Tier Three: NTRIP/Cellular-Assisted Quick Establishment
Where even a weak cellular signal reaches a site, a different correction path both establishes a base faster and produces a more accurate result than HAS does, by leaning on a real, calibrated regional network instead of a global satellite broadcast.
Applying a network correction, a traditional real-time network, a statewide CORS network, or PointPerfect Flex, during base establishment typically converges within seconds and reaches centimeter-level accuracy, faster and more accurate than HAS whenever cellular connectivity allows it.
Why a Little Connectivity Beats No Connectivity Here
A network-based correction has access to a regional model of atmospheric and orbital error, built from real, nearby reference stations, rather than a single global broadcast intended to work adequately everywhere on Earth at once. That regional specificity is what lets it converge faster and land closer to the truth, and it only requires enough of a connection to carry a modest data stream, not a strong or reliable one.
The advantage here is not connectivity for its own sake. It is access to a regional, rather than global, model of error, which a network correction can supply and a satellite-broadcast service structurally cannot.
Speed and Accuracy Moving Together, for Once
Most of this companion's methods trade speed against accuracy in one direction or another. This one does not. A network-assisted quick establishment is both faster to converge and more accurate in its result than HAS-assisted establishment, whenever the connectivity it requires is actually available. The only real cost is that requirement itself, a signal HAS was specifically built to work without.
Choosing between HAS and a network correction for base establishment is not a choice between fast-but-less-accurate and slow-but-more-accurate. Where cellular connectivity allows it, the network correction wins on both counts at once, and HAS's advantage is limited entirely to sites where that connectivity does not exist.
Choosing Which Network Correction to Use
A traditional real-time network, often shortened to RTN, is a wide-area system of continuously operating reference stations serving any rover within its coverage, rather than a single dedicated base, whether run by a private company or a state agency. It and PointPerfect Flex, delivered over RTCM, are both valid, practical, widely used sources for this method. A free statewide CORS network, covered in the fourth companion of this series, is typically an RTN in its own right, just publicly funded rather than commercially sold. Worth knowing, though not worth treating as a ranking: a traditional RTN and PointPerfect Flex solve the establishment problem through genuinely different underlying architectures, differential correction against a dense network versus SSR against a sparse one, covered at length in the fourth companion of this series. That architectural difference is a technical detail, not a statement about which option sees more everyday use. It is also worth being specific about which PointPerfect tier is meant here. The PointPerfect brand covers more than Flex, including Global, a pure PPP service, and Live, a traditional network RTK service, and only Flex is actually built on PPP-RTK. Checking whether a free, public RTN already covers a project's location is worth doing before assuming a paid subscription is required.
A Concrete Example: Network Correction to the Base, Radio to the Rover
A genuinely fast, genuinely easy version of this whole method looks like this in practice, and it is worth naming a specific, working pairing rather than only speaking in generalities: SparkPNT hardware connected to PointPerfect Flex over RTCM, the exact combination already established as a natural fit in the fourth companion of this series, over whatever marginal cellular signal the site has. Let that connection converge to a centimeter-level coordinate in a matter of seconds, then broadcast that corrected position to the rover over a direct, local radio link, an inexpensive point-to-point radio built around the ESP-NOW protocol among them, also covered in the fourth companion, though its 2.4 GHz signal loses range quickly through foliage and obstructions in a way a lower-frequency UHF or LoRa radio would not. That marginal cellular connection only ever has to do one small job: hold together long enough for the base to converge. Once the base has its coordinate, the connection can drop entirely, and the base carries on broadcasting to the rover over its own local radio link, needing nothing from the cellular network again for the rest of the session. The rover, for its part, never needed a cellular connection of its own at any point.
Once the base has converged on a corrected coordinate, the rover's own connectivity needs drop to almost nothing. A direct local radio link is enough, which is exactly why this combination goes from nothing set up to a working RTK setup in under two minutes, convergence included, with very little infrastructure.
When Even That Marginal Signal Is Not There
A site with no cellular signal at all does not automatically fall back to HAS as the only option. A compact satellite internet terminal, or a vehicle's own built-in satellite communications system, both covered in the fourth companion of this series, can supply the same kind of connection an NTRIP-based correction needs, just delivered by satellite instead of a cell tower. The same rule applies here as it does with marginal cellular: that satellite connection only has to hold long enough for the base to converge, a matter of seconds, after which it can be switched off entirely while the base keeps broadcasting to the rover over its own local radio link. Where either satellite option is available, this method's speed and accuracy advantage over HAS still applies, even at a site with zero cellular coverage of its own.
Zero cellular signal at a site does not mean HAS is the only quick option left. Satellite internet, whether a portable terminal or a vehicle's own built-in system, can carry the same NTRIP-based correction this chapter describes, often faster and more accurately than HAS, even where no cell tower reaches at all.
A job site has only one bar of cellular signal, barely enough to load a webpage. Does that rule out a network-based quick establishment in favor of HAS? Not necessarily. NTRIP corrections require very little bandwidth, and a marginal signal is often entirely sufficient, even where it feels too weak to trust.
A network-based correction, applied during base establishment, converges faster and lands more accurately than HAS whenever any cellular connectivity is available, because it draws on a real, regional model of error rather than a single global broadcast. HAS remains the better choice specifically when that connectivity does not exist at all.
Both methods so far establish a coordinate in one continuous session. The next tier introduces a genuinely different technique: catching an error neither single session, however corrected, could ever reveal on its own.