Combination class H modulator - power supply by Bob, K1KBW.
This implementation uses 10 IRFP260N MOSFETs, 5 in the each leg of the modulator.
Modulators and Audio
In theory, any type of high level modulator can be used with a class
E RF amplifier to form a complete AM transmitter. The modulator can be
transformer, series or shunt coupled. However, the best type of
modulator to use is some type of series modulator. Such a modulator can
use any technology - class A analog, high efficiency analog, or digital
(Pulse Width), as long as the result is modulated DC applied to the RF
amplifier. Two types of modulators are described here: A high
efficiency analog modulator using class H and a Pulse Width modulator (the
most efficient of the modulators)
Modulator types and advantages of each
Modulators fall into two general classes - Analog and Digital (pulse width).
Both have advantages.
The analog modulators will operate over a wider range
of load currents and resistances, and are generally fairly straight-forward in design.
The audio quality is generally superb - smooth frequency response and low distortion.
Because of these characteristics, along with the
relative simplicity of operation, these types of modulators have been
extensively used by amateurs building class E transmitters.
The primary disadvantages of analog modulators are the size and weight and heat production.
Analog modulators are much larger than comparable digital (pulse width) modulators, and
have significantly larger power supply requirements.
Digital (pulse width) modulators are the most efficient (up to 95% efficiency is possible),
and can be made smaller, and require smaller and simpler power
supplies than comparable analog modulators. Pulse width modulators are designed
to work into a particular load resistance - the load in this case being the
class E RF amplifier. The class E amplifier must be operated within
certain parameters to properly match the design impedance of the modulator.
Digital modulators are designed with significant
filtering, by the nature of how they work, and therefore can offer better
control over the transmitted bandwidth than analog modulators. A properly
designed and implemented pulse width modulator can produce broadcast quality
audio, and the system is widely used in modern AM broadcast transmitters.
A Class H modulator is an analog, series modulator. This
design is much more efficient than a standard class A series
modulator, which would typically be around 30 or 35 percent efficient.
A system based loosely on this technology was used
in the Harris MW1 Solid State 1KW Broadcast Transmitter. This
technology is applicable to vacuum tube designs as well as other solid
The idea behind class H, and a related class, class G, is to run the
series modulator devices in the audio output at or near saturation.
The voltage supplied to the near saturation devices is increased when greater
output voltage is required. Otherwise, the when a lower output voltage is
needed, the saturated device behaves like any other series modulator.
In class G, the
supply voltage supplied to the almost saturated output amplifier is stepped,
and the number of power supply steps depends
on the particular design. In class H, the supply voltage is adjusted
linearly, rather than in a step function.
How Class H Modulators Work
A class H modulator starts with a source-follower series modulator (Q2)
operated very close to the saturation point most of the time, and a 2nd device (Q1) that
supplies additional voltage to Q2 when needed.
The collector voltage of Q2 is fed through a diode from the
50V carrier power supply. The supply voltage is usually somewhat
higher than the desired output voltage at carrier. Transistor Q1 is
operated at cutoff when no modulation is present (carrier only).
Since Q2 is an emitter-follower, the voltage appearing at the emitter
of the transistor follows base voltage. The base voltage is
set such that the output of Q2 is around 40 volts - about
10 volts less than the power supply voltage of 50V.
During the negative portion of the modulating waveform, the voltage
fed to the base of Q2 is driven lower, and Q2 acts as a normal emitter follower
series modulator. Voltage is fed to Q2 from the carrier power supply through
Q1 is not conducting during the negative portion of the waveform,
and is effectively out of the
circuit at that time.
During the positive peak modulation cycle, when the modulator is required
to deliver more voltage, the gate of Q1 and the base of
Q2 are both driven higher. Since the emitter voltage of Q2 will follow the base voltage,
the voltage drop across Q2 will decrease, and the Q2 emitter voltage, and the modulator
output voltage will increase. At the same time, the voltage fed to the
gate of Q1 is also increasing, and Q1 will begin to conduct.
Q1 will begin to supply more voltage to Q2 before Q2 saturates and
causes distortion. The output voltage will continue to rise, up to the positive peak power
supply voltage of 100 volts, as long as the base of Q2 and the gate of Q1
continue to be driven higher.
At all times, Q2 is the primary modulating device. Q1 simply
supplies additional drain voltage to Q2 as the output voltage
increases. It is important to note that during positive peaks, diode D1
connected between Q2 and the carrier power supply
is back biased. No current flows from the carrier power supply
at this time, and the supply is effectively switched out of the
A complete modulator using class H is featured elsewhere, in the Construction
Projects section of this document.
Pulse Width Modulators
A pulse width modulator is essentially a switching power supply, where the
output voltage of the supply can be controlled by an external input. In this
case, we feed audio into that input, and control the output of the switching
supply with the audio we're supplying. Audio amplifiers using pulse width
modulation are becoming quite common, particularly for high power amplifiers.
Brief discussion on how Pulse Width Modulators operate
Pulse width modulators operate by varying the duty cycle - the "on time"
as compared to the "off time " of a switching (square wave) waveform.
This switching waveform is produced using relatively simple low level circuits,
and is amplified using switching (either saturated or cutoff) amplifier stages,
to the desired output voltage.
The output of this amplifier is then filtered, removing the switching frequency.
After filtering, the output is the average voltage of the switching waveform.
By controlling the switching waveform on-time as compared to the off time, we
can control the output voltage, after filtering, of the amplifier. Because
the amplifier stages in a pulse width modulator are operated either at cutoff
or saturation (this is called switch mode, or class D), such modulators are
typically very efficient. 95% efficiency is achievable with practical
The advent of some very good Pulse Width Modulator ICs over the past few years has significantly
simplified the design and construction of these types of modulators. Excellent
results can be achieved with comparatively few components. The filter network
components are fairly small, at least up to 400 or 500 watts, and it is possible
to build a 400 watt pulse width modulator that you can hold on one hand.
More Information about Pulse Width Modulators
A complete explanation of Pulse Width Modulators, and how the PWM
Modulator works can be found in this Solid State PWM paper.
There is also a similar writeup on Vacuum Tube PWM.
Class E RF amplifiers and their respective modulators require a variety of
DC voltages. Some of these voltages need to be switched on and off, and in
a particular sequence.
Providing power to low level circuits
Low voltages such as +12, -12, +5, etc. are easily supplied using standard
7812, 7912, and 7805 linear regulators connected to a transformer/rectifier/filter combination.
It is usually advisable to separate the power supplies for various functional
units. For instance, the modulator should have its own low voltage power source,
independent of any other functional units such as RF amplifiers. Low power
supplies should be fused with a small fuse, appropriate for the particular
transformer in use.
High Voltage, High Current Supply
The power supply for the final RF and modulator stage will need to supply
hundreds of watts of steady-state power, and perhaps thousands of watts of
peak power under heavy modulation. Furthermore, this power supply will most
likely be switched on (keyed) during transmit, and switched off when
in receive mode.
Power Supply Filter
Due to the relatively high current (low impedance) of most class E
transmitters, the power supplies are generally capacitor input filters, and
most often this single capacitor is the entire filter. Unless the filter
capacitor is excessively large, there will be a small amount of ripple in the
output. Generally, the modulator will reduce or eliminate this ripple from the
RF output if it is designed correctly.
The power transformer must have sufficient capabilities to supply the required
power for the duty cycle of the transmitter. If the transmitter is operated
for long periods of time (do you make "old buzzard" transmissions?),
the transformer must be specified for almost continuous duty. If you make
relatively short transmissions, a ligher transformer may be used.
rating of the transformer and the current rating will determine how much power a transformer can
supply. If you have a transformer that can supply 90VAC at 5.5 amperes, you
have about a 500 VA transformer. With a capacitor input filter, under load,
you can expect to get around 115 VDC from the supply. Even though the transfomer
can supply 5.5 amperes, you CANNOT draw 5.5 amperes continuously from the supply at 115VDC
because you will be exceeding the power rating of the transformer. Furthermore,
the volt ampere rating of the transformer should be reduced somewhat when
using a capacitor input filter. The supply in this example should be able
to safely provide between 3.5 and 4 amperes of current at 115VDC.
Switching the High Voltage Supply
A discharged filter capacitor in a high voltage supply is a virtual short
circuit across the supply when it is first energized. There needs to be some
amount of resistance in the system, or rectifier failure will eventually result
from the current surge. If the power transformer is a very high current
transformer and adds very little resistance to the system, some type of step
start should be employed. A step start puts a low value resistor in series
with the power supply AC input for a short period of time, as the filter bank
is charging. The resistor is then shorted out by a relay.
The AC input to the supply, as well as the DC output from the supply should both
be switched. If the AC input is allowed to remain after the load is removed
from the supply, the supply voltage will soar.
If the DC output is not switched, you run the risk of a
catastrophic failure in your transmitter, as the entire stored energy
of the power supply will be discharged into the lowest resistance path
of the system. You will also be prevented from using an effective
overload system, as these systems rely on the ability of the DC to be
quickly removed from the transmitter.
Be sure to include a light
bleeder resistor to slowly discharge the filter capacitors when the supply
is in standby or turned off. It is highly dangerous to allow stored voltage
to remain for a long time after a supply is turned off. The switching system
should be so designed so as to not enable the high voltage supply unless the
low level circuitry is activated and functioning.
The Driver should be switched in a manner similar to the RF
amplifier. The driver power *must* come from a separate power
supply, used only for the driver. If the driver power is derived from
the final RF amplifier power supply, you will not be able to
effectively test or tune up the transmitter.
The output voltage of the supply should be metered at all times, and it is also
a good idea to meter the output current. A proper line fuse must be provided for the
high voltage supply. Remember, the power supplies used in most class E transmitters
produce dangerous voltages!!! Please read the Safety Notice. You cannot be too
Power Supply Wiring, Rectifiers and Relays
The rectifier should be rated at many times the expected steady state supply
output. If you are building a 20 ampere supply, your rectifier should be at
least a 50 ampere (and perhaps higher) bridge rectifier. The peak rectifier
current is many times the average.
With certain designs and under heavy modulation, the peak current required
from the supply is many,
many times the average output - in some cases 8 or 10 times as much. As an
example, you are using a pulse width modulator modulating a 320 watt class E
RF amplifier: 40 volts at 8 amperes. The power supply output voltage is
120 volts DC (as an example). The modulator is 95% efficient, so the steady-state current
from the supply is around 2.8 amperes. If the transmitter were suddenly
modulated to its full positive peak value, the full supply voltage will, in
essence, be supplied to the RF amplifier. The resistance of the RF amplifier
is 5 ohms (40 volts divided 8 amperes), so the power supply would be required
to deliver 24 amperes (120 volts divided by 5 ohms) for the modulation peak.
You should use at least number 12, and preferably number 10 wiring between
the supply and the load, and the interconnecting wires should not be too long,
as excessive voltage drop will result.
Any relays must be able to carry the full, peak current of the supply. Use
large, good contactors with heavy terminals and wide contacts. Relays used for DC
must be heavier than their AC counterparts, as DC is more difficult to swith,
and contact welding and arcing can be a problem. Motor control relays with strong springs work
well in this application.