Notes
Five technical notes on public data: three studies and two reviews. Each states its basis, its validation and what it does not check. Each note has its own page, with the PDF for download. Written on personal time, on a personally held SOFiSTiK licence, using public data only.
The MODE jointed deployable truss flown on STS-48 is rebuilt from public documents and reproduces the 1993 analysis of record to within 0.8% without calibration. Three quantities are then predicted with the flight data held out, and seven candidate mechanisms for the measured ground-to-orbit softening are eliminated, each with its modal selectivity and gravity driver quantified.
A declared three-panel wing, dimensioned after a publicly documented class of array, is carried through 20 configurations across three deployment sequences using a geometric-nonlinear chain and an independent planar multibody model. The study tests what happens when the damper representation used for the torque-margin budget is carried into the dynamics.
A Monte Carlo study propagates 22 tagged parameters across five independent channels into deployed shape and deployed frequency, run through two solvers on identical per-sample seeds. It asks which toleranced quantity actually governs deployed geometry.
A review separating binding requirement from convention across the NASA and ECSS structural verification documents, and identifying four places where the requirement thins to judgement: unverified intermediate deployment configurations, uncharacterised joint nonlinearity, no anchor for model credibility when full-scale test is impossible, and unprescribed tolerance combination.
The method for handling a located, deterministic imperfection is already written down — NASA SP-8007 Revision 2 defines an imperfection signature, ECSS prescribes measured and worst-case imperfection analysis, and Koiter supplied the theory in 1945. What does not exist is a signature for packaging. This note sets out what is known about fold- and stowage-induced imperfection in deployable structures, what is missing, and why a packaging signature is a function of the article's history rather than of its design.
PSE-TM-01 measures its ground-to-orbit result against a reassembly scatter of 0.54%, taken from the source literature. PSE-TM-03 then shows that figure to be misattributed and pooled over hardware the model does not contain, and revises it to 0.082% with a 95% confidence interval of 0.053 to 0.181% resting on three tests. The two notes therefore use different reference scatters, and PSE-TM-01's significance statement is being re-expressed against the revised figure. The correction strengthens the finding rather than weakening it. Recorded here rather than quietly reconciled, because the original error was an inherited citation and that is worth showing.