Value chain · CLM-1004
In plain terms
In 1980 the systems contractor's own table said magnetrons were about three billion times (95 dB) noisier than klystrons; a year later the same contractor was quoting the tube maker's claim that a magnetron run without its heater was quieter than a klystron, while also listing proper noise measurements as work still to be done. Anyone citing "the magnetron's noise advantage" from this period has to say whose figure it is.
The claim as recorded
The magnetron's noise performance was contested inside the programme's own 1980-81 literature. Rockwell's 1980 tube comparison table puts the magnetron's noise at 35 dB per kHz below carrier against the klystron's 130 and the amplitron's 40, with RF stability "poor"; Rockwell's 1981 systems volume then recounts that in 1978 "noise in the output spectrum of a conventional microwave oven magnetron was dramatically reduced when the tube was operated with no cathode heater power", cites Raytheon's measurements of "a carrier to noise level of at least 188 dB per Hertz at frequencies outside of a 60 MHz band centered on the 2450 MHz carrier", and its 1981 research plan still states that for magnetrons "accurate noise measurements have not been performed".
Evidence behind this: T3, the systems integrator's own three volumes, which disagree with each other and with the tube developer within eighteen months.
Scope and limits
THE 35 dB/kHz IS A GENERIC TABLE ENTRY WITH NO SOURCE ON THE PAGE AND ALMOST CERTAINLY DESCRIBES A CONVENTIONAL HEATED MAGNETRON; THE 188 dB/Hz IS RAYTHEON'S FIGURE FOR THE COLD-CATHODE MODE, QUOTED AT SECOND HAND; neither is reconciled by Rockwell, and the two are not the same measurement (dB per kHz below carrier against dB per Hz). Raytheon's own reports were read at the page on 2026-09-18 (CLM-1043 to CLM-1050): Rockwell's "at least 188 dB per Hertz" and its "60 MHz band" appear in neither (Rockwell attributes them to "Report No. 5 of Reference 38", a Raytheon progress report not held), Raytheon's own figures being 155, 160, 195 and 196.8 dBc/Hz at the 100 MHz ISM band edges with the heater off; Raytheon's design figure for the SPS tube is -120 dBc/MHz, which is -180 dBc/Hz, 17 dB worse than its best laboratory reading (CLM-1046); and Raytheon measured 81.7 per cent efficiency with error bars a year before Rockwell's table printed 70 (CLM-1048).** Rockwell's phrase has a source after all, and it is Brown's own: at the April 1980 Program Review (SRC-449, PDF p.358, printed 336, read at the render) the Raytheon paper says "signal to noise ratios of 180 to 190 dB/Hz in selected tubes have been measured. The measurement sensitivity is still limited at frequencies outside of a 60 MHz band centered on 2450 MHz by the residual noise level of the spectrum analyzer." So "188" is Rockwell's restatement of a 180-190 range, and the "60 MHz band" is the band around the carrier outside which the 1980 measurement was still instrument-limited (Brown: "limited at frequencies outside of a 60 MHz band centered on 2450 MHz by the residual noise level of the spectrum analyzer"), not a specification.
What it assumes
No assumptions are recorded against this claim.
Where it comes from
Table 2.6-2, "Performance Comparisons of Basic Tube Candidates", from the render: present CW output power limitation 100 kW, 50 kW, 50 kW; present efficiency limitations at 1 per cent bandwidth at least 70 per cent, 70 per cent, 60 per cent; noise level 130 dB/kHz, 35 dB/kHz, 40 dB/kHz below carrier; RF stability good, poor, fair; gain at least 30 dB, 10 to 20 dB, 7 to 15 dB - klystron, magnetron, amplitron. SRC-577 printed 3-174, normalised: "because of its low gain (5 to 10 dB) and reputation for having a noisy output spectrum, the amplitron lost out to the linear beam klystron... In 1978, Brown made two discoveries that catapulted the crossed-field device back into contention for SPS use. He found that noise in the output spectrum of a conventional microwave oven magnetron was dramatically reduced when the tube was operated with no cathode heater power"; printed 3-195: "Certain tubes have shown a carrier to noise level of at least 188 dB per Hertz at frequencies outside of a 60 MHz band centered on the 2450 MHz carrier". SRC-579 printed 5-60, raw: "For magnetrons, accurate noise measurements have not been performed, and a low-noise mode (using cold cathodes during operation, i.e., turning off the heater voltage) has only recently been discovered"; "Current klystrons are inefficient (less than 50%), and considerable improvement is required."
Basis: dB below carrier per kHz; dB per Hz carrier-to-noise
Related records
- Answers question
- QST-001 not published in this release
Cite this
energyaperture.com/c/clm-1004
CLM-1004, Energy Aperture, as of 1981.