One Base for Drones and Auto-Steer Tractors

Run UAV mapping and RTK auto-steer from the same correction source — season after season, field after field.

The farm's accuracy problem is a plumbing problem

Modern operations need centimetre positioning in three places at once: the auto-steer tractor holding AB lines season after season, the sprayer or spreader following prescription maps, and the drone flying NDVI or elevation surveys that generate those maps. Each traditionally arrives with its own correction plumbing — a network RTK subscription per terminal, a dealer-installed base on the shed, an averaged position under the drone base — and the datums quietly diverge. The symptom every agronomist recognizes: this year's guidance lines offset a hand's width from last year's, drainage plans that don't quite sit on the paddock, prescription zones smeared at the boundaries.

Root cause, almost always: multiple correction sources, none of them absolute. The fix is architectural, not agronomic — one correction source, absolutely positioned, feeding everything.

Coverage economics in rural reality

Network RTK was designed around infrastructure density that farmland rarely has. Baselines stretch past the 30–40 km comfort zone, cellular coverage thins exactly where paddocks are largest, and the per-terminal subscription bill scales with the machinery list: tractor, sprayer, harvester, drone — each a few hundred to a couple of thousand per year, forever. Dealer-installed fixed bases solve coverage for one farm but are usually positioned by averaging (inheriting the repeatability trap) and speak one brand's radio dialect.

A self-converging PPP base inverts the economics: one hardware purchase, one correction-service subscription, unlimited consumers over standard RTCM 3.x. It converges to ~1.5 cm in ITRF2020 anywhere on the property in about three minutes — corrections arrive by L-Band satellite, so the farm's cellular situation is irrelevant (how satellite delivery works).

One base, every machine: how the pieces connect

The broadcast reaches each consumer over whichever transport suits it. Auto-steer terminals — John Deere, Case IH, Trimble- or Topcon-equipped, or open systems like AgOpenGPS — take RTCM over the UHF radio across a ~5 km working radius, or over NTRIP where the machine has connectivity; the base's radio speaks the common over-air protocols (TrimTalk and transparent modes) that terminals expect. The mapping drone — a Mavic 3M for multispectral, a Matrice with LiDAR for drainage — consumes the same stream via Custom Network RTK in DJI Pilot 2 or over the RC link (setup walkthrough). A rover for checking tile drains or staking fence lines joins with five NTRIP fields. Every position on the farm now shares one absolute datum.

Season over season: the repeatability dividend

Agriculture is longitudinal by definition — the whole point of AB lines, controlled traffic farming, and multi-year NDVI trends is comparison across time. An absolutely-positioned base makes the comparisons legitimate automatically: this season's guidance lines are last season's to the centimetre, wheel tracks stay on established traffic lanes (compaction stays where it belongs), variable-rate maps from 2025 drone flights align under 2026 yield data without rubber-sheeting. Move the base between farms or paddocks freely; it re-converges to the same global frame wherever it sits, so “where the base was” disappears as a variable from the farm's data history.

The agronomy stack benefits downstream: when drone orthomosaics, soil-sampling points, yield maps, and as-applied records genuinely share a frame, zone analytics stop being smoothed guesswork at the edges.

A season's rhythm with the workflow

Pre-plant: base at the shed or paddock edge, converge, broadcast; tractors take AB lines from it; the drone flies bare-soil elevation for drainage and zone design. In-season: sprayer follows prescriptions on the same datum; scouting flights (multispectral, thermal) geotag against the same stream; problem areas staked with the rover match the imagery exactly. Harvest: yield data lands on the shared frame, closing the loop for next year's prescriptions. Winter: the base logs RINEX for any survey-grade jobs — laser-levelled bays, dam as-builts — and the same unit doubles for the contractor's machine control. One device's duty cycle, effectively year-round.

Choosing kit if you're starting now

If the farm already runs a dealer RTK network with acceptable coverage and the machinery list is one tractor, the incumbent path is fine. The moment the list grows — a drone program, a second terminal, paddocks beyond coverage, or a contractor's mixed-brand fleet — the standards-first base wins: RTCM 3.x out, radio + NTRIP + RC-link transports, absolute self-positioning, no per-consumer fees. Evaluation is cheap: power the candidate in the worst-covered paddock and time to FIX on the oldest terminal you own. The base that passes that test will carry the operation for a decade (broader selection criteria in the buyer's guide).

Retrofit reality: older tractors and mixed fleets

The economics sharpen further down the machinery list. Auto-steer retrofits — from dealer kits to open-source AgOpenGPS builds — turn on the same question: where do corrections come from? A farm base broadcasting standard RTCM over UHF answers it for every retrofit at once, including the twenty-year-old spraying tractor that will never justify its own network subscription. Contractors benefit doubly: arriving at a client farm, their machines tune to the resident base's radio or take the NTRIP credentials, and every pass lands on the farm's own datum — hand-back maps that actually overlay the owner's records. The correction layer, once absolute and standard, stops caring how old or mixed the iron is; that indifference is precisely what farm fleets need.

Measuring the payoff: three numbers to track

Skip the brochure ROI and track three farm-specific numbers for one season. Overlap percentage from as-applied maps: centimetre repeatable guidance typically pulls pass overlap from 5–10% (metre-class or drifting corrections) to 1–2% — direct chemical, fuel, and time savings across every operation. Re-work hours: passes repeated because guidance lines shifted or coverage dropped mid-field. And subscription line items retired: terminals moved off per-machine network fees onto the base. Most operations find any one of the three carries the hardware cost; together they make the correction architecture one of the rare farm investments that pays in the first season and keeps paying identically every season after.

One-line takeaway

One absolutely-positioned base broadcasting standard RTCM turns guidance, spraying, mapping, and harvest data into a single repeatable datum — per-machine subscriptions and drifting AB lines are architecture problems, and this is the architecture that removes them.

Further reading

Base selection criteria sized for farm fleets are in the buyer's guide; the satellite-delivery mechanics that make paddock coverage irrelevant are in L-Band PPP explained.

Further reading

The RC-link and radio mechanics that carry corrections across paddocks are compared in the transport guide.

Frequently asked questions

Will it work with my John Deere / Case / Trimble auto-steer?

If the terminal accepts RTCM 3.x over UHF radio or NTRIP — which covers essentially all modern auto-steer, including AgOpenGPS builds — yes, from the same broadcast the drone uses.

What accuracy does auto-steer actually need?

Pass-to-pass work tolerates 2–5 cm; controlled traffic and strip-till want the repeatable centimetre-class an absolute base provides year over year.

Do I need internet in the paddock?

No. Corrections reach the base by satellite and the machines by radio. NTRIP is a convenience where coverage exists, not a requirement.

Can one base cover multiple farms?

Yes — carry it. It re-converges in minutes at each property, and because coordinates are absolute in ITRF2020, every farm's data stays in one consistent frame.