
watts--the driving power is usually stated as zero on the tube
manufacturer's technical information sheet. However, because charging
and discharging a capacitor requires current flow, in the real world
of conductor RF resistance, charging/discharging the grid capacitance
at an RF rate consumes some power. In a typical HF amplifier, the
drive power required for Class AB1 MF/HF operation is roughly 1% to
2% of the output power. Thus, the typical power gain is roughly 50 to
100. As frequency increases, conductor resistance increases due to
skin-effect. More R causes more I^2 R loss. As frequency increases,
the amount of current needed to charge and discharge the grid
capacitance also increases--causing even more I^2 R loss. These
losses can only be compensated for by adding more drive power.
Drivers (usually a transceiver) require a resistive load--which the
capacitive grid does not provide--so a suitable resistance must be
connected from the grid to RF-ground..
The zero-signal anode-current [ZSAC] in a Class AB1 amplifier
is normally set to about 20% of the maximum-signal single-tone
anode-current.
Tubes that are designed for Class A and Class AB1 service produce
high peak anode-current when the instantaneous grid-voltage is zero.
Typically, the peak anode-current is about three times the maximum
rated (average) anode-current. Most of the tubes that are used in
Class A and Class AB1 RF amplifier service are tetrodes and
pentodes--devices that have the advantage of grid-to-screen
amplification. Triodes are seldom used because the only grid designs
that can produce high anode-current with zero grid volts are those
that have a Mu of 2 to 5. Low Mu triodes require much more driving
voltage than a comparable tube with a screen requires. Since imposing
a high RF voltage across the capacitive grid is difficult, low Mu
triode Class A and Class AB1 power amplifiers are only practical up
to a few hundred kHz.
The most common configuration for Class AB1 operation is
grid-driven. Since grid amplification as well as grid-to-screen
amplification takes place, the resulting power gain is high. Class
AB1 amplifiers can also be cathode-driven if the tube is a tetrode or
pentode. The grid is tied to the cathode. Thus, the grid-voltage is
always 0V--so no grid-current flows. The screen is grounded. The
input signal is applied to the cathode/grid. Because input signal
voltage is applied between the grid and the screen, grid-to-screen
amplification takes place. However, since the grid is tied to the
cathode, no grid amplification takes place. Although the power gain
is relatively low, linearity is excellent. The Collins 30S-1 is an
example of a cathode-driven Class AB1 amplifier.
The Class AB1 amplifier is like a 2-cycle, single-cylinder engine. The power stroke is roughly half of each crankshaft revolution--it has a flywheel (the tank circuit) that supplies power between power strokes--and it is more efficient than a gas turbine (Class A).
Class AB2 is similar to Class AB1 except that the grid is driven into the positive voltage region during a part of the anode conduction period. A Class AB2 amplifier can be grid-driven or cathode-driven.
When the grid is driven positive, it attracts and accelerates the
cathode's electrons. Some of the electrons stick to the grid,
resulting in grid-current. Electrons that miss the grid travel to the
anode. The accelerated head start causes a sharp increase in
instantaneous anode-current--and a sharp decrease in linearity. The
distortion products from a single-ended, Class AB2 grid-driven
amplifier are roughly 1/100 part [minus 20db]. In SSB
service, this level of distortion is virtually certain to cause
interference to other stations using adjacent frequencies. However,
by limiting grid-current and by adding an unbypassed, low-L cathode
feedback resistor (to develop an out-of-phase {negative} feedback
voltage) it is possible to achieve acceptable linearity in
grid-driven Class AB2 operation--but only if the grid current
produced is small.
An unbypassed cathode resistor is also useful for improving linearity
in Class AB1. For example, the 4CX250B has a somewhat objectionable
distortion level in Class AB1, SSB service. Adding a 25 Ohm resistor
between the unbypassed cathode and chassis ground improves linearity.
The trade-off is that slightly more grid drive voltage is needed to
achieve the same output level. Cathode negative-feedback is also
useful with TV sweep tubes--devices that were originally designed for
switching--the opposite of linear amplification. When an appropriate
cathode resistor is used with a sweep tube, reasonable linearity can
be achieved.
Even though Class AB2 cathode-driven/grounded-grid operation
produces grid-current, it is never the less fairly linear due to the
laundering effect of negative-feedback. This is the result of the
input and output signals being in series with each other and out of
phase. Due to the negative-feedback, the distortion level in Class
AB2 grounded-grid service is low--typically about 40db below PEP.
High-Mu triodes work well in Class AB2 grounded-grid operation.
Medium-Mu triodes can be used, but they have less power gain.
Tetrodes and pentodes usually work well in grounded-grid operation.
Since tetrodes and pentodes typically have a grid-to-screen
amplification factor of about 5, its easy to assume that they offer
an advantage over triodes in Class AB2 grounded-grid operation.
However, RF-grounding the grid and the screen stops grid-to-screen
amplification. Applying DC screen-voltage does NOT increase gain
because grid-to-screen amplification can not take place unless input
signal voltage is applied between the grid and the screen.
The maximum available power in Class AB2 is roughly double the
anode-dissipation rating.
Class B is defined as a conduction angle of 180 degrees. Class B RF amplifiers produce unacceptable distortion in SSB operation.
Class C is defined as an anode conduction angle of less than 180
degrees. In Class C, the amplifying device is deliberately not
operated linearly. Instead, it is operated as a switch in order to
reduce resistance loss. The anode conduction angle in Class C
operation is usually made as short as is possible. In effect, the
tank circuit makes the RF output sine wave--like a bell that is
struck at a constant rate by a hammer. This is similar to the
principal behind the spark transmitter.
The efficiency of a typical Class C amplifier is high. When compared
to a Class AB1 or Class AB2 amplifier operating at the same power
input, a Class C amplifier will deliver a received signal increase of
about 1db--in other words, 1/6 of 1 S-unit. However, significant
trade-offs are required to achieve that 1/6 of 1 S-unit. As is the
case with Class B operation, the distortion from Class C operation is
so high that SSB operation is precluded. Only CW, FM or FSK operation
is practical. The harmonic output level from a Class C amplifier is
substantial. Extra filtering is usually needed to control harmonic
radiation.
The maximum available power in Class C operation is roughly three to
four times the anode-dissipation rating of the electron tube.
Class D is used up to about 1.6MHz--mostly in AM broadcast service. In Class D operation, the amplifying device rapidly switches on and off at a fixed rate--like a switching power supply--except that the output voltage varies at an RF rate. The amplitude of the RF is controlled by varying the on period of the switch. Smoothing is accomplished by a complicated filter that converts some of the odd-harmonic energy from the rectangular waves back to fundamental frequency energy. Class D is highly efficient--but it is limited in frequency capability and frequency agility.
Arriving at an amplifier design that will give years of
surprise-free service involves many considerations. Merely copying
circuits from published amplifier designs or commercial amplifiers is
not necessarily the best approach. Doing so may result in copying
someone's mistakes. The best approach is to learn what you can about
each section of an amplifier, discuss it with others--then reach your
own conclusions.
Basic prerequisites for getting a handle on what's going on inside an
amplifier are an understanding of Ohm's Law, inductive and capacitive
reactance, impedance, resonance, how gridded electron tubes function,
and some knowledge of L and pi networks.
A useful book on amplifier design is Eimac®'s Care and
Feeding of Power Grid Tubes. I will minimize the discussion
of topics that are covered adequately in this book.
When the first RF power FETs were introduced, it was commonly
thought that FETs would eventually replace bipolar transistors and
gridded electron tubes in HF power amplifiers. Since RF power FETs
work better at 50V than they do at 12V, FETs have not replaced
bipolar transistors in 12V mobile applications. Another difficulty
with FETs is cost. A pair of FETs that can produce 1200W PEP at 29MHz
cost about six times more than an electron tube, or tubes, that can
do the same job. The FETs' input power requirements are 50V at 50A,
i.e., 2500W--so there's considerable heat to dispose of. Meeting the
cooling requirement is not nearly as easy as it is with tubes because
tubes operate quite happily at surface temperatures that destroy
silicon devices.
In low power applications at room-temperature, solid-state devices
can last 100 years. However, at the junction temperatures encountered
in high power applications, the P and N doped layers slowly diffuse
into each other--thereby steadily eroding the device's amplifying
ability. A relatively-large rigorous cooling system is needed to
achieve a reasonable operating life from high power solid-state RF
amplifying devices.
Another difficulty with solid-state high power RF amplifiers is their
power supply requirements. Tubes are quite tolerant of moderate
variations in their anode supply voltage. Transistors, however, are
fatally sensitive to over-voltage. It is much easier to build a
3000V, 0.8A unregulated supply for a tube than to build a regulated
50V, 50A power supply with over-voltage, over-current and
over-temperature protection circuitry for high-power solid-state
devices.
The bottom-line is that 1500W, HF, gridded electron tube amplifiers
are more efficient, more forgiving, easier to cool, more compact,
weigh less, are more tolerant of high SWR and are less costly than
1500W HF semiconductor amplifiers. For instance, a pair of
legal-limit FETs cost about $800 from Motorola®. The efficiency
is about 10% less that what one can achieve with gridded electron
tubes.
For at least the last three decades, the vast majority of amateur
radio amplifier designs have been Class AB2 cathode-driven--a.k.a.
'grounded-grid'. One reason for this is simplicity--or at least the
appearance of simplicity. Ground the grid(s), drive the cathode. Only
three supplies are needed--the T-R switching supply, the filament
supply and the anode supply. Neutralization is theoretically not
needed because the grounded grid(s) shield the output element, the
anode, from the input element, the cathode. This theory works
almost perfectly. Grounded-grid amplifiers are virtually
always stable at the operating frequency because the reactance of the
feedback C is too high at HF to allow regeneration. This is fortunate
because there is no way of neutralizing a single ended grounded-grid
amplifier. Another advantage is flexibility. Almost any tetrode,
pentode, or high-Mu triode from the junk box will work. Linearity is
usually good and the typical power gain--10db to 14db--is acceptable.
So far, so good. Now for the trade-offs.
What goes on inside a grounded-grid amplifier is not as simple as it
looks. The AC component of the anode-current and the grid-current,
i.e., the RF cathode current, passes entirely through the cathode
coupling capacitor and the tuned-input circuit--so the input circuit
is in series with (and out of phase with) the output circuit. The
components in the tuned circuit must be able to handle a substantial
amount of RF current. Manufacturers of tubes that are designed for
grounded-grid operation typically recommend using a tuned input
pi-network with a Q of 2 to 5. To maintain an acceptable SWR and Q
when the operating frequency changes appreciably, all three
reactances in the tuned input must change proportionally. However, if
Q is allowed to change, L can be left as is providing that C1 and C2
are retuned. [For more information on this problem, see the
section titled "Tuned Input Circuits."]
Even though HF grounded-grid amplifiers are stable at their operating
frequency, at VHF the grid looses its ability to shield the input
from the output. HF grounded-grid amplifiers have a
less-than-pristine reputation for VHF stability.
For wide frequency coverage, the Class AB1 grid-driven amplifier
requires a much simpler tuned input than a grounded-grid amplifier
requires. Typically, grid-driven amplifiers have more power gain than
grounded-grid amplifiers. One Class AB1 grid-driven amplifier has
about as much gain as two Class AB2 grounded-grid amplifiers in
series. The trade-off is two additional DC supplies--a grid bias
supply and a screen supply. Both of these supplies need to be
adjustable. HV power FETs make this task easy.
There are two types of cathodes--directly-heated and
indirectly-heated. In a directly-heated cathode, a ditungsten carbide
layer on the hot (c.1800 degrees K) tungsten, alloyed with about 1.5%
thorium--a.k.a. 'thoriated-tungsten', filament wire emits electrons.
In an indirectly-heated cathode, the filament (a.k.a. heater) heats a
metal cylinder that is coated with strontium oxide and barium oxide.
This coating is relatively frangible--but highly emissive.
Ditungsten carbide is commonly formed by heating tungsten in an
atmosphere of acetylene (C2H2) gas. Carbon
atoms in the gas break their electron bonds with hydrogen atoms and
bond with tungsten atoms to form ditungsten carbide on the surface of
the filament wire. Since it is atomically linked to the underlying
tungsten, the ditungsten carbide layer is very durable. During use,
the process reverses. Ditungsten carbide gradually looses carbon and
changes back to tungsten. Extra heat exponentially accelerates this
process. A cathode is worn out when the carbon is mostly used up.
After their cathodes grow tired of emitting electrons, large
external-anode amplifier tubes are commonly "recarburized" with
acetylene, vacuum-pumped and resealed. This restores full emission.
Although it is possible to recarburize a 3-500Z, doing so is not
economically feasible. The smallest tube that is currently being
recarburized is the 3CX1000A7.
Each type of cathode has advantages and disadvantages.
Indirectly-heated cylinder [8877] and planar
[3CX100A5] cathodes have much less inductance than a
directly-heated cathode made from wires. Thus, indirectly-heated
cathodes are more frequency-capable. Some indirectly-heated cathode
tubes can perform satisfactorily at 2500MHz. The 3CX100A5 is an
example.
Directly-heated/thoriated-tungsten cathodes are more resistant to
damage from electrons that bounce off the anode. It's possible to use
up to 22kV with the larger thoriated-tungsten cathode tubes.
Electrons that have been accelerated by such voltages move at very
high velocities. When they strike the anode, they produce X-rays.
A thoriated-tungsten cathode typically warms up in one second, while
few indirectly-heated cathodes can warm up safely in one minute--and
three to five minutes is not uncommon. For HF operation,
indirectly-heated cathode tubes have a much higher cost to watt ratio
than thoriated-tungsten cathode tubes. For VHF and especially for UHF
operation, indirectly-heated cathode tubes are often the only choice.
For super-power HF operation, thoriated-tungsten cathode tubes are
the only choice.
Cathodes deserve respect. Filament-voltage and filament inrush
current are the prime areas for concern.
For optimum life from a thoriated-tungsten cathode, the
filament-voltage should be just above the voltage where PEP output
begins to decrease. As a thoriated-tungsten cathode ages,
filament-voltage needs to be increased incrementally to restore full
PEP. By using this technique, commercial broadcasters typically
achieve an operating life of more than 20,000 hours from
thoriated-tungsten cathode tubes.
According to Eimac®'s Care and Feeding of Power Grid
Tubes, "every 3% rise in thoriated-tungsten cathode
filament-voltage results in a 50% decrease in life due to carbon
loss." Each additional 3% rise in filament-voltage decreases the life
by half. Thus, cathode life is proportional to [E1/E2]^23.4
where E1 is the lowest filament-voltage at which normal PEP output is
realized--and E2 is the increased filament-voltage. However, for
heater-type oxide cathodes, if the heater potential is allowed to
fall below the specified level, the emissive material may flake off
of the cathode, and cause a cathode-grid short. On the other hand,
excessive heater potential causes barium migration to the grid -
which results in primary grid emission.
It's simple to make the filament voltage adjustable when the
filament is powered by its own transformer. All that's needed is a
small rheostat in series with the primary. For dual voltage, dual
primary transformers, a dual ganged rheostat is required. However,
when the filament is powered by a winding on the HV transformer,
making the filament-voltage adjustable is more difficult since a dual
ganged, very low resistance, high current rheostat must be connected
to the low-voltage high-current secondary winding. The typical value
needed for a pair of 3-500Zs would be (2) adjustable 0.01 Ohm @
30A--most definitely not a common rheostat. A reasonable substitute
can be made from a double pole, c.10 position, 30A rotary switch and
short lengths of resistance wire or ribbon bridging the fixed
contacts.
An indirectly-heated cathode can be ruined by operating it below the
rated minimum filament-voltage. When operated above its maximum
filament-voltage rating, an indirectly-heated cathode boils off
emissive material (principally barium) onto the grid and other parts.
This results in decreased cathode life and undesirable grid-emission
when the grid warms up during transmit. This condition is indicated
when the output power steadily drops off in AØ (max. signal,
key-down, a.k.a. NØN) operation. The decrease in power
normally begins within two seconds.
For maximum cathode life in HF communications service, an
indirectly-heated cathode should be operated at the rated minimum
filament-voltage. This can be accomplished best with a regulated DC
supply.
In a typical amateur radio amplifier, the filament-voltage rises about 5% during receive due to decreased load on the electric-mains. Not only is a 5% increase in filament-voltage during receive useless, it uses up thoriated-tungsten cathode emission life about three times as fast as during transmit. However, if an appropriate resistor is placed in series with the filament transformer primary, and shorted out by a relay during transmit, the filament-voltage will not change appreciably between transmit and receive. An appropriate type of relay for shorting the resistor on transmit is a 'power' reed-relay.
Thoriated-tungsten filaments commonly consist of two vertical
intermeshing helices (coils) of tungsten wire that are suspended by
their ends. (see Sept. 1990 QST, p.15) The conductance of tungsten at
room temperature is about 8.33 times the conductance at the normal
operating temperature. Thus, the start-up current for a 15-ampere
filament can exceed 100 amperes. Needless to say, 100 amperes makes
for a dandy electromagnet.
In a high amplification triode such as the 3-500Z, the filament
helices clear the grid cage by a matter of thousandths of an inch. If
the position of the filament changes, a grid-to-filament short may
result. Therefore, it is prudent to limit filament inrush current in
order to minimize thermal and magnetic stress.
Since the grid-to-cathode clearance in an indirectly-heated cathode
is not affected by movement of the heater inside the rigid cathode
cylinder, indirectly-heated cathodes are not affected by inrush
current.
For many of its smaller thoriated-tungsten cathode amplifier
tubes--such as the 3-400Z and 3-500Z--Eimac® recommends that
filament inrush current be limited to no more than double the normal
current. This rating is easily exceeded unless a special
current-limiting filament transformer or a step-start circuit is
used.
END PART 1