RINEX Logging and the PPK Backup Workflow

Log raw observations at the base while flying RTK — and rescue any mission in post-processing.

The insurance policy that weighs nothing

RINEX — the Receiver Independent Exchange format — is GNSS's universal raw-observation language: every epoch's carrier-phase and code measurements, satellite by satellite, written to a file any processing software can read decades later. Logging it at the base while you fly RTK costs one toggle and a few megabytes; owning it means any mission the live link betrayed can be reconstructed to full accuracy afterwards. Crews that log religiously describe the habit the same way: it is the cheapest insurance in surveying, and the premium is zero.

The claim events are mundane and universal: radio shadow behind a stockpile for ninety seconds, a cellular handover that spiked correction age, an NTRIP credential that expired mid-mission. RTK writes decimetre FLOAT geotags through every such gap — silently. The base's RINEX log is the difference between a reflight and a coffee-break reprocess.

PPK in one paragraph

Post-Processed Kinematic replays the mission on a desk. Two raw logs — the base's RINEX and the rover's (a drone's onboard log, or the flight controller's) — are aligned in software that resolves ambiguities forward and backward through time, producing corrected positions for every camera event. Backward smoothing is PPK's quiet superpower: an ambiguity resolved at minute five constrains minute four retroactively, so gaps that crippled the live solution often process clean. Accuracy matches RTK's centimetre class; the trade is latency (results after landing) and one office step (the full method comparison lives in PPP vs RTK vs PPK).

Setting up base logging correctly

Three settings decide whether a log is useful. Rate: 1 Hz is the field standard — drone events interpolate cleanly between epochs; 5 Hz buys margin for fast, low flights at 5× the file size. Format and version: RINEX 3.x with multi-constellation observations (GPS, GLONASS, Galileo, BeiDou) so the processor sees everything the receiver tracked; vendor-raw formats (like UBX) convert fine but native RINEX skips a step. Session hygiene: start logging before takeoff, stop after landing, and let files carry timestamps in their names. Storage maths are trivial — 1 Hz multi-constellation runs a few MB per hour, so an 8 GB base stores weeks — and pulling files over Wi-Fi at day's end takes minutes. UAV Mate's console exposes exactly this: a logging toggle beside the base's convergence status, so the habit costs one glance (where it sits in the M350 workflow).

The base position: PPK's inherited truth

PPK positions the rover relative to the base with centimetre precision — and then plants the whole result wherever the base coordinates say. A base averaged to half a metre yields PPK output half a metre wrong, immaculately. This is the same absolute-versus-relative trap that haunts RTK (dissected here), and it is why the logging habit pairs naturally with a self-converging base: the PPP-established coordinates (~1.5 cm, ITRF2020) written into the RINEX header make every downstream PPK run absolute automatically. Crews using averaged bases have a workaround — submit the same log to OPUS, AUSPOS, or CSRS-PPP, receive authoritative coordinates next day, and re-run PPK against them — which works, at the price of a day's latency the PPP base never charges.

Software: the free tier is genuinely good

Emlid Studio processes drone PPK free — base RINEX plus rover log plus camera events in, geotagged image list out, with support for DJI, Autel, and generic logs; it made PPK mainstream almost single-handedly. DJI Terra performs one-click PPK for the Matrice 4E's logs inside the mapping pipeline. RTKLIB (and its friendlier forks) remains the open-source workhorse for arbitrary receivers and forensic jobs. Up the ladder, the photogrammetry suites and survey-office packages fold PPK into their project flows. The practical guidance: match the tool to the rover's log format first — the base's standard RINEX is welcome everywhere — and keep one operator fluent in a second tool as the tiebreaker when a stubborn dataset argues.

The hybrid habit: fly RTK, keep PPK loaded

Professionals do not choose between RTK and PPK; they stack them. Fly on RTK for live FIX confidence, terrain-follow, and instant QA; log RINEX at base and aircraft throughout; post-process only what needs it — a FLOAT leg behind the highwall, a mission where correction age spiked, or a client demanding an independently reproducible trail. The cost is a toggle and a few megabytes; the payoff arrives the week a link fails over unrepeatable conditions — the demolition mid-blast, the flood peak, the harvest morning. Some operations go further and reprocess one mission per month as an audit: RTK geotags versus PPK positions versus checkpoints, three independent legs holding up one accuracy claim (the checkpoint discipline).

A rescue, step by step

The anatomy of a save: a corridor mission drops to FLOAT for two minutes in radio shadow. Back at the office — pull the base RINEX and the drone's raw log; load both into Emlid Studio or DJI Terra with the camera-event file; confirm the base coordinates in the header (PPP-converged, so they stand); process. The software resolves through the gap using both time directions, flags residual quality per event, and writes a corrected geotag list; swap it into the photogrammetry project and rebuild. Elapsed desk time: fifteen minutes. The alternative — remobilising a crew two hours up a forestry road — is the number to weigh the toggle against.

Logging beyond rescue: the uses that accumulate

A RINEX archive quietly becomes infrastructure. Static submissions: any base log doubles as an OPUS/AUSPOS check on the PPP coordinates — a free, independent audit trail regulators love. Boundary-of-record: years later, a disputed as-built can be reprocessed from raw observations with modern software, something no geotag list allows. Multi-rover forensics: when two instruments disagree, their logs against the common base settle it arithmetically. Even receiver health shows up — cycle-slip patterns in the logs flag a failing antenna cable before it fails a job. None of these justify the habit alone; together they are why mature operations treat raw logging as default-on and deletion as the decision requiring a reason.

One-line takeaway

Log RINEX at the base on every mission, anchor it with absolute base coordinates, and PPK stops being a workflow choice — it becomes a standing safety net that costs a toggle and pays in saved reflights, audit trails, and arguments you never have to have.

Frequently asked questions

Do I need PPK if my RTK worked perfectly?

Then the logs simply archive. You need PPK the day RTK doesn't work perfectly — which is unannounced. Logging is the option premium; it costs almost nothing.

What files does a drone PPK run need?

Three: base RINEX, the aircraft's raw GNSS log, and the camera-event (timestamp) file. Flight apps and processing suites pair them automatically.

How long can the base log for?

At 1 Hz multi-constellation, a few MB per hour — an 8 GB store holds weeks of sessions. Rotate files to the office server as routine housekeeping.

Is PPK accuracy better than RTK?

Comparable at short baselines; PPK can edge ahead through gaps and long baselines thanks to backward smoothing. The decisive difference is operational robustness, not the centimetre count.

Can I PPK against someone else's base?

Yes — any RINEX source with trustworthy coordinates works, including CORS archives. Baseline length and data rate govern quality; your own on-site base at 1 Hz is the gold standard.