A .zip holding one RINEX obs file per station of the network. The stations are solved together: every pair observed at the same time forms a baseline (RTKLIB), each station is tied to the nearest Sabah CORS, and the whole network is adjusted by least squares. You get a network adjustment report (baselines, loop closures, residuals, 95% accuracies) plus the usual SGD2020 report for each mark.

What to observe. Not sure how far your marks are from a CORS? Open the planner map to check each mark and see which case applies.

• All stations within 10 km of a CORS: at least 2 stations, observed at the same time. Each is tied to its nearest CORS and to the others, so together they close a loop (CORS → A → B → CORS) that checks every mark.
• Any station further than 10 km: at least 3 stations, and one of them (the hub) observed for 6 hours or more. The hub is solved against the CORS network by double-difference processing; the other marks are tied to it and to each other.
• Each baseline needs 15 minutes of common observation up to 10 km, plus 6 minutes per km beyond. Baselines over 20 km are not processed by fast-static.
• Loops are recommended, not required. A mark on a closed loop is checked by it. A mark joined by only one baseline — straight to a CORS, the hub, or a mark on a closed loop — is still solved, as a hanging mark: nothing checks that baseline, so it must fix its ambiguities, and the report flags the mark NO LOOP CHECK. A mark hanging off another hanging mark is left out. Observe a mark against a second station for a checked coordinate.
• Record at 1–15 s (at most 30 s), dual-frequency GPS, with a calibrated antenna. Night sessions are markedly better than midday ones in Sabah: the daytime ionosphere is the dominant error over long baselines.