Introduction
An RTK base station is the foundation of a differential GNSS system. Every rover connected to it inherits any errors in the base station's position, antenna setup, or signal quality. A poorly set up base station does not produce obvious errors at the rover - it produces consistent but systematically wrong positions that may not be discovered until results are compared to independent reference data. Getting the base station right is the most important step in any precision GNSS operation.
Antenna Placement: The Most Critical Decision
The base station antenna must have a clear, unobstructed view of the sky above the elevation mask (typically 10–15 degrees). This is non-negotiable. Even small obstructions that block a few satellites intermittently affect the quality of the corrections broadcast to rovers. The antenna should be placed:
- Away from buildings, walls, trees, and any structure that can block or reflect signals at satellite elevation angles
- As far as practical from vehicles, metal fencing, containers, or large flat surfaces that cause multipath
- On a stable, non-vibrating mount - a concrete pillar, heavy tripod, or permanent monument
- At a sufficient height above the ground to minimise ground-reflection multipath
For permanent base stations, a geodetic-grade choke ring antenna on a forced-centring pillar is the professional standard. For temporary field bases, a survey-grade patch antenna on a heavy tripod with a ground plane disc is acceptable, provided the site is chosen carefully.
Survey-In vs Known Position
Survey-In (Self-Survey)
Most modern GNSS receivers include a survey-in mode that averages single-point position measurements over a defined time period (typically 1–30 minutes) to estimate the base station coordinates autonomously. Survey-in is fast and requires no external data, but the resulting coordinates are only as accurate as the underlying single-point GNSS - typically 50 cm to 2 m. All rover positions will be offset from true ground truth by this initial error.
Survey-in is appropriate for relative positioning applications where only the relative positions of points within the project matter, not their absolute coordinates. For absolute positioning - connecting survey results to a national grid or delivering coordinates that will be used in GIS systems alongside other data - survey-in coordinates are generally insufficient.
Known Position (Control Point)
The highest accuracy is achieved by establishing the base on a known control point - a monumented geodetic mark whose coordinates are precisely known in the project datum. The base receiver is configured with these coordinates exactly, and the corrections it broadcasts encode the true atmospheric and clock errors at that point. If no physical control point is accessible, the base coordinates can be established by:
- A static GNSS session tied to a national CORS network (typically 1–4 hours, achieving 1–3 cm accuracy)
- PPP processing of a static session (requires 30–60 minutes observation time, achieves 2–5 cm accuracy post-convergence)
- Using VRS network coordinates if the system provides a reliable known-position output
Common RTK Troubleshooting Scenarios
| Symptom | Likely Cause | Diagnostic Step |
|---|---|---|
| Rover shows Float, not fixing | Insufficient satellites, long baseline, high multipath, or correction stream issue | Check satellite count (>6), PDOP (<4), correction age, and baseline distance |
| RTK Fix but positions are wrong by 10–50 cm | Wrong fix, base coordinate error, or wrong antenna height | Compare to independent control; verify base coordinate and antenna height entry |
| Corrections received but Fix not achieved | RTCM message type mismatch, incorrect mountpoint, or constellation mismatch | Verify base RTCM output matches rover's expected message types; check constellation settings |
| Correction stream dropping out | NTRIP connection unstable, cellular signal intermittent, or radio link blocked | Check data connection; monitor correction age; consider UHF radio backup |
| Position jumping intermittently | Cycle slips resetting ambiguities, NLOS satellite inclusion, or stale corrections | Review fix/float transitions in log; check individual satellite C/Nâ‚€ and residuals |
Verification and Quality Assurance
Before beginning any precision survey or operation, the base station setup should be verified by occupying at least one known control point with the rover and checking the reported position against the known coordinates. A discrepancy of more than 2–3 cm (horizontal) and 5 cm (vertical) in good conditions indicates a problem with the base setup, correction stream, or rover configuration that must be investigated before data collection begins.
During operations, monitoring the correction age, satellite count, PDOP, and fix status on both the base and rover provides early warning of developing problems. Many survey-grade receivers can be configured to log all this diagnostic information for post-session review - enabling forensic analysis of any anomalies discovered after the fact.
Vital Points
- Base antenna placement is the single most impactful setup decision - poor placement creates systematic biases in every rover measurement without triggering any quality alarms.
- Survey-in provides only single-point accuracy (50 cm to 2 m) - use a known control point or a static survey tie for absolute coordinate accuracy.
- Verify the base setup by checking a known control point with the rover before beginning data collection - do not rely on the receiver's Fix status alone.
- Common RTK problems (float not fixing, wrong fix, intermittent corrections) all have systematic diagnostic approaches - work through satellite geometry, correction stream, and configuration settings methodically.