Value chain · CLM-1064
In plain terms
The best transistor anyone had measured in 1980 was 72 per cent efficient at one watt and 60 per cent at five; the satellite designs assumed 80, and were heavier than the tube designs.
The claim as recorded
The measured solid-state state of the art the programme held in April 1980 was well short of its 80 per cent design figure: RCA's "Twenty-five devices of seven types have been tested to date. The best results are 60% efficiency at 5.1 watts and 72% efficiency at 1.2 watts. Two amplifiers were delivered in April, meeting the Task 1 goals of 5 watts, 50% efficiency, and 8 db gain", under the paper's own heading question "can 80% conversion efficiency be obtained?"; JSC reported that with off-the-shelf FETs "a 71% power-added efficiency was achieved. However, tests of similar devices from the same manufacturer but rated for slightly less output power resulted in efficiencies of much less (~60%)"; Rockwell's two solid-state array concepts used "36-watt amplifiers" and "4,096 elements in each 5m x 5m subarray" with a coaxial distribution network of "0.67 million kilograms for the total array" and "approximately 7 dB ohmic loss (in addition to 36dB splitting loss)" (the render at zoom 5 reads "36dB" where the text layer prints "3QdB"; six 4-way levels give 4^6 = 4,096 elements and 10 log10(4,096) = 36.1 dB); and the contractors' comparison put the solid-state satellites at 7.7 and 8.5 kg/kW against the reference 6.2.
Evidence behind this: T3, contractor and JSC status reports of measured solid-state device efficiencies on small device counts, with Rockwell's array design figures.
Scope and limits
Task 1 results from a $100K contract; device counts are small and the paper says the discrepancy between similar devices "are under intensive investigation". The 7 dB ohmic loss is a network figure that is not carried into any efficiency chain in the volume.
What it assumes
No assumptions are recorded against this claim.
Where it comes from
PDF p.366 / printed 344, normalised: "The key question today is, can 80% conversion efficiency be obtained? Analysis of numerous device concepts available for consideration eliminates all but a very few. The gallium-arsenide field effect transistor is today's first choice ... the GaAs FET has the potential for 80% dc to rf efficiency. Size and hence power levels must be limited (1 to 10 watt range) for this to be so"; the results sentence as quoted; "The waveform analysis has revealed a significant realtionship between efficiency and gate resistance." PDF p.304 / printed 282: "The Gallium Arsenide field effect transistors appear to be the most promising at the present time to achieve the high efficiency, high power (5 watts) required for SPS applications"; the 71 per cent sentence as quoted; RCA contract "$100K ... awarded by the Marshall Spaceflight Center through Rockwell International". PDF p.362 / printed 340: "in the end-mounted system 36-watt amplifiers are mounted on the ground-plane, whereas in the sandwich the amplifiers are elevated to the dipoles, and their waste heat is dissipated by beryllium oxide discs"; PDF p.363 / printed 341: "Six levels of 4-way corporate divisions provide equiphase feeding to the 4,096 elements in each 5m x 5m subarray ... weight of 0.67 million kilograms for the total array using UT-47M; 250 C temperature capability; approximately 7 dB ohmic loss (in addition to 36dB splitting loss)" (render read at zoom 5: "36dB"; the text layer prints "3QdB"); "The total required filter isolation is 70 dB"; "~100 dBw pilot power". PDF p.120 / printed 97, render read: solid-state end-mounted 7.7 kg/kW, sandwich 8.5, reference 6.2.
Basis: per cent DC-RF or power-added efficiency; watts per device; kg; dB
Related records
- Answers question
- QST-001 not published in this release
Cite this
energyaperture.com/c/clm-1064
CLM-1064, Energy Aperture, as of 1980-04.
Connected records
Other published records linked to this one. A list, not a judgement.
- Also from SRC-449
- CLM-0964To build two of these satellites a year, about 300 tonnes of material had to arrive at the orbital construction site every day. Getting it...CLM-0974The 1978 reference design shows every loss between sunlight and the power grid as a chain of numbered boxes, and the boxes multiply out...CLM-0975Depending on which published assumptions you take, a satellite needs between about 2 and 6 square kilometres of solar blanket for every...CLM-1058The often-repeated claim that the magnetron is 188 decibels quieter than its carrier traces to a conference abstract in which the tube...CLM-1059By spring 1980 the engineers, the reviewing panel and the ionosphere experimenters were all saying the beam-intensity ceiling the whole...CLM-1060The 1980 review did not agree with itself on how much of a satellite's cost was rockets: one paper said over a third, the programme's own...CLM-1061In one 1980 volume the satellite cost anywhere from about 2,300 to 3,800 dollars a kilowatt depending on who added what - and the programme...CLM-1062In 1980 the launch vehicle that everything was costed on was said by its own designers to be arbitrarily big, and they proposed a rocket a...More on the source page.