How to Choose a GNSS Receiver: Accuracy, Compatibility & Corrections

What Makes a GNSS Receiver "RTK-Capable"?

Before getting into how to choose a GNSS receiver for your project, you need to know what separates an RTK-capable receiver from a regular GPS module. Not every GPS module can deliver centimeter accuracy. The term "RTK-capable" means the receiver meets three technical requirements:

1. Carrier-phase tracking. Standard GPS receivers measure the travel time of satellite signals (code-based positioning). RTK receivers go deeper: they track the phase of the carrier wave itself. This is what makes millimeter-level measurement resolution possible, which the RTK algorithm then refines into centimeter-level positions.

2. RTCM 3.x correction input. The receiver must accept correction data from an external source. The industry standard format is RTCM 3.2 (or 3.3), streamed in real time. If a receiver cannot parse RTCM messages, it cannot perform network RTK, regardless of how many frequencies it tracks.

3. An RTK engine (on-board or external). The correction data and raw observations need processing. Most modern receivers include an on-board RTK engine that computes the corrected position internally. Some lower-cost modules output raw observations for processing on an external computer or flight controller.

There is an important distinction between three tiers of GNSS receivers:

The jump from meter-level to centimeter-level is not incremental. It requires fundamentally different hardware and a real-time correction source.

GNSS Accuracy Tiers: From Meter-Level to Centimeter-Level

Receiver accuracy depends on two factors: the hardware's frequency capability and the quality of corrections applied. Here is what each class actually delivers in practice.

Why dual-frequency is the practical minimum for RTK. Single-frequency receivers can achieve centimeter accuracy, but only under ideal conditions: clear sky, short baseline to the base station, and no signal interruptions. The moment conditions degrade (urban canyons, tree canopy, longer baselines), single-frequency RTK struggles to resolve the integer ambiguity and loses its fix. Dual-frequency receivers resolve ambiguities faster and maintain fix more reliably because the second frequency allows the receiver to model ionospheric delay directly.

The introduction of the u-blox ZED-F9P in 2019 was a turning point. It brought dual-frequency, multi-constellation RTK to a module price under $200, making centimeter accuracy accessible outside the traditional surveying market (u-blox ZED-F9P Product Summary). Before the F9P, comparable capability required modules costing $2,000 or more.

Adding a third frequency (L5) further improves performance, particularly for initialization speed and robustness against interference. However, for most field applications, dual-frequency delivers the accuracy needed at a fraction of the cost.

The 5 GNSS Receiver Specs That Actually Matter

Receiver datasheets list dozens of specifications. When figuring out how to choose a GNSS receiver, these are the ones that directly affect your RTK performance and workflow.

Constellation Support

Modern GNSS means more than GPS. Four global constellations are operational: GPS (US), GLONASS (Russia), Galileo (EU), and BeiDou (China). A receiver that tracks all four constellations sees 2-3 times more satellites than a GPS-only device. According to the European Space Agency's Galileo performance reports, multi-constellation tracking improves position availability by up to 30% in obstructed environments such as urban areas or forest canopy (ESA Constellation Information).

Minimum recommendation: GPS + Galileo + GLONASS. Preferred: all four plus SBAS (WAAS/EGNOS).

Update Rate

How often the receiver outputs a position. Common options:

• 1 Hz (1 update/sec): Sufficient for static surveying, agriculture auto-steer, and most mapping.
• 5-10 Hz: Recommended for drone mapping and moving vehicles.
• 20 Hz: Required for high-speed applications like autonomous vehicle testing or racing (see the PWR Racing Team's RT16e for a real-world example at 20 Hz).

RTCM 3.x and NTRIP Support

This is the compatibility checkpoint. If you plan to use any network correction service (rather than your own physical base station), confirm two things:

1. The receiver accepts RTCM 3.2 or 3.3 correction messages as input.
2. The receiver (or its connected device) can establish an NTRIP client connection to stream corrections over the internet.

Many integrated GNSS rovers handle NTRIP natively via built-in cellular modems. Board-level modules like the u-blox F9P require an external device (phone, tablet, or single-board computer) to handle the NTRIP connection. Both approaches work. For a deeper explanation of the protocol, see our NTRIP documentation.

Communication Interfaces

Consider how the receiver connects to your workflow:

• Serial (UART/RS232): Standard for embedded and machine-control applications.
• USB: Common for laptop-based field work.
• Bluetooth: Convenient for tablet/phone-based surveying with apps like SW Maps or FieldGenius.
• Ethernet/Cellular: Found on higher-end rovers for direct NTRIP streaming without a phone.

Antenna: Built-in vs. External

This specification is often overlooked, but antenna quality has a direct effect on RTK performance. Integrated antennas on compact boards are convenient for prototyping. For field work requiring reliable fix in all conditions, an external survey-grade antenna with a proper ground plane is strongly recommended.

Which RTK Correction Service Should You Use? Cloud NTRIP vs. Base Station

Here is a fact that receiver marketing rarely emphasizes: deciding how to choose a GNSS receiver is only half the picture. A $20,000 survey-grade receiver without corrections delivers the same meter-level accuracy as a $50 consumer GPS module. The receiver determines your accuracy ceiling. The correction source determines whether you reach it.

Four correction methods exist. Three of them are edge-case answers. One is the default choice for the overwhelming majority of teams.

Why Cloud NTRIP is the default for most teams

The math is straightforward when you compare Cloud NTRIP against the obvious alternative, your own base station, over three years of normal operation:

• Upfront cost: $0 vs. $5,000-$15,000 (receiver, antenna, radio, surveyed reference pillar).
• Time to first fix on a new site: minutes vs. a full day of hardware setup and a static survey for the reference point.
• Site flexibility: works the same in Munich, Lagos, or rural Texas. A base station works only within ~30 km of where it was set up.
• Operational overhead: zero. Your team carries no extra hardware, no batteries, no LTE modem for the base, no insurance against theft on remote job sites.
• Accuracy: identical when you are inside the cloud provider's coverage area. Both deliver 1-2 cm horizontal on a properly configured RTCM 3.2 stream.

This is why RTKdata and similar cloud NTRIP providers have become the standard for surveying firms, drone service operators, agritech companies, and robotics teams that work across multiple sites. RTKdata's network covers 140+ countries through 20,000+ reference stations, with sub-second latency on any NTRIP-compatible receiver. Subscriptions start at $32 per month and include a 30-day free trial so you can verify performance at your actual work site before committing.

When the other methods are the right answer

Cloud NTRIP is not the answer in every scenario. The honest edge cases:

• Own base station wins when you operate at one fixed site, day in and day out, with no need to move (large mining operations, fixed agricultural research stations, dedicated test tracks). The amortized cost over years can drop below subscription pricing.
• PPP is the right choice for long-baseline geodesy, offshore work, or any application where you have no internet connectivity at all but can wait minutes for convergence.
• PPP-RTK is a niche compromise. If you need global coverage, cannot tolerate the convergence time of pure PPP, and cannot rely on terrestrial RTK network density, it sits in the middle.

If your team moves between sites, ships products with embedded GNSS, or runs operations across more than one country, Cloud NTRIP is the simpler answer.

Which GNSS Receiver Fits Your Use Case? A Quick Decision Matrix

Theory is useful. Practical decision-making is better. Here is how to choose a GNSS receiver class and correction method for the most common applications.

Real-world examples

When ADVOLARE SkyView needed centimeter-accurate drone surveying in the Swiss Alps, receiver choice was only part of the equation. The alpine terrain made base station placement impractical, so they paired their dual-frequency drone receiver with cloud-based RTK corrections from RTKdata to maintain fix even in narrow valleys.

In a very different environment, AeroAg Drone Services in South Dakota switched their agricultural drone mapping workflow from PPK post-processing to real-time NTRIP corrections. The result: a 10-minute setup reduction per flight mission, with rock and weed detection maps delivered same-day instead of next-day.

Both cases illustrate the same principle: the correction source shapes your workflow as much as the receiver hardware does.

5 Common Mistakes When Buying a GNSS Receiver

After supporting thousands of receiver setups across our correction network, certain patterns emerge. These are the mistakes we see most often when teams figure out how to choose a GNSS receiver for their workflow.

1. Buying a single-frequency receiver for professional RTK work.
Single-frequency modules cost less, but the cost savings evaporate when you spend extra time waiting for fix, re-initializing after signal loss, or repeating measurements that did not converge. For any application where accuracy and uptime matter, dual-frequency is the baseline.

2. Assuming every receiver supports NTRIP natively.
Board-level modules (like the u-blox F9P) accept RTCM corrections, but they do not connect to the internet on their own. You need an NTRIP client running on a connected device: a phone, tablet, Raspberry Pi, or the receiver's companion app. Before purchasing, verify whether the receiver has a built-in NTRIP client or whether you need to provide one.

3. Ignoring antenna quality.
A dual-frequency receiver connected to a poor antenna will underperform a well-designed single-frequency setup with a quality antenna. The antenna is the sensor. Invest in a multi-band antenna with a ground plane for any field application. Patch antennas on development boards are fine for bench testing, not for production work.

4. Overlooking correction costs in total cost of ownership.
A receiver module at $200 looks inexpensive. Add a $12,000 base station and the picture changes. Alternatively, a cloud NTRIP subscription at $40/month over three years totals $1,440: a fraction of base station hardware. Calculate total cost of ownership for your expected usage period before deciding on a correction strategy.

"The question we get asked most often is: 'Will my receiver work with your service?' In almost every case, the answer is yes, as long as it accepts RTCM 3.x corrections via NTRIP. That is the universal compatibility layer. The receiver brand does not matter."
— Konstantin Nidens, CEO & Co-Founder, RTKdata

5. Locking into a proprietary ecosystem.
Some manufacturers bundle receivers with their own correction service and make it difficult (or impossible) to use third-party corrections. Before purchasing, confirm that the receiver allows you to configure a custom NTRIP caster address. Hardware-agnostic receivers give you the freedom to switch correction providers or use multiple sources.

GNSS Receiver FAQ: Accuracy, Cost, Compatibility

Your GNSS Receiver Decision Checklist

Before purchasing, work through this checklist on how to choose a GNSS receiver that fits your operating reality:

• Accuracy requirement defined? Know whether you need meter, decimeter, or centimeter accuracy. This determines receiver class.
• Dual-frequency minimum? If centimeter accuracy is the goal, confirm L1/L2 support.
• Multi-constellation? GPS + Galileo + GLONASS at minimum. All four constellations preferred.
• RTCM 3.x input confirmed? Check the datasheet. No RTCM input means no network RTK.
• NTRIP client available? Either built into the receiver or provided via companion app/device.
• Correction source decided? Own base station or cloud NTRIP. Factor the cost into TCO.
• Antenna quality matched to application? External antenna for field work, patch antenna only for prototyping.
• Update rate sufficient? 1 Hz for static work, 5-10 Hz for drones and vehicles, 20 Hz for high-speed applications.
• Not locked into a proprietary ecosystem? Confirm you can configure custom NTRIP server settings.