It seems like a bare-essentials natural: a Pierce "aperiodic" oscillator driving a non-neutralized power amplifier for a multiband Morse code transmitter that requires just one coil and minimum adjustment. And then one day you happen to unplug the oscillator tube only to find that it still works! This page explores that topology, the Boosted Pierce (Figure 1), from its Great Depression beginnings through its canonization as a ubiquitous "glowbug" design that just won't go away. Fundamental to understanding the circuit is that it's an aperiodic-anode tri-tet oscillator that uses separate triode and screen-grid (tetrode or pentode) tubes.

Figure 1—Boosted Pierce transmitter based on the version described by Don Mix, W1TS, in October 1968 QST. Some variations use more or less voltage, regulated or not, on the oscillator; some variations include a cathode bias resistor in the final amplifier and some don't; some key both stages and some key only the final amplifier; this version uses a voltage divider to supply final screen voltage rather than the series resistor used in many versions. All have the same aim: better performance than an oscillator-only transmitter without the fuss and bother of neutralization—but it turns out that neutralization is well worthwhile for multiple reasons. Test point T allows measurement of the final grid current, with every 0.1 V across the associated 100-ohm resistor equating to 1 mA of grid current in the 5763—a value worth measuring, because underdrive is likely at 7 MHz and higher frequencies in every Boosted Pierce design that does not overstress its crystal. RS1 and RS2 are proportioned to set the screen voltage to 204 V, key down, in my test version. (Schematic symbols [Tubepad] and original Mix "Novice Special" Boosted Pierce transmitter schematic by Gary Johanson, WD4NKA)

November 1937: Radio Describes a Transmitter Based on the Pierce Oscillator

In "Airline Transmitter Adaptable for Amateur Use" (Radio, November 1937, pages 32–38 and 81), G. E. Smith, W4AEO, described a commercial transmitter design based on a 42 pentode Pierce oscillator:

     This oscillator, with crystal between control grid and plate, is commonly called the Pierce Circuit. If a coil-condenser tank is substituted in place of the crystal, it will be seen at once that it becomes similar to an ultra-audion or Colpitts circuit.
     This circuit was selected after all types of crystal oscillators had been tested, complete sets of readings had been tabulated, and the advantages and disadvantages of each taken into consideration. Each type of oscillator was keyed by various methods and the shape of the keying was checked on the oscilloscope.
     The only disadvantage of the Pierce oscillator is low output as compared with the more common oscillators, pentode, tri-tet, 6L6, et. But a crystal oscillator should not be used for power output, as it has one primary function and that is to control the frequency of a transmitter. The r.f. crystal current should be kept low at all times to minimize frequency drift. Why make the crystal do all the work when, by the addition of another small inexpensive tube, the crystal can be run at low output and the burden shifted to the second tube? An ordinary receiving tube is much cheaper than a crystal. Through the use of the Pierce circuit no more tanks are required; in fact, one less tank is used in this circuit that is used in a tri-tet. The output from the two arrangements will be about the same. Another advantage of the Pierce is that the plate circuit is untuned and there are consequently no critical tuning adjustments which must be made for proper keying.
     This oscillator is by far the most active of any circuit that was tried. To prove this during the preliminary tests, some very poor grade of X-cut crystals that refused to oscillate in other circuits, including the tri-tet, all oscillated actively when put in the Pierce oscillator; it was even possible to key these heretofore inactive crystals at high speeds.

*   *   *
     Some notes on the operation of the Pierce circuit will be presented. C1 plays an important part in determining the grid excitation on the oscillator tube. For frequencies normally used by amateurs, a value in the vicinity of 150 μμfd. is correct. For lower frequencies, such as may be used in a test oscillator, 250 μμfd. should be used. The r.f. choke in the plate circuit of the oscillator should be about 3.5 mh. for amateur frequencies. For lower frequencies one from 10 to 60 mh. should be used. Of course, one of 10 mh. is all right for the higher frequencies, but is not necessary. In the unit should, a National R-100 is used. Another important item is C7. This must be at least .01 μfd. At first this does not look correct; in fact, on one of the first experimental models a conventional .0001 μfd. coupling condenser was used and the r.f. crystal current was found to be as high or higher than in a normal oscillator. The coupling condenser was then increased to .01 μfd and the crystal current was lowered to a normal value. The idea seems to be to couple tightly and thus to pull the current out of the crystal stage. It is also important not to use over 200 volts on the plate of the oscillator and not over 100 volts on the screen grid, regardless of the tube used.
     It will be mentioned that during the preliminary tests various types of tubes were tried—41, 42, 6L6, 6L6G, 807, W.E. 307A, etc. All gave good results, but the 42 type was selected as it was short enough physically to be mounted in an upright position in the 5-inch chassis to be used. The 6L6 could have been used, but at the time the unit was developed this was a comparatively new tube and there was some doubt about how it would stand up under r.f. use. The 6L6G could have been used, but the envelope was too large. The 42 also had been proved to have a long useful life in a similar circuit that had been used previously.
     Should this circuit be tried for amateur use, it is suggested that an octal socket be used which would allow the use of a 6F6G, 6V6G, or 6L6G without any circuit change. All have been tried with excellent results, especially the 6L6G.

January 1938: Jones's 6C5-807 160-Meter 'Phone

From Frank Jones, W6AJF, "A New, Simple 160-Meter Phone" (Radio, January 1938, pages 130–134):

     A modification of one of the first crystal oscillator circuits provides a single-dial control 160-meter or 75-meter 'phone. The untuned crystal oscillator provides enough power to drive a screen grid tube such as an 807, RK25 or 802. An 807 tetrode is used with a tuned circuit, which has the only tuning control on the set....
     The crystal oscillator uses a 6C5 triode in a modified Pierce circuit in which the quartz crystal is the oscillator tuned circuit. The crystal is connected between the grid and plate in place of the tuned circuit of a Colpitts or ultra-audion oscillator. A .0001 μfd. mica condenser from grid to cathode provides the correct amount of feedback in the oscillator circuit. (It depends upon the total effective grid to cathode and plate to cathode capacity ratio.) The plate circuit is resistance coupled to the 807 grid. The plate resistor was split and bypassed as shown in the circuit diagram in order to avoid any undesired r.f. feedback. [The dc resistance in the 6C5 plate circuit is 30 kilohms, 10 kilohms of which is bypassed to ground. The 807 grid leak, which is surprisingly not seriesed with an RF choke and so therefore acts as part of the 6C5's plate load, is 100 kilohms. The coupling/DC-blocking capacitor between the 6C5 plate and the 807 grid is 100 pF.—W9VES] The actual plate voltage on the 6C5 tube is about 250 volts and the output is sufficient to light a small neon bulb or to drive the 807 grid circuit hard enough for class C operation. This simple oscillator circuit is sure-fire with 160- and 80-meter crystals of any kind.

This circuit is the first publication of the Boosted Pierce transmitter—that is, a Pierce oscillator that drives a non-neutralized amplifier feeding an antenna. So Frank Jones can be considered to be the inventor of the Boosted Pierce.

For the first few years after the 807 was introduced, its manufacturer (RCA) recommended operating the tube at no more than 400 to 500 V (officially upped to 750 V later) and that cathode bias from a resistance on the order of 400 ohms be used in addition to any additional bias provided by a grid-leak resistor and/or negative DC supply. As a result, the power-amplifier design in this very first Boosted Pierce implementation in effect includes Class A bias—not to protect against the underdrive later to become endemic in Boosted Pierce designs operating at higher frequencies, but rather in accordance with RCA's recommendations for proper application of the 807. At the higher operating frequencies covered by later implementations of the Boosted Pierce, shunt capacitance in the oscillator-plate-to-final-grid circuit, and increased feedback/feedthrough caused by the final's unneutralized grid-plate capacitance, would reduce the driving power at the final grid to less than that necessary for true Class C operation. Many of these implementations nonetheless omit final-amplifier cathode biasing, resulting, in many cases, in underbias that causes excessive plate dissipation in the final—a condition commonly overlooked because measuring output power was relatively problematic, and because the FCC regulations of the day required that transmitter power be evaluated in terms of power input.

September 1938: The Norfolk Radio Amateurs Prepare for Emergency with the Boosted Pierce

December 1938: Goodman's 6C5-6L6 Transmitter

1938: The Fifth Edition of the Radio Handbook Names the Boosted Pierce

Four of the seven "Multi-Tube Exciters" described in the "Exciter Construction" chapter of the October 1938 Radio Handbook use Pierce oscillators: three based on Frank Jones's harmonic Pierce "regenerative" oscillator—which, as in its original description in Radio, he studiously avoids describing as a Pierce circuit—and two based on the classic triode Pierce. The first circuit in the section, "Two-Tube, Two-Band Exciter with Single Tank Circuit," coins the name I use for the circuit in the terse caption for its schematic (Figure 2): "Wiring diagram of the 'Boosted Pierce.'" (note the quotation marks; that was the Radio crew's name for it, too.) Its principal attraction and its principal oddity are described in the same sentence: "No neutralization is needed for working on the same frequency as the crystal; the 6L6G sort of 'takes over' and acts somewhat as a regular 6L6G crystal oscillator."

Figure 2—A write-up in the fifth edition of the Radio Handbook (1938) gave the Boosted Pierce its name. Coil data was supplied for 160 through 10 meters; "the output will vary from 3 to 15 watts, depending on the frequency of the crystal and supply voltages....With 10-, 20- and some 40-meter crystals, the coupling must not be too tight, loading the 6L6G too heavily, or the crystal may act sluggish or refuse to oscillate." Considering that most variations on the Boosted Pierce do not drive their power amplifiers hard enough to develop class C bias, the inclusion of a cathode-bias resistor for the 6L6G is a good thing.

1941: Bacon's Portable-Emergency Transmitter

A test version consisting of a 12V6GT tetrode-connection Pierce oscillator (213 V plate and 105 V screen, both regulated, with 470 ohms of cathode resistance and 51 pF of additional grid-to-ground feedback capacitance) driving a shielded 6417 (12-V 5763) beam tube (360 V plate, key down; 267 V screen [from 18-kilohm series dropping resistor], key down, 22 kilohm grid leak with no series choke, 340 ohms of cathode resistance) produced 7.5 to 7.7 W at 80 meters, 5.3 to 6.0 W at 40 meters doubling from an 80-meter crystal, and 4.7 W at 40 meters using a 7.01-MHz crystal. The 6417 grid current was 3.2 mA with my 3.57-MHz crystal and 2.51 mA with my 7.01-MHz crystal, indicating sufficient drive for Class C. With the 3.57-MHz crystal, short-circuiting the final's cathode resistance increased grid current to 3.58 mA and increased output to 8.18 W. Simultaneous cathode keying of the oscillator and amplifier was used.

Unlike many other Boosted Pierce implementations, this one drives its final sufficiently hard for Class C operation. In every configuration, however—straight through on 80, doubling to 40, straight through on 40—there was moderate, long-term yoop, indicating considerable crystal heating that must be considered dangerous. (And these were pre-FT-243 crystals—mainly FT-171B units intended for use in the Hallicrafters HT-4/BC-610 transmitter.) Bacon's keying method of letting the oscillator run while making and breaking the final screen supply would largely mask this effect, but the thermal-cycling stress on the crystal would still occur, if less often. Because of this, and especially considering that modern crystals can tolerate even less drive that the relatively large crystals of the 1930s and 1940s, I can recommend the Bacon Boosted Pierce only in its lower-power, 1Q5GT-based form—pending investigation of the crystal stress in that configuration.

1946: The QST Longfeller

1947: The Eleventh Edition Radio Handbook's Longfeller "Me-Too"

Like its predecessor in 1946 QST, the Radio Handbook's postwar Boosted Pierce (Figure 3) promoted low-cost wood construction.

Figure 3—Boosted Pierce transmitter schematic from pages 287 and 288 of the 1947 (11th edition) Radio Handbook. This design, the power amplifier of which lacks protective cathode bias but includes a 47-ohm resistor between its screen and screen bypass capacitor to stop negative-resistance-caused parasitic oscillations, is intended to produce 3.5- and 7-MHz output with only 3.5-MHz crystals—that is, to work 40 meters by doubling from 80. Only the power amplifier is keyed. "The transmitter will deliver between 10 and 15 watts to the antenna on the 80-meter band and 5 to 8 watts on the 40-meter band with the same crystal and coil being used on both bands." Notice that half of the final coil's 20 turns are shorted out for 40, which seems logical for doubling frequency—until we recall that because a coil's turns-v-inductance ratio is nonlinear, halving the turns of a coil does not halve its inductance.

1950: QST's 6AG7-6*6 Beginner Transmitter

195x: The Aperiodic Oscillator Goes to Market as the Heath AT-1

1966: McCoy's Mighty Midget

Described in pages 54–57 of February 1966 QST by Lewis G. McCoy, W1ICP, the Mighty Midget transmitter uses a 6GW8 triode-pentode in an oscillator-amplifier design that, seemingly in attempting to overcome the insufficient drive problem endemic to the Boosted Pierce, combines a curious misapplication of oscillator fundamentals with insufficient attention paid to issues of amplifier stability. The result is a transmitter that should not be duplicated and used on the air without considerable circuit modification.

The Mighty Midget's troubles begin with its oscillator, a common-anode triode Colpitts circuit that nonetheless operates with its anode held above RF ground by the imprecisely resonant circuit formed by one RF choke (40 meters) or that RF choke plus another choke in series (80 meters) paralleled by the capacitance to ground of the triode anode and pentode grid and their interconnecting wiring. The use of the Colpitts arrangement, the building-in of oscillator-anode parallel resonance, and the use of an RF choke in series with the power amplifier grid-leak resistor—a usage more or less obsolete as of the mid-1950s—suggest that in designing the Mighty Midget, McCoy had worked to avoid the power amplifier underdrive of the Boosted Pierce: Eleven years earlier, beginning on page 36 of October 1955 QST, he had modified an aperiodic-oscillator-based commercial design to overcome exactly that problem by adding oscillator-anode tuning and power-amplifier neutralization in "More Power with the AT-1."

The presence of oscillator anode (amplifier grid) resonance without the addition of amplifier neutralization results in a serious flaw: With its anode and grid tuned to roughly the same frequency by design, the transmitter's high-gain pentode power amplifier is unstable to the point of acting more like a locked oscillator than just an amplifier. Suspecting that this might be true with my version of the Mighty Midget after I encountered hysteresis in its amplifier plate tuning at 40 meters—I couldn't achieve the same output power peak, and maximum output occurred at different settings, when tuning through resonance from above and below resonance—I disconnected the cold end of the oscillator cathode choke from the KEY jack and discovered that the Mighty Midget put out more RF power with its crystal oscillator disabled than with its crystal oscillator connected and operating! Adjusting the amplifier plate tuning across a wide range below, through, and above resonance and back again, I heard a series of grunts and squawks in my receiver as the final amplifier found and rejected varying modes of oscillation, some purer than others. Holding my key down at the circuit's highest power output setting, I discovered a relatively pure but highly frequency-unstable signal at about 7065 kHz. Moving a hand near the oscillator-anode chokes swung the frequency wildly. Keyed on and off, the amplifier-cum-oscillator chirped like the proverbial bird. This is a transmitter suitable for neither beginner nor Novice.

At least two options are available for taming the Mighty Midget transmitter. One would be to abandon its triode oscillator, rewiring the oscillator's Colpitts capacitive voltage divider to the grid and cathode of the pentode and moving the triode's cathode RF choke to the cathode of the pentode. This turns the transmitter into an electron-coupled, "grid-plate" pentode oscillator-only design. Keying will be excellent at 80 meters (4.1 W output measured in my test version) and acceptable (2.8 W output) at 40 meters. The other option is to rewire the 6GW8 triode as a Pierce oscillator, removing the low-value RF chokes used for oscillator-plate resonance in the original design and moving the oscillator cathode RF choke to the oscillator plate circuit. With this arrangement, output with 350 V on the amplifier plate will be nearly 5 W at 80 meters (excellent keying) and 4 W at 40 meters (acceptable keying [that is, keying with some yoop, which I attribute to crystal heating]).

1968: Mix's 6C4-5763 "Beginner & Novice" Transmitter

In March 2008, peripheral to new experiments with the Caringella 5763 transmitter, I briefly tested the Mix 6C4-5763 Boosted Pierce circuit at 7 MHz under these conditions:

6C4 (oscillator) plate voltage: 180, regulated
5763 (amplifier) plate voltage: (key down): 360
5763 plate current:
5763 screen voltage (key down): 204
5763 grid current: 1.6 mA
5763 plate pi net inductance:
Output power:
Efficiency of 5763 stage:

The arrangement of the test circuit allowed me to key only the 5763 or the 6C4 and 5763 simultaneously. This was useful because the crystal environment created by the oscillating 6C4 stage resulted in considerable yoop (frequency shift across multiple Morse Code elements [dots, dashes]) at oscillator turn-on. Keying both stages simultaneously would cause this frequency shift to restart many times during a transmission, whereas turning on the 6C4 and keying only the 5763 amplifier stage would cause the yoop to happen only at the beginning of a transmission, allowing me to render it practically inaudible to other stations if I waited to start sending until after the oscillator had stabilized.

In practice, keying both stages resulted in considerable yoop that differed somewhat from the turning-on-only-the-oscillator yoop, likely because of feedback and load shift contributed by the 5763. Turning on the oscillator stage at the beginning of each transmission and keying only the amplifier stage gave yoopless keying, but with two artifacts: (a) because the amplifier stage was not neutralized, a neighborhood-audible backwave was present that I knew, from listening long ago to a mile-away friend's taped reception of my 1970s version, would be audible whenever the key-down signal was sufficiently strong, and (b) an FSK effect, audible if the backwave was audible, caused by shifting oscillator loading as the amplifier stage was keyed.

Although the on-air effect of oscillator-startup yoop can be minimized by turning the oscillator on and off only once per transmission, I consider yoop to be an indication of crystal heating that is potentially damaging to the crystal whenever, and however often, it occurs. But even with the oscillator stage producing 1.6 mA of grid drive, the unneutralized 5763 is not yet overdriven—increasing oscillator plate voltage above 180 increased the amplifier output, indicating that overdrive had not yet been reached—so reducing oscillator plate voltage below 180 would also reduce the transmitter's output power and efficiency.

Bottom line: At least at 7 MHz and higher frequencies—I have not yet tried the circuit at 3.5 or 1.8 MHz—the 6C4-5763 Boosted Pierce is a poor transmitter design that endangers its crystal and sacrifices signal quality for the purpose of avoiding the easily implementable best practices of (a) using a tetrode or pentode oscillator circuit that operates as a triode-tetrode, thereby isolating the crystal from feedback and load shift attributable to the amplifier, and (b) neutralizing the amplifier, making the stage easier to drive and practically eliminating backwave by canceling key-up leakage.

Glowbug Variations: The "One-Tube" Triode-Pentode Boosted Pierce

2003: Johnston Builds the Mix 6C4-5763 Boosted Pierce

[to be continued...]