Value chain · CLM-1312
In plain terms
In 1979 the contractor simulated the whole reference-distribution tree. Small, random timing errors were added at each branching level, and the simulation measured how much beam power was lost. To keep 96% of the ideal beam, the errors could add up to about 12 degrees, or about 6 degrees at each of four levels.
The claim as recorded
LinCom ran SOLARSIM on the 4-level, 102,400-tip reference phase distribution tree in 1979. With equal, independent errors per level, 96% main-beam gain efficiency tolerates about 11.8 deg total RF rms phase error, which is 5.9 deg per level for four levels or 3.93 deg per level for nine. At 90, 85, 80 and 75% the totals are 18.8, 23.0, 27.0 and 30.5 deg. The tolerance is an output for an assumed per-level sigma, not a prediction of what the hardware adds.
Evidence behind this: SOLARSIM simulation outputs in a NASA contractor report. Table values read at render.
Scope and limits
The per-level sigma is an input; the source derives no hardware value for it. 5 deg per level is used as the typical case (printed p.95, PDF 103). The source is internally inconsistent on the per-level figure. Printed p.107 (PDF 115) and Fig. 10.14 (printed p.108, PDF 116) give "12 deg total or 4 deg per level" for the baseline. Every per-level entry in Fig. 10.14 is the total divided by 3 (19/3 = 6.33, 24/3 = 8.0, 31/3 = 10.33, recomputed), which is the 9-level split, while the 4-level split would be 6.0. The tree node counts also conflict: level 1 has 4 PCCs (Fig. 4.3, printed p.35), 16 (Fig. 10.1) or 25 (printed p.95). Converting the total to millimetres is a register step, not the source's: 11.8 deg at 2.45 GHz is 4.0 mm free-space equivalent (lambda 122.4 mm, recomputed). The tree itself runs at 490/980 MHz, so physical cable tolerance would differ.
What it assumes
Equal sigma per level; independent Gaussian errors; the 10 dB taper baseline; RF-equivalent phase (IF build-up about one fifth, 2.4 deg at 96%).
Where it comes from
Printed p.89 (PDF 97, render): four levels, 102,400 nodes, phase noises introduced at each level by the MSRTS. Fig. 10.1 (printed p.90, PDF 98): branching 16/16/16/25, recomputed 102,400. Printed p.101 (PDF 109, render): total rms error defined as build-up with each level adding the same amount and independent nodes. Fig. 10.11 (printed p.103, PDF 111, render): the table quoted in the claim. Printed p.104 (PDF 112, render): 5.9 deg per level for four levels, 3.9 deg for nine (paraphrase). Recomputed: 11.8/sqrt(4) = 5.9; 11.8/sqrt(9) = 3.93; exp(-(0.2059 rad)^2) = 0.9585, against the source's gain approximately exp(-sigma^2) (printed p.107, PDF 115).
Basis: deg rms RF-equivalent phase error, total and per tree level; percent main-beam gain
Related records
- Answers question
- QST-013 not published in this release
Cite this
energyaperture.com/c/clm-1312
CLM-1312, Energy Aperture, as of 1979-03.
Connected records
Other published records linked to this one. A list, not a judgement.
- Also from SRC-647
- CLM-1311The 1979 design sends a sample of the reference back up each cable and adjusts until the far end matches the source. That makes cable...CLM-1313An error near the antenna elements is different at each element. It averages out in pointing but scatters power. An error near the centre...