Oral History

M.J.B. Scanlan


The Radar News

Radar theory

RAF Radar

Radar Personalities

Oral History Home

Gentlemen, that reminds me......

Radar Jargon

Help Wanted!

How it's done!


Radar, Service and Cold War links

Contact the Editor

FAQ (Frequently Asked Questions)


Chain Home Radar - A Personal Reminiscence

These three valves were CATs (cooled anode transmitter), the anodes being so heavy as to require a chain and pulley for lifting. The purpose of this elaborate cooling, pumping and lifting system was to allow the valve filament to be replaced. The filament began life as a thick (5 mm?) hair-pin of thoriated tungsten, 30cm long, and ended its life when it became so thin as to fracture under the inevitable vibration. The power consumed by these filaments was enormous: 2.6 kW for the driver, 5.5kW for each output valve. Since even the standby transmitter ran with its filaments hot, the total filament power for the six main valves alone was over 27kW, as compared with the mean RF power output of about 200W (25 pulses per second, l6us pulse width and a peak RF power of 500kW). When all the other power supplies are added in, together with power for vacuum pumps, water pumps and blowers, it will be appreciated that the overall efficiency (that is, the ratio of mean RF out-put to mains input power) was only about 0.5%.

Replacing a filament was a lengthy and error-fraught procedure, which I went through only once: fortunately, my C.O. at the time was an experienced CH engineer, who had done this job before. The valve was allowed to cool, the pumps switched off and the anode lifted against its weight and the atmospheric pressure acting on it. With the anode removed, the screen and control grids could also be removed and the new filament bolted in place: finally, the grids and anode were replaced and the vacuum pumps restarted. Absolute cleanliness was essential, since the grease from a single finger mark would delay final evacuation for hours, besides leading to the suspicion of a leak: the preferred cleaning agent was ether. The final vacuum seal was between machined metal surfaces, and the anode must be lowered exactly in place and square to its mating surface: needless to say, any grit or lint between these surfaces was fatal to any hopes of an adequate vacuum. Then a ring of 'Apiezon' (the name conceals a neat Greek pun) was added around the anode base seal, the filament and grids 'conditioned' by heating to high temperatures, and the valve was once more ready to run. Considering the novelty and complexity of the vacuum systems (most vacuum pumps up to and including the 1930s used mercury vapour), the successful replacement of a filament was something of a triumph: fortunately, it was not something to be carried out every week or month.

The CH receiver was comparatively orthodox, apart from its bandwidth: however, the receiver cabinet, perhaps 2 m high by 2 m wide by a metre deep, also contained all the timing and time-base circuits, the CR0 and its power supplies, and the goniometer and range switches. It progressed through various stages of increasing complexity: in 1943, the current versions were the R.F. 6, 6A, 7 and 8 (R.F. stood for 'receiver fixed', as distinct from airborne). The circuitry (all thermionic, of course) was arranged on shallow trays, valves above and other components below; the trays were attached to front panels, which might carry switches, meters and controls, and which were attached by their vertical edges to the uprights of the receiver frame. Viewed from the front, these panels formed an unbroken surface. The display, a 33cm oscilloscope tube, protruded at the left, with its face perhaps a metre from the floor, and sloping at 30o to the horizontal. Thus, the seated observer viewed the tube squarely, with the gonio knob to her left hand and the range knob to her right. Various controls, for example, for switching the transmitter from main beam to gap-filler, or for switching the gonio from direction-finding to height-finding, were operated by push-buttons within easy reach. Also within reach were the bandwidth switch (three bandwidths were available, 500kHz, 200kHz and 50kHz, to conform to the pulse-width in use) and the receiver tuning controls (all three of the RF stages were tunable and the tuning was checked frequently). Since the display CRT was electro-static, it was relatively easy (at least by comparison with a magnetically-deflected display) to have a very linear and easily-expandable time-base. This was important, especially with manual plotting, when the time-base was aligned as well as possible with a linear scale, from which the operator read off the range of the target.

The time-base was calibrated with range markers generated from a 9.3kHz crystal oscillator, which gave marker pips at 10 mile (16km) intervals. Range accuracy was of considerable importance in the CH system, since despite the undoubted skills of the operators, bearing accuracy was unreliable on very small echoes at extreme range. However, if the same target was picked up by two or more stations, the filter room could often take a 'range cut' on the various plots and so establish the position of the target with some accuracy. In any case, of course, early warning was more important than extreme accuracy: the accuracy of position-finding, even from a single station, would improve as the target grew closer and the signal strength increased.

The signals received on the dipole arrays at the top of the tower were fed, via at least 100m of coaxial cable, to the stators of the goniometer, picked up by the rotor and transferred via slip-rings to the receiver input: inevitably there were severe losses along this path, losses which might seem at first sight to invalidate the whole system. However, the loss in signal-to-noise ratio would be much lower, because of the very high level of galactic noise present at 30 MHz.


Previous page

Top of page

Next page

Constructed by Dick Barrett

Email: editor@ban_spam_radarpages.co.uk

(To e-mail me remove "ban_spam_" from my address)

ęCopyright 2000 Dick Barrett

The right of Dick Barrett to be identified as author of this work has been asserted by him in accordance with the Copyright, Designs and Patents Act 1988.