*** The Ampliphase Ancestry ***
Latest Update August 2002
 | |
The first RCA "ampliphase" transmitter, the BTA-50G was released onto the market in 1955, but the outphasing principle was already 20 years old at that time. Outphasing was first described by Henry Chireix in his paper "High Power Outphasing Modulation" in the IRE proceedings, November 1935, no. 11, Volume 23. This was just nine months before W. H. Doherty patented his amplifier design.
Information on early outphased transmitters is now very scarce, however, by 1948 station KFBK in Sacramento, California had successfully home-built a fifty kilowatt version. This station was operated by McClatchy Broadcasting who also operated KOH in Reno, Nevada, and a five kilowatt set was subsequently built for that site. The May 1951 issue of "Electronics" magazine carries a three page article,
describing in great detail the design and operation of the five kilowatt transmitter, including a complete circuit diagram. This design used a single modulator stage, followed by a tripler (as used by RCA on the shortwave and five and ten ampliphases, fifteen years later), to provide the required phase shift at the carrier frequency. However, it appears the original KFBK system used three cascaded stages, as per the 50G and 50H.
The PA stages at KOH were a pair of 3CX2500F3's, driven from a pair of 4-250A's. To provide for linearity correction and drive regulation, the RF drive to the PA's was progressively reduced
as the phase moved towards 180 degrees, by means of a simple resistor connected between the PA grids. With 180 degree difference, a reasonable amount of drive power was lost in the resistor due to the current flow, but as the difference reduced, there was less current flow and therefore less power lost. At 0 degrees difference there would be no current flow in the resistor, and no power lost, though of course,
the minimum phase difference of the system is 90 degrees. The original fifty kilowatt transmitter had a "rather complex vacuum tube RF voltage regulator", as did RCA's version when it was released several years later.
McClatchy Broadcasting also operated KMJ Fresno, and KBEE Modesto. It is possible that outphased transmitters may have been used in at least one of these stations, too, though no firm information on that is currently available. It is believed that having pioneered outphasing modulation, McClatchy persuaded RCA to build these on a commercial basis and the ampliphase was born. The exact details of this "handover" are also unknown to this author at the present time.
The 50G, the first example of which was shipped to WINS in New York on 28th December 1955, inherited a number of features from it plate modulated predecssor, the BTA-50F. This included the 807 oscillator block, and the 5671 PA tubes. Whereas the 50F had four such tubes, two in the modulator and two in the PA, plus two installed spares, the ampliphase had just one pair. This PA tube was over two feet in length, 16 inches in diameter, and weighed in at ??lbs (??kg). Putting one of these into
service was considered a two man job, and a small trolley mounted crane was provided to assist with this operation. The driver tubes on the 50G were type 6076, and the pre-driver a pair of 4-250's as used on the KOH/KFBK transmitters and as used on the later 50H and 50J systems. In 1955 solid state rectifiers were in their infancy, so all the main power supplies to the 50G utilised mercury arc rectifier tubes.
The high voltage was 16.2KV provided by a total of twelve 6894 rectifier tubes. These were arranged as pairs, in a conventional six way three phase configuartion. Six 8008 tubes were used in another 3-phase full-wave configuration to provide the low/intermediate voltages of 2500V, and 5000V. Each rectifier (or pair of rectifiers) required its own filament transformer, which had to be insulated to withstand the 16KV HT voltage.
As Mercury rectifiers will not strike in low tempeatures (the vapour will condense), a 400 watt heating panel had to be fitted into the rectifier cabinet to prevent the tube temperature dropping below 20 deg C. Once the transmitter was up and running, a blower motor had to circulate air through the cabinet in the normal manner, to prevent the tubes from overheating. A selenium-metal rectifier was used to provide the -400volt bias.
RCA made big claims in 1955 about the compact size and lightweight of the 50G, which was approx half the length of the 50F, and with no modulation transformer or reactor was a third of the weight.
Although produced as a 50KW transmitter for the domestic market, there is evidence that 100 and 250 kilowatt versions were manufactured and shipped to Iran and Mexico. No further information has as yet come to light on these variants.
Early pictures of the prototype units show that a modulator, with most of the components laid in a single horizontal line was fitted, and that the Crystal oscillators were fitted at the bottom of the cabinet, in the space which was later used to house the drive regulator. There are reasons to suspect that this used a single phase modulator, followed by a tripler, but it is not known if this modulator was purely a prototype or if it was shipped with early units. Later versions of the 50G had
the "conventional" cascaded modulator, arranged in four vertical columns, with the drive regulator at the bottom of the cabinet, and the oscillators to the side. Approximately 30 examples of the 50G are thought to have been shipped.
|
| Fig 1: Diagram of KOH homebrew ampliphase, 1951 | |
| | |
 | |
Fig 2: 50G Handbook Cover
- spot the printing error! | |
| | |
 | |
| Fig 3: 50G Mercury Arc rectifiers and heater panel | |
| | |
 | |
| Fig 4: The early 50G "horizontal" modulator | |
| | |
 | |
Fig 5: 50G installed at
WLKW Providence | |
| | |
 | |
The 50H, released in late 1960 in turn inherited features from the 50G, including the four column cascaded modulator, drive regulator, general electrical and physical layout and that ubiquitous 807 oscillator block. However, the 50H contained many "new" features, such as silicon rectifiers throughout, an updated tube line up, and facilities for remote control. The massive 5671 PA's were replaced by a pair of 6697's, and the 6076 drivers were replaced by a pair of 4CX5000A's.
The 4-250 drivers were retained, but now ran from a 1000 volt supply, and not the 2500 volt supply of the 50G. The intermediate and high voltages remained the same at 5KV and 16KV, though the bias was increased to -950 volts to suit the new PA's. RCA claimed the use of silicon rectifiers allowed operation as low as -20 degrees, although no claims were made for how well the engineer would operate at that temperature! See my "ampliphase" page for a detailed description of the 50H.
A later variant of the 50H, the 50H1 was produced sometime in 1966, though the technical differences of that to the earlier unit are not too clear at the moment. However on the cosmetic side, the unit benefitted from RCA's "New Look" marketing, and was shipped in a two-tone grey and blue finish. When RCA later produced a solid state modulator, transmitters could be field upgraded by RCA engineers to become a "50H1S", which included the later alterations to the "H1" (presumably without the new colours),
plus the solid state modulator designated by the "S". It is believed the first unit upgraded this way was WNBC in New York. Upgraded units were given a new serial number plate, showing the new designation, though it seems as if they retained the original serial number. It appears that a considerable number of units had the solid state option fitted, as the ageing and drift of the 19 tubes in the modulator was a frequent cause of poor performance. The prototype of the solid state modulator is rumoured to have been built by Radio Pakistan, and the design sold back to RCA.
The development of this unit must have been a considerable project for a broadcaster, so it is possible they operated several of these transmitters. My investigations into this unit and its development are continuing.
A good many 50H transmitters were produced, and a fair number of these continue in standby service into the new century. It also appears there are a handful of stations using them as their main.
The 50H was succeeded by the 50J in 1970, and included the solid state modulator as standard. This transmitter also had the "New Look" and the new square "RCA" logo (introduced in 1966, following the appointment of Robert Sarnoff to head RCA, replacing his father David Sarnoff) and was shipped in single tone blue, but had solid doors, without the sight windows, and new, rectangular style meters. RCA proudly claimed "Only 6 tubes in transmitting circuits" though another four formed part of the monitoring and line protection unit. Information on these units is fairly scarce, and even RCA's own informations shows that only about 12 were shipped. The last 50J left the factory in May 1977.
Advances in tube technology by the 70's meant that screen-grid and pulse-duration modulation were commercially viable, and transmitters such as the Harris MW50 and Continental 317 started to dominate the market. RCA Broadcast catalogues of the late 70's/early 80's show the only AM transmitter available from RCA as a 1kw model.
By 1962 RCA was also producing a ten kilowatt Ampliphase, the BTA-10J. This unit featured a pair of single phase modulator stages, followed by frequency doublers. It also featured a small sized crystal oscillator (at one half of the carrier frequency) and not the 807. The oscillator, modulator and driver were built into a single pull-out drawer style chassis, with the PA's in a cabinet above the drawer. The whole transmitter was essentially built into a single width cabinet, with the active components in the middle and metering panels and controls down both edges.
BTA-10J transmitters were used by a number of UK offshore stations during the 1960's, notably Radio Scotland onboard the MV Comet with a pair of 10J's, Radio 270 onboard the MV Ocean VII and Radio 390. The 390 transmitter was later sold to the Mebo company for use onboard the Radio Northsea ship, and the 270 transmitter was acquired by the Captal Radio project onboard the MV King David.
A BTA-10L was subsequently introduced featuring the solid state modulator - this transmitter was also available in a 5KW version, BTA-5L.
I would appreciate any photographs of installed 5 or 10 KW ampliphases.
|
Fig 6: The 50H in conventional RCA brown
- WMAQ Chicago. | |
| | |
 | |
Fig 7: An ex-Canadian 50H in blue
- Radio Caroline | |
| | |
 | |
| Fig 8: The 50H1 in striking "New Look" two-tone colours | |
| | |
 | |
| Fig 9: The windowless 50J in blue with RCA's new meters | |
| | |
 | |
Fig 10: The 10J in brown
- Radio Scotland 242 | |
| | |
 | |
A one hundred kilowatt shortwave ampliphase, the BHF-100A was manufactured in small quantities, the first of which shipped in 1963. The design of this was significantly different than the AM version, due to both the frequencies involved, and the need to change frequency on a regular basis. Four 6697 PA tubes were used in a grounded grid arrangement, with two in parallel on each side of the modulation chain. The modulator was single stage, followed by a tripler: it had to operate wideband, from 1-9 Mhz, with minimal tuning prior to triplication.
A 6EA8 triode-pentode was used for this, the triode functioning as a Belaskis modulator, and the pentode section as multiplier. The entire frequency range was accomodated in four bands, with a motorised control adjusting six inductor slugs. Each PA driver stage comprised a pair of 4CX10000 tetrodes, with an unusual bias and drive regulator arrangement. At carrier power and below only one driver tube operated, with the second biased beyond cut-off. As the required carrier level increased, the second tube was brought into operation thus
increasing the drive power to the PA stages. (Sitting here 40 years later, I can't help but think that two tubes, one for carrier, and one for peak has similarities to a rival modulation system....). The output of the transmitter was a balanced line, of 300 or 600 ohms. In the "RCA Broadcast News" article of August 1962 in which the transmitter was described, the concluding paragraph stated that a diplexing system was available to combine two transmitters to give 200KW of carrier. A device RCA called a "computer module" was used to tune
the 8 PA tubes for maximum efficiency in this configuration. From reports received, it seems these transmitters were manufactured in Italy by a sub-contract manufacturer, and not by RCA in Camden, NJ - more information on this arrangement is being sought. It seems that RCA made an engineering oversight with the mechanics of the transmitter - being intended for an international market, the assembled units would have to be shipped by air freight to their destinations. However, the units were too bulky for most planes of that period, and even the largest commercial freight planes could accomodate the units with just a few inches to spare.
See my "anthology" page, for details of the small number of HF ampliphases which were made. Three of these are still in use today.
A one hundred kilowatt standard AM ampliphase was produced from the time the 50H went into production, the BTH-100B. This basically consisted of four 50H PA cabinets with a single exciter/modulator cabinet - indeed, RCA documentation shows this was handled in production as two separate 50H models. One of these, serial number 2, was used onboard the Radio Northsea International ship, MV MEBO 2, anchored off the UK and Dutch coasts from 1970-1974. It is probable that only a small number of these were made, all for the export market. The RNI unit, which was most likely bought 2nd hand at only 7 or so years old apparently suffered overheating problems when running on full power. Other BTH100B's may have been shipped to Finland and Nigeria.
|
| Fig 11: Front view of BHF-100A | |
| | |
 | |
| Fig 12: HF100 Block Diagram | |
| | |
 | |
Fig 13: The 100KW Ampliphase onboard Mebo 2.
Photo: Martin Stevens/Gerry Bishop | |
| | |
The BBC and Marconi undertook experimental work with ampliphase technology in the late 40's, which culminated in the trial of a 60KW set at Daventry for the 3rd Programme on 647kc/s from 15 March 1950. This was to implement the post-war "Copenhagen" frequency plan, and saw the 3rd programme move from 1149khz at nearby Droitwich. The trial lasted until a higher-power set (200kw, Marconi) was installed on 8 April 1951 and the ampliphase was used for a short while as standby.
Although the set was noted for being remarkably small and performed satisfactorily during the trial, the BBC was not in the market for new TX's at that time and the set was returned to Marconi. No more were installed in the UK though no further information on overseas use of the Marconi ampliphase system has been uncovered at the time of writing.
If you can shine any light on any version of ampliphase or wish to correct any of the above information please feel free to contact me. In particular I would appreciate photographs for the five and ten.
I really appreciate all feedback to these pages. Thanks!
With many thanks to Duffy Egan, John Stanley, John Byrns, George Saunders, Roger Stubbe, Dave Porter and all contributors for their help !
Many of the pictures on these pages were taken from RCA publicity material of the day.
My E-mail address is : "ampliphase [at] radio [dot] fm" or use the links to the feedback form. Thanks!
This page first isssued March 2002 (C) Last Major Update August 2002.
|
