The ContinentalsThe second most popular make of transmitters amongst the offshore stations was Continental Electronics. These were installed on both the original Radio Caroline ships, Frederica and Mi Amigo (the ones on the Mi Amigo were originally fitted by Radio Nord in 1960) as well as on the second Veronica ship, the MV Norderney and the Laissez Faire, home of Radio England and Britain Radio, et al. Whereas RCA, the most popular make used the "ampliphase" technique to develop high power broadcast signals, Continental used a technique known as "Screen Grid" modulation on transmitters upto 10 kilowatts. For higher powered sets this was originally coupled to an external high powered amplifier system, using a technique known as Doherty to produce a linear (low distortion) signal whilst remaining electrically efficient. In later high power models Continental combined the two systems and applied screen-grid modulation directly to a Doherty power amplifier to obtain a very efficient transmitter, with simple circuitry and a low tube count. Whereas the RCA BTA-50H transmitter (released 1961) required 32 valves (of nine different types) to produce 50,000 watts of power, the Continental 317C (released 1964) needed just nine (of four different types). This obviously gives rise to substantial savings in manufacturing, operation and maintenance.
A Suffix letter, such as A, B, or C indicates generation of design.
The 315B and 316B were released in the late 1950's, and used the then new technqiue of screen grid modulation of the output valve to produce 5,000 or 10,000 watts. A separate unit, the 317B was simply a linear amplifier using
Doherty circuits to raise the power of either of the smaller units to 50,000 watts. The driver transmitter remained independent (though with some control circuits interlocked) and could be put on air directly if required, at the press of a switch.
This was advantageous for day/night power changes and for maintenance purposes. The models were updated in the mid-60's, by the 315C/316C which, electrically, was fairly similar to the earlier sets but in a radically different cabinet layout,
and by the 317C which was a complete 50kw sender and thus dispensed with the need for the earlier driver transmitter. These transmitters were available as both medium wave and shortwave variants, with very little difference in circuitry, and many hundreds,
if not thousands have been built and shipped around the world since the 1960's. The basic design readily extends itself to super high power operation, and Continental have manufactured these upto the one megwatt level (1000 kilowatts). The shortwave variant
is still manufactured into the new millenium, and the design has been licenced to other manufacturers over the years.
The perenial problem with generating AM, is to efficiently superimpose (modulate) the audio signal onto the RF carrier. In very early sets this was done at a relatively low powered level, simply by connecting an audio amplifier valve in parallel with an RF amplifier valve, and feeding both stages with a common voltage supply through an inductor. As the applied audio signal caused the audio valve to take more current, it "starved" the RF amplifier of supply thus reducing its output and vice versa. Although very crude, this Heising modulation was the only practical method in the earliest days of AM broadcast in the 1920's. (an even earlier and cruder method was to place a carbon granule microphone in series with the antenna feed cable!). To generate more power than could be produced by these simple modulator stages they were followed by "linear" RF amplifiers as the amplitude of the the modulated RF signal had to be accurately amplified to provide listeners with a low distortion signal. Unfortuntely simple linear amplifiers, using Class A or Class B stages are very inefficient, typically 20-30%, meaning lots of wasted electricity and surplus heat. Thus for a 50KW linear amplifier, there could be over 150KW of heat which needed to be removed from the transmitter. By the 1930's more effective techniques were under devleopment, such as Class B high level plate modulation, still regarded today as the "classic" way to generate AM, whilst engineers such as Cheriex and Doherty looked for more efficient ways to produce AM without having to apply high level modulation to the RF power amplifier. If you haven't done so already, I recommend following the links in the left hand menu bar to the "Ampliphase Theory" (Cherix) and "Five and Ten" (high level plate modulation) pages on this site. Dohertys solution to the problem, which he patented in 1936 was to design a high efficiency linear amplifier which could be used to amplify the output of a low powered transmitter of any type. Although often referred to as Doherty Modulation, this is incorrect as it is not a modulation technique in its own right, it is just an amplification technique.
For a plate supply of say, 1000 volts, setting the screen grid of a hypothetical valve (a hypode?) at 500 volts and driving the control grid with RF will cause the plate to switch cleanly between 1000 volts and 500 volts. Although little power is lost in the switching part of the cycle, a fair amount of power will be lost in the valve when it is switched on, as it will still have 500 volts drop. If the screen grid voltage is now modulated, the output appearing on the anode will vary in sympathy for a continuous drive condition on the control grid. This isn't quite technically correct, but gives a good simplified analogy of the process. Note that in the above example, with 500 volts on the screen grid in the unmodulated "carrier" condition, the tube will dissipate power due to the 500 volts dropped from its anode to cathode and thus be much less efficient (about 40%) than a typical class C power amplifier (about 75%). However, as the modulation is increased (towards either peak positive or peak negative), the valve will tend to operate more in a Class C condition, being either fully on or fully off, thus giving rise the effect that as the average modulation increases the efficiency of the transmitter will also increase! This is the opposite of systems such as plate modulation, where high levels of modulation produce more losses within the components. This is why SGM is often known as efficiency modulation. As radio stations tend not to transmit silent carriers very often, then the overall efficiency of screen grid modulation is on-par with conventional modulation systems.
Fig 3: The Continental 317B Screen Grid Modulated Transmitter with Doherty Linear amplifier. (C.E. sales brochure)
The above diagram shows the 50,000 watt 317B, which in effect is a 5KW 315B, coupled to a Doherty linear amplifier using a pair of 6697 power triodes. Note the switching to enable the 5kw stage to be put directly on air, bypassing the 50kw stage. As the gain of tetrodes is relatively high, the RF drive requirements are minimal, and two stages of amplification after the oscillator are sufficient to drive both 4CX5000 tubes. Audio modulation is applied to the PA screen grids from three parallelled 4-65 tubes, which function as a simple "series regulator" type circuit. Note the simplicity of the audio circuit, there are no iron cored modulation transformers required with their attendent compromises on audio quality and expense. As only a hundred watts or so of power are required to modulate the screen grids, the audio stages require no real high power technology or complicated techniques. The system is also able to cope well with over-modulation and similar transients due to the lack of any high power audio systems. Carrier regulation is also good, as there is no audio amplifer consuming power from the power supply, and is further enhanced by a regulator section within the modulator. Also, the "efficiency modulation" analogy lends itself to improved carrier shift, as the current drawn by the PA stage is much more constant between carrier and peak positive than a conventional transmitter - the higher apparent current at carrier being lost as heat in the reduced efficiency of the PA. The mechanical design of the 315/316B ws such that it could be built within the enclosure space of the larger 317B thus giving the impression of a single transmitter.
In a simple view, there are two amplifier tubes, both operating at parts of the modaltion cycle as Class B linear amplifiers and at others times as Class C. At the unmodulated carrier condition, the "carrier" tube is normally biased to operate in Class C and provide approx 95% of the output power, and the peak tube to provide about 5%. As the input drive is increased, the carrier tube saturates, and contributes no more voltage swing across the load (but remains in class C), however, the peak tube is then driven further into conduction (Class B) and thus adds power to the output. As the input is driven below the carrier point, the peak tube is cut off and produces no output, and the carrier tube is gently lowered from Class C to Class B status thus producing less output voltage swing as the modulation heads towards peak negative. However, as the modulation of a radio station will always lie between the positve and negative extremes and for symmetric modulation will always average at carrier level, then the Doherty amplifier will always tend to operate in an efficient manner. IE, at carrier, the carrier tube is Class C, at positive peak, the carrier tube is saturated and the peak tube is in Class C, whilst at negative peak, the peak is tube is cut-off and the carrier is barely conducting. However, the above is an over-simplification of the process and the real trick of the Doherty amplifier is how the two tubes are driven and then combined into the common load.
The peak tube is connected "directly" to the load (via the plate tank circuit, output filters, etc) whereas the carrier tube is connected to the same load through a 90 degree (1/4 wave) network. When the carrier tube is operating alone, this combining network has little effect other than to provide an incidental phase shift. However, without the phase shift circuit, as the peak tube starts to conduct, the "apparent" impedance seen at the combining point will start to rise:- As the genuine load impedance has not changed, and the plate voltage swing is the same the power to the load remains the same, but now two devices are supplying power, thus each tube would be required to supply less power, hence it will "see" an apparent increase in load impedance. However, with the impedance inverting properties of a 1/4 network taken into account, as the peak tube starts to supply power, and the apparent impedance increases, this increase will be reflected through the 1/4 section as a lowering impedance on the carrier tube anode/plate. As the plate voltage swing is the same, and the load impedance is effectively decreasing, the effect is for the carrier tube to supply power to the load than at carrier. When this is added to the power supplied by the peak tube, the power to the actual load also increases. The circuit is so arranged that at 100% positive modulation equal power is supplied by both tubes. In order for the circuit to work correctly, one last feature needs to be added. As the power delivered by the two tubes at the combining point is 90 degrees out of phase, an additional 90 degree "phase advance" circuit must be incorporated between the grids of the two tubes, such that the outputs appear in phase. Finally, suitable bias must be applied to the control grids of the two tubes to ensure that they are driven to the right conditions at carrier level.
On the 317B as shown in the schematic above, the basic Doherty circuit as described was modified by a patented Continental innovation known as the "Weldon grounded grid" (Mr James O. Weldon, 1905-1993, was Continentals founder) in which the carrier tube was operated in grounded grid configuration, with the drive applied to the cathode. This holds further efficiency benefits in that the drive power is passed through to the load circuits and not wasted as heat dissipation in the grid structure as in a conventional grounded cathode amplifier stage. Also, neutralisation of a triode is not required when operating in grounded grid, thus making associated cost savings. On the 317B the Doherty tubes were 6697 triodes, as later used by RCA on the 50H. The screen grid modulated 5kw driver stage used two parallel 4CX5000's, again as later used by RCA on the 50H. The 10kw 316B used three parallel 4CX5000's. I am unsure as to the exact release date of the 315/316/317B, but a good guesstimate would be 1957, a year after RCA launched the first ampliphase the 50G. All rectifiers on the "B" series were solid state, either silicon or selenium.
Type 317C transmitters as shown above were first released by Continental in 1964/65 and quickly became accepted within the indutry as a workhorse transmitter. Many of these are still installed and working to this day around the world. The design was a convergence of the two types of earlier transmitter technology, with a modern tube compliment. A Doherty power amplifier was designed around a pair of the then new high powered RF tetrodes, 4CX350000, with screen grid modulation applied directly to this amplifier stage. An incredible small number of tubes were required to acomplish this, giving rise to substantial cost savings in manufacture and operation. Audio amplification and modulation were simple and similar to the earlier lower powered transmitters, using the same "series regulator" type of arrangement. The actualm modulator tubes were a pair of paralleled 3X3000's to ensure an adequate amount of power was available to the screen grids. The two PA tubes were both driven by the same RF drive (with the use of a a 90 degree phase advance network, as described earlier) and a common control grid bias of -600 volts. However, to bias the tubes to the correct condition for Doherty operations, the PA's were fed with different static bias to the screens. The peak tube ws normally operted with -200 volts on the screen to ensure it was not conducting at carrier, whilst the carrier tube was operated at +750volts on the screen to ensure it ws operating in Class C. The plates were connected to the combining point of the load through a the 90 degree network, as per the 317B. Note that both tubes were operated in conventional grounded cathode configuartion, the earlier "Weldon" grounded grid arrangement had been dispensed with. Application of modulation was made simultaneously to both screen grids. As the carrier tube was already operating in class C, increasing its screen grid higher with postive going modulation had little effect. However, increasing the screen grid of the peak tube from the static -200 volts towards the +2000 volts which fed the modulator would bring this tube into greater conduction thus delivering power to the output. Under negative modulation, the peak tube is already cut-off whilst the lowering screen grid voltage on the carrier tube reduces the power delivered from this tube. Using tetrodes with their attendent high gain on the output stage means that a very simple RF drive is required, and only a few hundred watts, easily supplied by a single 4-400 tubes was enough to drive the PA stages. Compare this with the 5kw required to drive the PA stage on the earlier 317B using triodes in the output stage. A small oscilloscope was built into the transmitter to aid with alignment and set-up the operation of the PA stages.
Three Continental 317's were used amongst the offshore radio fleet. Two were fitted onboard the Laissez Faire, home of Swinging Radio England and Britain Radio in 1966 (these two eventually went to Transworld Radio in Swazilan where they apparently still operate into the new Millenium), and one was installed on the Mi Amigo for Radio Caroline when she was repaired in early 1966 following the beaching. A 317C was a natural compliment to the two 316's already installed on the Mi Amigo as well as the two 316's installed on the Caroline North ship, Frederica. There are a number of stories that a 317C was also purchased for the North ship and was never fitted, but I have not been able to confirm this story one way or the other. The set which went onboard the Mi Amigo was not the one originally ordered by Caroline, however, as that was commandeered by the Uk government "Diplomatic Wireless Service" and used for broadcasting into Rhodesia. Ironically this very same transmitter was later used for jamming the broadcasts of RNI from the Mebo 2. See the "Other ships, Other Tx's" page via the links on the left menu bar for this fascinating tale. Other early 317C's went to XETRA in Tijuana, Mexico, WRKO in Boston and KOMA in Oklahoma. A major design flaw in the original high voltage plate transformer saw a number of these fail in service with spectacular and frightening results, but the unit was quickly re-designed by Continental and gave no further problems after that. Over the following 20+ years the design was updated several times with solid state low power stages, and these became the 317C2, and 317C3 making this family of transmitters one of the most long lived and enduring models ever built.
The design of the 317 was also emulated by other manufacturers, such as Marconi, and a very large number of their B6034 transmitters were installed by the BBC as part of the November 1978 WARC frequency changes. The Marconi design was almost a carbon copy with solid state drives, though they used a valved modulator stage for flashover protection of the semiconductor stages. The BBC's use of these transmitters (909, 693, 1215, 1089, 1053, Brookmans Park, Droitwich, etc.) involved tri-plexing three units to gain 150KW of power, and then diplexed these with other services into a common antenna. Prior to purchasing these Marconi sets, the BBC home-built a small number of 50kw Doherty sets in the early 70's, using an inhouse design. This unit, which appears to have no model number, used some very elaborate circuitry to avoid infringing any patents, and rather than using screen grid modulation, simply mixed RF drive and AF at the control grids of the two 4JC301J PA tetrodes. It ws not particularly efficient, consuming more power at 45% modulation than the 317 did for 100%, - the BBC declined to provide a figure for 100% output. It also suffered a number of problems with spurious emissions, which could sometimes be embarassing. Most of these were taken out of service in November 1978, though two were modified for the 200khz long wave services from Westerglen and Burghead, but even these were eventually replaced.
The 315/316 C was an updated 10KW version of the 315/316B, and went into production mid 1960's. One of these was installed on the second Radio Veronica ship, MV Norderney. The electrical configuartion was not unlike the earlier 316B, but benefitted from using the latest tubes. A single 4CX10000 formed the PA stage on the 5kw whilst on the 10kw two were used. Using a PA tube rated for 10,000 watts of dissipation on a 5kw amplifier would ensure a long, easy and stressfree life for the device, thus assuring minimum maintenance costs. An even later version of the 316, the 316F is generally regarded as being the finest sounding AM transmitter ever built, with exceptionally low distortion and noise levels and excellent phase characteristics. It was a 316F that was used as a bench mark when many years later,
Harris developed their modern day legendary DX50 digitally modulated solid state transmitter.