QRP technical TX notes

c23. QRP experimental techniques II,
"transmitters"

Low cost solidstate power amplifiers
When considering the design of solidstate power amplifiers with all their problems of accurately-matched loads, susceptibility to transient overloads, parasititcs and other forms of destructive self-oscillation, not to mention the high cost of devices capable of high-power operation, most amateurs tend to play safe and follow closely the techniques closely the techniques currently favoured by professional designers. This is understandable but is not necessarily the best policy for those seeking ways out of using the older rf power devices that can now be acquired for a fraction of the cost of "state-of-the-art" devices. With some shopping around it is possible to build a power amplifier of 10W rating at better W/£ ratio than by using newly-bought valves and high-voltage psu (though I stick to my formerly expressed beliefs that the valve approach is still a good deal easier at and above this power level).
Almost 20 years ago, TT, and subsequent editions of ART, drew attention to what was then described as an "improved transistor amplifier circuit" in the following terms:
Most transistor power amplifiers use the common-emitter arrangement, corresponding most closely the the standard valve (grounded cathode) circuit, on account of the high gain; others use the common-base configuration on account of its higher maximum frequency for a given transistor; or the common-collector circuit wich allows direct metallic contact between the transistor case and the heatsink or chassis.
"The distribution of the desirable characteristics among the three standard circuits has encouraged some designers to try and produce hybrid arrangements combining as many possible good points into a single circuit".

Fig.1. A hybrid form of solidstate power amplifier using grounded collector. This was described for use with early pnp power transistors in 1964.

One such approach is that described by Irving Gottlieb in Electronic Industries August 1964: Fig.1. This circuit, is claimed, has been developed using 2N1908 pnp device, and the author states that he obtained about 10W output power power on 7MHz band using a transmitter consisting of a 2N697 co, 2N1907 ba and 2N1908 pa. The efficiency and thermal electrical characteristics of this circuit are claimed to be outstanding. The author admits that some experimentation will usually be needed to achieve best results but states that, once the input and output impedances have been optimized, stability is almost unaffected by driver or load variations, or by changes in the supply voltages.
It was also claimed that such amplifier could be amplitude modulated alomst 100% by means of a transformer winding inserted in the positive dc lead, provided that the supply voltage was suitably reduced. However, I have not heard of anybody using this type of amplifier for many years. Indeed little attention has been given to either this 'hybrid' configuration or the basic common-collector arrangement during the pasta decade. G3VA, Technical topics, Radcom January 83 pp41.

Grounded-collector 14MHz linear.
In view of the above notes, it was all the more interesting to receive from LA8AK a description of a 14Mc/s solidstate linear power amplifier using the 'grounded collector' configuration. This amplifier was assembled in order to test what could be achieved using low-cost BD139 transistors which a sometimes available at low cost, an alternative is BD135. LA8AK recalls that while the grounded-collector arrangement was occasionally used by amateurs 15-20 years ago, it is seldom found mentioned. To the best of his recollection, he has never seen it proposed for linear operation. In practical terms the amplifier is not very different from conventional grounded-emitter configuration, but offers useful advantages.
The LA8AK test amplifier is shown in Fig.2, and incorporates a number of ideas suggested originally by LA7MI Stein Torp, LA8AK writes:
"Instead of an rfc in the supply lead, it uses a low-Q resonant circuit to provide greater stability; LA7MI burnt out several BLY89 devices on hf before adapting this arrangement. For 3.5MHz, C2 would be 2200pF.
"L2-C2 plus stray-capacitance is resonated by means of a GDO with the supply voltage connected with base-to-emitters short-circuited and L1 removed. TR5 is diode-connected and provides further protection of the power devices. L1 consists of 7t, 15mm inner diameter. The input transformer, T1, is three times 8t of enamelled copper wire, trifilar wound on unknown ferrite core with 15mm OD.
"In the test set-up the transistors were just below cut-off, and some further work on this might prove rewarding. Gain was measured as roughly 10dB, linear rf output with four BD139 devices is10W, saturation is 12-15W rf. Efficiency is about 50%.
"The four transistors are mounted in a 'cross', with bases connected together (and collectors to ground). The emitter leads should be of equal length to ensure equal power dissipation; no emitter ballast resistors were used in the prototype but could be 1 ohm, 0.25W. It is important to achieve equal power sharing either with equal length leads or emitter resistors. The amplifier is mounted on a brass plated screwed to a cooling fin, and in appearance looks more like a 144MHz amplifier than an HF design. The -24V dc supply was chosen because many readily-available surplus power-supply units can provide this.

G3VA, Technical topics, Radcom January 83 pp41-42

Fig2. LA8AK's experimental 14MHz grounded-collector linear amplifier using four low-cost BD135 (BD139) npn devices and providing about 10W rf output.

Back

Last update 2004.12.26