GNSS Interference Is Rising. Is Your BVLOS Drone RF and Timing Chain Ready?

Why resilient flight depends on the complete RF and timing chain, not one ‘hero’ component.

Permission is not the same as readiness

The CAA's CAP3182 roadmap supports the UK objective of routine BVLOS operations by 2027, with operational scenarios extending beyond that date; CAP3269 describes how the supporting airspace architecture could evolve. In the US, the FAA's proposed Part 108 framework had not taken effect as of 8 September 2026: a final rule was reported under OIRA review, so the new framework was not yet operative and existing operating authorities continued to apply. That is where the physical signal chain matters.

A capable autopilot and compliant radios can still lose mission confidence if the RF and timing chain runs out of margin. GNSS arrives extremely weak: the GPS Standard Positioning Service Performance Standard specifies a minimum received L1 C/A-code signal power of −158.5 dBW under the conditions defined in IS-GPS-200. Put that receiver centimetres from telemetry, video and propulsion, and the integration problem is obvious.

Authentication helps, but does not replace RF engineering. Galileo OSNMA enables compatible receivers to authenticate Galileo navigation-message data. That adds an important check on data authenticity, but it does not make the RF signal immune to jamming, guarantee reception, or by itself authenticate the resulting position solution. A genuine signal still has to reach the receiver with usable C/N₀, while the receiver front end must retain sufficient dynamic range in the presence of blockers.

One airframe. Four engineering jobs.

Underneath those jobs sit four engineering budgets: wanted-signal/noise, blocker/linearity, timing/phase-noise and link/fallback. Treating these as explicit budgets makes resilience measurable and exposes conflicts early. Full calculations sit in the companion application note.

1. ACQUIRE: protect the signal before the receiver sees it

GNSS filtering is not just about low insertion loss: the designer must decide which bands must pass, what else transmits at the same time, and whether the first LNA stays linear in the real blocker environment. A filter before the LNA improves blocker immunity, but its insertion loss spends part of the receiver's weak-signal margin by raising the cascaded noise figure. Putting the LNA first reduces that penalty, but the LNA must then tolerate the coupled RF environment. There is no universal order.

Where TechPoint Golledge fits: GSRF-GN/GSDX-GN provide GNSS RF filtering and band separation; GS-FEM-GN combines gain and filtering in an integrated front end.

2. REJECT: your own drone may be the strongest interferer

A video or telemetry transmitter does not have to be faulty to desensitise GNSS; it only needs enough coupled power or an unfortunate intermodulation relationship, and the same applies to payload radios, NTN terminals and switching regulators. Record TX power, duty cycle and measured isolation for each emitter, then compare against filter and receiver blocking limits. Simply saying “the frequencies are different” is not enough.

Where TechPoint Golledge fits: GBAW-HB extends the chain into 2.4, 5 and 5.8 GHz high-band communications, where compact, high-selectivity BAW filtering supports video, mesh and payload-link coexistence. BAW is not automatically better than SAW; selection follows the application.

3. HOLD: surviving vibration is not performing under vibration

A high-G or shock-survival rating asks whether a device survives a mechanical event; low acceleration sensitivity asks how much its frequency and phase move while acceleration is present. NIST has documented how vibration-induced phase modulation can degrade oscillator performance in UAV and other dynamic-platform applications; in a multirotor aircraft, propulsion-related vibration can be a sustained in-flight condition rather than a one-off shock event.

GTXO-LG is the vibration-tolerant local reference feeding a clock/LO synthesis stage that produces the RX/TX oscillators and sampling clocks the radios use. It is not itself the final RF LO or sample clock, and does not replace GNSS positioning or create navigation holdover. Supporting clocking is separate: GXO-95 covers SDR, FPGA and high-speed digital clock trees, kept structurally apart from GTXO-LG.

4. LINK: design the fallback before the primary path disappears

BVLOS communications is not just a range number: it is route coverage, coexistence with GNSS and payload RF, fade margin, and a fallback plan for when the preferred link degrades. NTN or satellite connectivity can be one option, not a universal answer.

3GPP/ETSI satellite-access specifications define FR1 NTN operating bands including n255 and n256. Compact UAV terminals must evaluate the applicable uplink and downlink frequencies, antenna configuration, coupled transmitter power and front-end linearity alongside GNSS. Spectral proximity does not make interference inevitable, but frequency separation alone is not sufficient evidence of coexistence.

Where TechPoint Golledge fits: GSRF-NTN/GSDX-NTN provide NTN/satellite filtering and duplexing, while GBAW-HB covers high-band payload paths.

The test most teams miss: combine the stressors

Expensive problems often surface only after four separate tests have passed: GNSS reception on a clean bench, a filter on a network analyser, an oscillator on a vibration table and a radio in a range test. Then the real aircraft flies with transmitters active, motors at speed, GNSS degraded and link margin moving all at once.

A stronger sequence builds the budgets first, checks filter and linearity performance, surveys the powered platform spectrum and vibration behaviour, then combines everything into one mission-representative test. The stress levels and pass/fail criteria should come from the intended mission, the applicable regulatory basis and the aircraft safety assessment, not a generic checklist.

Explore the TechPoint Golledge UAV signal chain

TechPoint Golledge also supports broader frequency-control needs, including TCXO, VCXO, OCXO and differential clocking.