Scope and limits

This is a finding about SRC-085's internal consistency, NOT about the model and NOT a correction to any of the three numbers individually. The calculation cannot say which of the three is wrong, only that they do not hang together. Its value is that it removes the error budget from the argument: a uniform error factor cancels out of a peak-to-mean ratio, so the discrepancy cannot be reconciled by any assumption about losses, which is what makes it different from the peak-value disagreement recorded elsewhere. Depends on reading the source's taper as 10 dB over a 1 km aperture, which the source specifies. THE MAGNITUDE IS CONDITIONAL ON A FRAME ASSUMPTION WORTH MORE THAN THE FINDING, established by VER-1092. Peak-to-mean cancels a uniform error factor exactly, which is what this claim rests on and which verification confirmed algebraically - but it is NOT invariant to the frame. SRC-085's Figure 18 labels its x axis "Ground radius, m" and plots power density at the rectenna, so both the 23 mW/cm2 and the 5 km are GROUND quantities, while the formula divides by pi*5000^2, the area of a CIRCLE. An inclined GEO beam lands as an ELLIPSE filling pi*a*b, and the ratio scales by exactly b/a: 3.0547 becomes 3.6656 at an axis ratio of 1.2, 3.9711 at 1.3 and 4.2766 at 1.4. A frame mix is therefore worth about 30 per cent against a 5 to 7 per cent finding. The DIRECTION survives and is strengthened - the source's numbers become MORE over-determined, not less - but the 5 to 7 per cent must not be quoted as a magnitude without the frame settled. That is the same beam-normal ambiguity the register already carries unresolved against CLM-0020. The verifier did not check SRC-085's own rectenna axis ratio, so the figures above are a SENSITIVITY, not a correction. Closed 2026-09-19: p.138 makes the beam circular and only the footprint elliptical, and the axis label is exact along the east-west axis, so the b/a sensitivity does not apply (CLM-1092).