Thermocouples and the Marconi TF144

[WME_bill]

Followint the thread on Marconi TF144 - CT452A RF Generator.
Thermocouples for measuring RF output have quite disappeared. They are extremely fragile, and burn out when you look at them. Back in 1930-1940, they were usually open construction, but later they were sealed in a vacuum bulb, which improved the response time, as there was no heat loss, and also protected the fragile elements.
I attach a page from a 1967 STC catalogue, which gives details of range then produced by Ormandy and Stollery Ltd of Colchester. Not a good copy. If wanted, I will try to find the original and re-scan it.
Don't think you can get a replacement now though, as the firm had disappeared by mid 1990s. Derek Stollery held several patents for these devices and I suspect may have developed the vacuum sealed idea. He died in 2009.

I believe the TF144H used the UHF "acorn shaped" version. Perhaps someone can check and let us know. I sold my TF144H, with working thermocouple many years ago, as I had run out of space to keep it.

I am quite bemused by the enthusiasts who want to replace a broken unit with a true rms power IC. Surely this is a huge over-kill.
The output is working into the constant impedance of the attenuator, so a voltage indicator will be valid and much simpler.
When the original TF144 was introduced, in about 1935, there were no crystal diodes or vhf diode valves, so the thermocouple was the only possible device.

Silicon crystal mixer diodes were developed in the War by BTH at Rugby for radar use. The BTH / AEI CS2A was the usual one. Known also as CV101, CV102, CV103, CV253, selected for various microwave noise and performance levels. USA similar were 1N21 and 1N23 series.
After the War, they were used for all the Marconi generators. I attach an extract showing the key parts of the circuit for several models of VHF generators: TF1066-1957, TF801D-1960, TF801A-1950, TF995A-1960.
Nowadays, with germanium and silicon and schottky diodes, there is almost too much choice. In the early 1960's, the alternative to the CS2A was the GEC GEX66 (which became Mullard AAZ13) and AEI / BTH CG91H, as used in the Bradley CT471 voltmeter. and the Hughes HD1870 / HD1871, used by Advance and elsewhere in the TF144H/1. All small fast point or gold bonded ge diodes, rated at only 8V PIV. Now we would use a hot carrier like HP2800, as the TF144H/2. The CS2A seems to have been only 4V. I have one only, and don't really want to test it and risk destruction.

The meter is 100ua I expect, though it may have rather low damping. When a schoolboy in early 1950's. I bought a meter in Lisle St to make myself a meter unit. It had what I now know was a bare wire thermocouple inside which, not knowing anything about it, I scrapped, and made a multirange meter of the rest. But the meter movement was undamped, and swung about a lot. Eventually one Christmas I was given a Universal Avominor, which was a huge improvement. The permanent shunt on all the Avo range provides a good damping.
I still have both meters, but no thermocouple. wme_bill m0wpn

[Dekatron]

Could any these be of any use to you?