Sunday, September 18, 2022

A Technological Appendix to Postblogging Technology, June 1952: Sidebands, AI Mark 17 and AIRPASS


Two 11,000lb Sapphires! Did we mention that?  Two posts back with Flight's insufferable smugness and I am starting to root for metal fatigue. Yes, Aviation Week is louche as all get out, but at least it isn't afflicted by whatever is bothering Flight. (Could it be an inferiority complex?) 

So this week we have word that the Javelin has been ordered as an emergency super priority, to give the the RAF the "all-weather fighter" capability it so desperately needs. The ad promises Hawker Siddeley shareholders even more: "Capable of continuous development in many roles," which doesn't exactly pan out. All that power, almost triple that available to the F-104 comparing both power plants at full afterburner and high altitudes, and the Javelin can't even make it past Mach 1.0. While the Javelin faces more onerous endurance, crew, and payload requirements, the fact remains that it needs significant aerodynamic improvement into the "thin-wing Javelin" to accomplish those "many roles," and that will be overtaken by the Sandys Report. 

All of this is perhaps less relevant in June of 1952 than the still-classified radar going into the Javelin, AI 17, Ferranti's winner in a competition against GEC for a "lock and follow" successor to the abortive wartime AI Mark IX. I have previously discussed around here because of the choice to replace it on the earlier Javelin models with an American radar, in which I might perhaps have taken too much of an "Imperial sunset" perspective. 

It seems, in fact, that the Air Ministry has tired of all that old stuff and has its eyes on something shiny and new. AIRPASS, which began design the year before, will make its first flight in 1958, and looks to crowd the Javelin's lifespan. Maybe an "interim" radar is a better approach than an expensive attack on AI 17's current problems?

Oh, and there's that bit about "American supermen are our superiors." Always good for laughs!


AI 17 is a conical search radar, which means that the radar beam is emitted in a beam slightly offset from the centre of the antenna, on a feed horn that rotates around the centre, alternately painting the target with more or less flux, producing a signal that varies in strength. It is fairly simple to design a circuit to control the motor that moves the feedhorn so that it registers the maximum signal strength, and then, once  that has been ascertained, or "locked on," directs the feedhorn to move to maximise that strength. 

All of this is easier said than done, especially leaving out damping. Given the grotesque failure rate of contemporary vacuum tubes, it is not surprising that early examples of AI 17 were highly unreliable. All that said, the IATA's report this month, indicating a strong preference for Single Sideband radio equipment, signals a much more important weakness in AI 17, the failure to exploit the potential additional information gained by treating each lobe of radar signal separately.  

AM transmissions rely on mixing an aural signal with a carrier wave to produce one of those wavy, warbly lines that we're familiar with as a visual representation of a radio broadcast.  The bits of the wavy line that go above the straight-line base of the carrier wave are one sideband; the ones that go below, are the other. Somewhat mysteriously (but the math says it is true!) both sidebands carry all the information from the aural signal. This allows for the opposite strategy to the one I sketched for radar, in which the equipment suppresses the redundant sideband and saves on . . . sidebands, I guess. No, actually, it reduces the power requirements, since the mechanics involve suppressing the carrier frequency to eliminate one sideband,. This means putting something into the receiver to regenerate the carrier, but the power gains remain real and SSB broadcasts are also less vulnerable to jamming.

All of this said, radar has no aural signal mixed in to create symmetric sidebands. All feedhorns will produce a lobed transmission to some extent, but getting information out of them requires polarising the input signal and favourable feedhorn geometry. Given that, the monotone radar emission becomes a four-lobed transmission, each lobe having a distinct pair of polarised signals. Monopulse radar returns thus automatically contain datum indicating the antenna's vertical and horizontal error off the target, greatly reducing the time required for lock on, although the modern emphasis is on increased accuracy.  I can certainly visualise early applications making the "following" driver circuit simpler and more reliable. It also prevents the target from spoofing the signal, since the signal received is of unknown polarisation. This (briefly) reduced defensive ECM's range of options, as already noted by IATA in highlighting SSB's greater resistance to jamming. 

Somewhat annoyingly, monopulse radar is yet another invention claimed for Robert Morris Page of the Naval Research Laboratory based on his, and his institution's willingness to apply for US patents. Apparently, he invented it in 1943, just like he invented radar in 1934. I found a fuller discussion of the NRL work here, For Ferranti's process in developing AIRPASS, later AI 23, we have first a photo, kindly posted by Paul S. Rose on Flickr, which I, less kindly, have scraped for your viewing pleasure, and secondly a very slow burn threat at secretprojects, including the pilot's notes, and an extract of a Ferranti house journal, mainly interested in the numerically-controlled milling machines that turned the feedhorns out of a solid block of aluminum with little insight into the history of the electronics engineering research behind it. I did learn about Ferranti Crewe Toll, the Edinburgh factory built in 1943 to produce gyro gunsights for the Air Ministry, which is where the research work was done, and helps explain why AI 23 was integrated into the first head's-up display. This, combined with the rapid development of terrain-following radar from the AI 23, suggests a connection with the ongoing Autoland research project. 

For anyone interested in following up on my brief and no doubt hideously wrong summary, I found an entire monograph on monopulse radar still up on Google Books by some miracle. (Totally unrelated, I ran into a 1971 article from the Journal of Church and State on the "Yale apostates" of 1722 that was embargoed at JSTOR the other day. I'm definitely going to pay to see that, guys!) So, anyway, there you go. All the details about monopulse you ever wanted to hear. Except for the history. That can just go in as citation references.  

No comments:

Post a Comment