Why GNSS antenna design is harder than it looks
Getting a position fix seems simple — the satellite signals are there, the receiver processes them, and coordinates appear. But the signals arriving at the antenna are extraordinarily weak: roughly -130 dBm at the Earth’s surface, which is below the thermal noise floor. The GNSS antenna’s job is to capture as much of that faint signal as possible while rejecting interference from every other direction. A poor antenna choice can add metres of positioning error, or prevent the receiver from acquiring satellites altogether.
Unlike cellular or Wi-Fi antennas, GNSS antennas must handle signals from four different constellations — each with its own frequency bands, modulation schemes, and orbital geometries. The design challenge is not just sensitivity but multi-band compatibility, phase centre stability, and multipath rejection.
The four constellations and their frequency requirements
Modern GNSS receivers are multi-constellation, meaning they track satellites from multiple systems simultaneously to improve accuracy, availability, and time-to-first-fix. Each constellation operates on different frequencies, and the antenna must cover all of them.
GPS (USA) operates on L1 (1575.42 MHz), L2 (1227.60 MHz), and L5 (1176.45 MHz). L1 carries the civilian C/A code that every receiver uses. L2 and L5 provide additional accuracy and ionospheric correction for dual- and tri-frequency receivers.
GLONASS (Russia) uses FDMA (frequency-division multiple access), with each satellite transmitting on a slightly different frequency around L1 (1602 MHz) and L2 (1246 MHz). The antenna must cover a wider bandwidth — roughly 1593-1610 MHz for L1 — to capture all channels simultaneously.
BeiDou (China) transmits on B1 (1561.098 MHz), B2 (1207.14 MHz), and B3 (1268.52 MHz). The B1 frequency is close to GPS L1 and GLONASS L1, so a well-designed antenna can cover all three with a single radiating element. B2 and B3 require additional bandwidth.
Galileo (Europe) uses E1 (1575.42 MHz — same centre frequency as GPS L1), E5a (1176.45 MHz — same as GPS L5), and E5b (1207.14 MHz). The alignment with GPS frequencies simplifies antenna design: a dual-band antenna covering L1/L5 can simultaneously receive GPS, Galileo E1/E5a, and BeiDou B1/B2.
A practical antenna specification for full multi-constellation support needs to cover at minimum 1561-1610 MHz (L1/E1/B1/GLONASS L1) and 1164-1300 MHz (L5/E5a/B2/B3/GLONASS L2). Many commercial “multi-band” GNSS antennas achieve this with a single dual-band patch element.
Patch antennas: the workhorse of GNSS
The vast majority of GNSS receivers use ceramic patch antennas — typically 25×25 mm or 18×18 mm squares of high-dielectric-constant ceramic mounted over a ground plane. They are compact, low-profile (4-8 mm thick), and can be manufactured to cover the required frequencies with good efficiency.
Ceramic material matters. High-permittivity ceramics (εr ≈ 25-50) allow the antenna to be physically small relative to wavelength. The trade-off is narrower bandwidth: a high-εr patch may only cover 20-30 MHz, which is sufficient for single-band L1 but may not span all GLONASS channels. Lower-permittivity ceramics (εr ≈ 10-15) offer wider bandwidth but require larger patch dimensions.
Ground plane is essential. A patch antenna needs a ground plane of at least 70×70 mm to achieve its specified gain and pattern. Smaller ground planes — common in handheld devices — reduce gain by 2-4 dB and distort the upper hemisphere pattern, which is exactly where GNSS satellites appear. If space is limited, use a passive element or decoupling structure to create an “electrical ground plane” larger than the physical one.
Active LNA is almost always integrated with the patch antenna. The LNA (low-noise amplifier) is mounted directly behind the patch, connected by a short microstrip line, and provides 20-30 dB of gain with a noise figure of 1-2 dB. This amplification is critical: without it, the cable loss between the antenna and receiver would degrade the signal-to-noise ratio below usable levels. Always power the LNA at the voltage specified by the manufacturer — overvoltage increases noise figure, undervoltage reduces gain.
Phase centre stability: the hidden accuracy factor
GNSS positioning works by measuring the time of flight from satellite to receiver. For high-accuracy applications (RTK, PPP), the measurement is made to the antenna’s phase centre — the apparent point from which the signal radiates. If the phase centre moves with satellite position, temperature, or nearby objects, the measurement error changes accordingly.
Phase centre offset (PCO) is the physical displacement between the antenna’s mechanical reference point and its electrical phase centre. It is typically 1-2 mm for L1 and is characterised by the manufacturer. High-precision GNSS antennas provide calibration data that allows the receiver to compensate for PCO.
Phase centre variation (PCV) is the change in phase centre position as a function of satellite elevation angle. A good geodetic antenna has PCV below 2 mm across all elevation angles above 10°. Cheap patch antennas can have PCV of 10-20 mm, which introduces centimetre-level errors in RTK positioning.
Multipath rejection is achieved through choke rings, ground planes, or specialised antenna elements that attenuate signals arriving from below the horizon. Reflected signals from buildings, vehicles, or the ground itself can be as strong as the direct signal and cause the receiver to compute a biased position. For survey-grade accuracy, use a geodetic antenna with a choke ring or multipath mitigation ground plane. For consumer applications, ensure the antenna is mounted with a clear view of the sky and away from reflective surfaces.
Single-band versus multi-band: when to pay for complexity
Single-band L1 antennas are sufficient for navigation applications where 3-5 metre accuracy is acceptable — fleet tracking, asset monitoring, personal locators. They are cheap (under $5 in volume), compact, and power-efficient.
Dual-band L1/L5 antennas enable ionospheric correction, which removes the largest single source of error in GNSS positioning (2-10 metres depending on solar activity). L5 also offers higher chipping rate, which improves multipath resistance. Dual-band antennas are increasingly common in smartphones and are essential for precision agriculture, drone navigation, and autonomous vehicles.
Multi-band (L1/L2/L5) antennas are required for survey-grade RTK and PPP receivers that need centimetre-level accuracy. The additional frequency allows the receiver to resolve integer ambiguities more quickly and reliably, reducing convergence time from minutes to seconds. The antenna cost premium is typically 3-5× over a single-band design, justified only when accuracy demands it.
GNSS antenna selection checklist
- Constellation coverage: confirm the antenna covers all frequencies your receiver will use (L1 alone, L1/L5, or full L1/L2/L5)
- Ground plane size: at least 70×70 mm for specified performance; larger is better
- Active LNA included: verify gain (25-30 dB) and noise figure (<2 dB); ensure power supply matches receiver
- Connector and cable: SMA is standard; keep cable as short as possible — every metre at L1 costs 0.7 dB
- Mounting: magnetic for temporary, bolt-through for permanent; ensure waterproofing (IP67 minimum)
- Phase centre: required for RTK/PPP; not relevant for navigation-grade applications
- Multipath environment: urban canyon or near reflective surfaces? Consider a choke ring or geodetic antenna
- Temperature stability: ceramic patches drift with temperature; check frequency shift specification (±1 MHz over -40°C to +85°C is typical)
The GNSS antenna is the sensor that defines the ceiling of your positioning system. No amount of receiver sophistication can overcome a poor antenna. Match the antenna to your accuracy requirement, your constellation plan, and your physical mounting constraints — and the rest of the system will perform as designed.
Leave a Reply