This page describes the
General Radio Model 857a  UHF Oscillator
article and images by ralph klimek vk3zzc
This unit was purchased by me for $15  at the Rosebud Hamfest in 2013. It is a piece of world war two heritage.  I generally do not collect or want WW2 relics because my shed is already too full of junk  and any further junk addition must perform a usefull function for me, as I do not run a museum.   It did not know when I bought this that it was constructed in 1944 according to the General Radio Catalogue.  The active ingredients in this unit are the butterfly resonator  and an instance of the famous Western Electric  Door Knob triode, manufactured originally in the early 1930s for the microwave research community.  Mine is a slightly more recent  General Electric  war version.  

So after a long sleep of 70 years , would this antique work again ?
Yes !

It was possible to download the GR catalogues and a service note from IET laboratories that maintain the GenRad heritage. The filament of this antique triode is a thoriated tungsten directly heated cathode that wants 2.2  volts at 3.5 Amperes,  an odd voltage requirement to be sure.  I had to hand wind 30 turns through a surplus toroidal mains transformer  ( a great way to light up old Transmitter Bottles with odd filament voltage ).  The original unit was meant to be operated with the GenRad custom power supply which included a grid leak resistor and metering of grid current with a magic eye tube, this being used due to war shortages rather than a more usefull  moving coil meter. This could make the instrument behave like a grid dip meter.

It tunes from 90 Mhz up to the dizzying heights of 515 Mhz  which was considered   microwaves in those days.  The anode on the triode is made from molybdenum  and can dissipate about 5 watts of power. The oscillator was rated to produce up to 1 watt of output power.  At this power level the anode will glow dark red.  The real gem in this unit was the resonator that permits 2 octaves of tuning range without band switching !  Thats why the resonator design  was patented ! Just try tuning
100-500 Mhz with a variable capacitor or varactor...it cannot be done.



The oscillator despite its single triode has remarkable frequency stability due to its solid cast metal construction and the extremely high Q of the butterfly resonator.  After 1/2 hour of warmup this can be netted on a narrow UHF chanel and it just stays there. Absolutely amazing ! Even better with a stabilized power supply.  Frequency stability is still a function of output coupling as you would expect.


Powering up for the first time in 70 years.

The filament is meant to run at bright yellow heat and emits a lovely bright warm glow in operation. Original cost in 1944 was $250 US dollars, a small fortune at that time.  I let the filament condition itself.  Really antique thoriated filaments should be run at their rated current for a few hours  without applying high tension.  This gives time for the thorium to migrate to the filament surface and permit outgassing.  It is the thorium  atom layer that supplies the electrons, not the tungsten, which is an electron emitter only at white heat.  The initial current draw at 250Volts  was 70mA  which is  way too high. This stabilized by itself in about 30 minutes.  I expect that this was due to filament outgassing  and the tramp gas before making its way to the getter.  Do not permit anode hight tension to remain if the oscillator stops.  Otherwise ,Substantial and destructive anode current will flow because grid leak bias will not appear to limit anode current.  The anode is Molybdenum and will tolerate a some abuse. 
The output coupling link can be rotated to vary the output power.  It should not be rotated all the way in, this will result is too much power being removed from the resonator quenching the oscillation, particularly  at the upper frequency limit of the unit.
Frequency stability after a long warm up period is superb.

The General  Radio instructions specify 350V for the anode high tension,  mine works fine at 250V.

I used carburetter cleaner to  dissolve and remove the ancient, completely dried out grease on the gearing  and put a little dab of modern teflon grease on the worm. The sealed ball bearings  were still serviceable.


This oscillator would have been used for  deci-metric radar development, it would have been too expensive for day to day field alignment work. I wonder what it was doing in Australia ?  It will find a second life in the shack for tuning my UHF and VHF antennas.



thumb.imgp7044-gr857-osc.jpgthumb.imgp7046-gr857-osc.jpgthumb.imgp7047-gr857-osc.jpg
output port has a PL259compliance plate
thumb.imgp7049-gr857-osc.jpgthumb.imgp7050-gr857-osc.jpgthumb.imgp7051-gr857-osc.jpg
output coupling loop can turn to adjustoutput coupling, and goes in and out on the screw thread  for optimum resultsATT , Bell Labs owned the butterfly resonator patent and were not letting anyone else have a go. I think the state overode this restriction in this case. I think the term  is Sovereign Right in times of war
thumb.imgp7053-gr857-osc.jpgthumb.imgp7054-gr857-osc.jpgthumb.imgp7055-gr857-osc.jpg
massive mechanical engineering on a solid cast aluminium chassisthe red grid leak resistor is my addition. The gearing has an anti-backlash design. Tuning is very smooth to permit the oscillator to work without the custom GR power supplymassive mechanical engineering
the resonator is at anode potential so dont touch !
thumb.imgp7056-gr857-osc.jpgthumb.imgp7058-gr857-osc.jpgthumb.imgp7059-gr857-osc.jpg
nice view of the anode here.  IT is a huge bottle containing a very tiny triode. No socket, connections made directly to the valve pins.butterfly resonator at near max frequency. The small bit with the slot is an adjustment trimmer to align the dial with reality.butterfly resonator at near min frequency
thumb.imgp7060-gr857-osc.jpgthumb.imgp7061-gr857-osc.jpgthumb.imgp7062-gr857-osc.jpg
yet to come, extract from the General Radio Experimentor and Catalogue.  I think the butterfly resonator is described in the Bell Systems  Technical Journal from the mid 1930s.  Maybe I can still find the patent











2-april-2014  first draft