History of British radar tubes, part III

External-anode VHF / UHF power valves

In the early 1920s the tube industry started using external anode triodes to handle anode power exceeding about 1 kilowatt. Anode was a hollow copper block sealed to the glass bulb holding the grid-filament subassembly. Copper could be easily worked to any shape, but its extremely high thermal expansion coefficient obstacled sealing to any type of glass. The problem was solved by W. G. Housekeeper at Western Electric, U.S. Patent 1,294,466 released in 1919. The Housekeeper process involved tapering the copper edge, to form a ‘feather-edge’, and then sealing the glass to the thin edge, flexible enough to safely relieve stresses due to the different thermal expansion of the glass. Depending upon the power and the selected cooling system, forced-air or water, anode was surrounded by a finned radiator or by a water jacket. In radiation-cooled tubes anode must be large enough to dissipate heat and the maximum heat transfer is reached when it becomes very hot, up to cherry or even to bright red. Due to the excellent heat conductivity of copper, in external-anode tubes heat is completely transferred to any suitable external radiator. Anode size can be then much smaller, a useful feature when trying to design electrode structures of high frequency power tubes.

Fig. 9 - Pre-WWII external anode power tubes. A) UV-207, also known as 207 or VT-34, was one of the early external anode power triode, introduced since 1923. With external water jacket its plate could dissipate 10 kW. B) The 891R was quite similar, intended for forced-air-cooling. Its heavy finned radiator was suited to dissipate 4 kW. C) In the 1930s GEC was one of the leading suppliers for external-anode transmitting tubes. This STC CV28, without radiator in the photo above, was equivalent to GEC ACT9. D) Early in the 1930s GEC also introduced a line of smaller triodes with external anode, referred to as ‘Catkins’. Here a prototype with a small finned radiator screwed over the copper plate. E) One of the ‘catkin’ tubes was this NT39, a compact 75 W transmitting triode operating well over 30 MHz.

In 1938 at GEC a group leaded by Robert le Rossignol designed an external anode triode intended to replace the ‘silica valves’ as the NT57, whose demand by military far exceeded the production capacity. In order to operate in the VHF region were used small electrodes, anode active length being reduced to about 50 mm The E960, approved as VT58, operated up to 23 kV anode voltage, dissipating 750 W with forced-air-cooling. Its tungsten filament required 12.6 V at 58 A. The tube was specified for operation up to 100 MHz but it could be used up to 250 MHz with good efficiency. Not only it replaced NT57 in the existing ground and naval radar sets but it was used at 200 MHz for the prototypes of the coast watching CHL system. A pair of VT58s generated 30 kW typical pulses.

From September 1939, the tungsten filament was replaced by a thoriated-tungsten one. Emission of the new VT98 raised to 25 A at less than one half the heating power of VT58. Although electrodes were only 50 mm high, much less than in other transmitting tubes, with the VT98 the pulse power of the CHL and MB2 transmitters increased to 100 kW. The new triode was widely used not only in England, but also by Canadian REL in its Coast-Defence radar, CD, similar to the British CHL. Lots of the tube were then manufactured as REL-5 by Canadian Westinghouse, some with the dual mark REL-5/VT98. The CD system was also supplied to the United States for the surveillance of the Panama Channel, and other tubes were manufactured by the U.S. Westinghouse, some units with the proprietary marking WL-533.

Fig. 10 - A) GEC built several thousands units of VT98 / CV1098 during the war. In the British tubes anode radiator was made of solid perforated copper. B) This REL-5 / VT98 was made in the U.S. by Westinghouse. C) and D) Details of the Westinghouse tube, showing the double marking, REL-5 / VT98 in this case, and the realization of the finned radiator in place of the British perforated one.

After the VT58, a larger tube with similar construction was sampled by GEC around August 1939 to replace the NT60 tetrode: it was the E1024, approved as VT114 with thoriated-tungsten filament granting 70 A emission. In a push-pull oscillator a couple could deliver 400 kW pulses at 50 MHz.

- The ‘micropup’ triodes’, 1939 onwards

Fig. 11 - Draft showing the construction of a ‘micropup’ triode and enlarged view of the ‘squirrel cage’ grid of a ‘milli-micropup’ designed to operate up to 1.25 GHz.

In 1939 at a demand of the Air Ministry GEC introduced a smaller triode with external anode, the E1046, designed by the group of le Rossignol and approved as VT90. The triode was intended to replace the 4304CB in AI and ASV radar sets. Its design was innovative, anode being a small copper tube, both ends being flared and sealed to two glass bulbs. A ‘parrot cage’ grid, made with short molybdenum wires welded along the edge of a small dish, was supported by a rod sealed through one of the bulbs. The cathode was a spiral of thoriated-tungsten supported by two tungsten rods sealed through the opposite bulb. At the end of the manufacturing process, the design granted accurate and close spacing between electrodes. The finned radiator was then brazed around the anode. VT90 was rated for 100 W plate power dissipation and 300 MHz maximum operating frequency. Anode voltage could reach 9 kV pulsed, emission being greater than 5 A.

VT90 was favorably accepted from December 1939 onwards. It was used in the AI Mk II and in the ASV Mk II. About 18.000 ASV-like sets were also built in America after the Tizard Mission by Philco and by Canadian REL. They were known as SCR-521-A or SVC and SCR-521-B or ASE. Many tubes were therefore built by local firms, Western Electric, Northern Electric, National Union, Amperex and RCA.

Worth of note is the CV15, a conduction-cooled micropup without the finned radiator. The slender shape of its central anode among the two glass bulbs well explains why these tubes were referred to as ‘micropups’.

Fig. 12 - The VT90 and some of its copies and variants. A) Sample of the British VT90. B) In the U.S. RCA, Amperex and National Union built the same triode as 8011. The sample in the above image was made by Amperex. C) Western Electric marked its production with the proprietary code 710A. The above sample was made by National Union for WE. D) This REL Type #1 was made by Canadian Northern Electric, related to WE. E) A sample of the CV15, supplied with a rectangular mounting block to provide conduction-cooling. The block is 15 to 17 mm high, hence we can assume that the anode active length should not exceed 15 mm.

After the war RCA proposed more powerful variants of VT90 intended for industrial heating, the 6C24 and the 8014. They were uprated respectively to 600 and 400 W dissipation and specified for CW operation. The size of them both was larger than in the original one.

Fig. 13 - RCA industrial heating triodes A) A sample of 8011 - VT90 as reference. B) The 6C24 shows larger anode diameter, heavy radiator and double CT filament. C) 8014A shows the same large structure.

- Oxide-coated cathodes

Until June 1940 cathodes for transmitting tubes were strictly filamentary, at most made of thoriated-tungsten wire. It was common belief that ion bombardment, due to residual gas particles accelerated by the high voltage used in the transmitters, would have destroyed the oxide-coating layer. Right at GEC in July 1940 Eric Megaw, with his magnetron E1189, had demonstrated that oxide-coated cathodes worked properly in pulsed applications. In fact, heavy ions could never gain harmful speed, provided that high-voltage pulses were short enough.

As a consequence, GEC launched the E1232 design, with the more efficient cathode. First samples were released in April 1941, while production started in July. Internal drawing can be appreciated in fig. 11-A. The cup-shaped nickel cathode was heated by the tungsten spiral inside. The ‘parrot-cage’ grid, similar to the one in fig. 11-B, was tightly screwed to the connecting copper plate, sealed to a folded glass spacer. The whole assembly was extremely rigid, electrodes being very closely and precisely spaced. The tube operated up to 600 MHz, pulsed to 8 kV with 40 A emission. Two tubes in push-pull generated 150 kW at 200 MHz and about 100 kW at 600 MHz. Approved as NT98, its success was even larger than that of VT90. It was used in several ground, naval and airborne sets at 200 and 600 MHz. It was also used in Canada and in the U.S., where it was built by National Union and Central Electronics as 4C27, by RCA as 8026 and by Rogers as REL 7.

Fig. 14 - NT99 and its variants. A) The British NT99. B) A quite late sample of 4C27 likely made in the fifties by Central Electronics. C) A Canadian REL 7 made by Rogers. D) During the war RCA introduced the 4C28, a variant of NT99 designed for its SHORAN navigation and bombing system. The system was used after the war for oil exploration. E) 4C29 was made in Canada by Rogers for REL and was registered also in the U.S.

In April 1944 GEC started designing a new micropup triode, released shortly before the end of the war and believed to be the largest one ever built. With its massive radiator the E1495, approved as CV240, was specified for 1 kW plate power dissipation, 15 kV anode voltage and 125 A minimum emission. It could be used as pulsed oscillator up to 100 MHz. Other structures were also defined, capable of operating at 600 MHz and up and readily adaptable to coaxial resonators. The CV288 was designed in 1943 and tested at TRE, generating 50 kW peak output. There was no operational use during the war but the tube was proposed again after the war as wide-band UHF amplifier for TV repeaters, with the commercial code ACT25. The same electrode arrangement was retained in the post-war CV2200/CVX2200, which could operate as pulse oscillator up to 1.5 GHz.

Fig. 15 - A) CV240 was the largest micropup triode ever made, with its 1.2 kg weight. B) The basic shape of the external anode triodes was modified to design tubes adaptable to coaxial resonators, as this ACT25 derived from the CV288. C) The same structure was retained in this CVX2200, capable to operate up to 1.5 GHz.

- Milli-micropups

A little known family of UHF triodes was the one referred to as ‘milli-micropups’, whose design started early in 1940. In response to the request of Air Ministry for a system operating at 10 cm wavelength, GEC proposed the development of power triodes capable of operation at 25 cm, or 1,200 MHz. First laboratory samples were ready since March 1940 and an experimental system was already installed on the roof of the Wembley Laboratory in the summer, when Megaw started testing his magnetron. Subsequent development stages up to the production were delayed, after Megaw's magnetron had run at 10 cm beyond all expectations. Production of E1190, tested and approved as CV55 for CW and as CV155 for pulse operation, started only in 1941. Although a pair of tubes could generate 40 kW at 1.2 GHz, they were used only in niche applications, such as IFF transponders. An high-gain variant was the E1458, rated both for CW and pulse operation. Based upon the E1458, National Union in America designed its own 3C27, with coaxial heater-cathode connector. 3C27 was modified to 3C27B, adding a radiator on the grid, and then to 3C37, which was advertised at the end of the war as capable of generating 10 kW pulses at 1.150 GHz.

National Union also registered the small 3C36, an improved design whose experimental code was R1001. Its was specified for pulse operation between 500 and 1500 MHz. Emission was as high as 50 A. It could accept a forced-air cooling radiator up to 200 W or a water jacket, power dissipation reaching the exceptional value of 500 W.

Performances of the milli-micropups stood unsurpassed until the appearance of power planar tubes. To understand how advanced this family was, we must consider that when it was designed, in 1940, the best German UHF triodes, the LS180 and the TS1, could operate up to about 600 MHz, RF power not exceeding 8 kW at the best. Western Electric itself in the U.S. had nothing better than doorknob tubes and the transmitter of the experimental radar CXAS in October 1940 could generate only 2 kW pulses at 700 MHz, using four triodes in its ring oscillator.

Fig. 16 - A) Anode of CV155, E1190 tested for pulse applications, was only 7 mm high. B) National Union 3C27 evolved in the 3C27B (C) with the top grid radiator, to prevent grid emission when heavily driven positive during pulses, and then in the 3C37 (D). E) A rare sample of the tiny National Union R1001, registered as 3C36.

Fig. 17 - The National Union ad for its 3C37, from April 1946 issue of Electronics.

 

To be continued....

To parts I and II

To thank the Author because you find the post helpful or well done.