1. What Is Unintended Electromagnetic Radiation?

Every piece of operating electronics radiates. Switching-mode power supplies chop DC at tens to hundreds of kilohertz, producing harmonic spurs that reach deep into the HF and VHF spectrum. Clock lines in digital processors couple into PCB traces and become inadvertent antennas. Current loops in unbalanced wiring harnesses re-radiate differential-mode switching noise. Reaction wheels, magnetic torquers, solar-array regulators, laser inter-satellite links, phased-array beamforming electronics — every subsystem inside a modern satellite contributes something.

Engineers call this electromagnetic interference (EMI). Radio astronomers, who measure its effect on the sky rather than on a neighbouring subsystem, prefer the term Unintended Electromagnetic Radiation (UEMR). The distinction matters because of how it is regulated: intentional emissions — the 10.7–12.7 GHz Ku-band Starlink downlinks, for example — are governed by the ITU Radio Regulations, coordinated through national administrations, and subject to hard power-flux-density limits in protected bands. UEMR is not covered by the ITU Radio Regulations at all. The spacecraft electromagnetic compatibility (EMC) standards that do apply — for example MIL-STD-461 and the ESA ECSS-E-ST-20-07C family — are designed to stop a satellite’s own subsystems interfering with each other or with the launch vehicle. They were never written to protect third-party radio services on the ground.

For a single satellite in geosynchronous orbit 36,000 km away, UEMR rarely mattered. For a constellation of ten thousand satellites orbiting 340–570 km above your head, the aggregate flux density is an entirely different problem. At any given moment a fixed station anywhere on Earth has dozens of Starlink satellites above the local horizon, each one moving through the antenna pattern and each one contributing its own broadband electromagnetic signature. The sky that boatanchor operators spent decades learning to listen to is being re-engineered in real time.

2. What the Radio Astronomers Have Measured

2.1   LOFAR, the Netherlands — Di Vruno et al. (2023)

Using the LOw Frequency ARray (LOFAR) radio telescope in the Netherlands, Federico Di Vruno (SKA Observatory), Benjamin Winkel (Max Planck Institute for Radio Astronomy) and their colleagues observed 68 first-generation Starlink v1.0 and v1.5 satellites across 110–188 MHz. Radiation correlated with satellite passes was detected from 47 of the 68 — a 69% detection rate. Spectral power-flux densities ran from 0.1 to 10 Jy for the broadband component and 10 to 500 Jy for narrowband features at 125, 135, 143.05, 150 and 175 MHz.

Translated to electric field strengths at the 10 m reference distance used in commercial EMC work, the measurements reached up to 49 dB μV m⁻¹ in a 120 kHz bandwidth. For context, the CISPR‑32 limit for class-B commercial electronics is 30 dB μV m⁻¹, and the ITU‑R RA.769-2 interference threshold for the 150.05–153 MHz primary radio astronomy band would require each satellite to stay below 11.7 dB μV m⁻¹. The first-generation Starlinks exceeded the radio astronomy threshold by 10–30 dB — that is, each satellite was radiating 10 to 1,000 times more UEMR than ITU limits say is tolerable.

2.2   LOFAR (Second Generation) — Bassa et al. (2024)

One year later, the same team led by Cees Bassa (ASTRON) turned LOFAR on the second-generation v2-Mini and v2-Mini Direct-to-Cell satellites across 10–88 MHz and 110–188 MHz. The results were alarming. Broadband emission covered 40–70 MHz and 110–188 MHz with spectral flux densities of 15 to 1,300 Jy between 56 and 66 MHz, and 2 to 100 Jy in 8-MHz bands centred on 120 and 161 MHz. After correcting for range, the v2-Mini satellites were found to emit UEMR up to 32× stronger than the first generation. Bassa noted publicly that this made the satellites “ten million times brighter than the weakest astrophysical sources” that the same instruments are designed to detect.

2.3   EDA2, Western Australia — ICRAR / Curtin (Grigg, Tingay & Sokolowski 2025)

Closer to home for Australian hams, the Engineering Development Array 2 (EDA2) — a prototype station for the SKA-Low telescope, located in the Inyarrimanha Ilgari Bundara / CSIRO Murchison Radio-astronomy Observatory in Western Australia — has produced the largest low-frequency survey of satellite emissions to date. PhD candidate Dylan Grigg, Prof. Steven Tingay and Dr Marcin Sokolowski analysed approximately 76 million full-sky images captured over 29 days of observing. The results, published in Astronomy & Astrophysics in July 2025, are sobering:

• 112,534 individual detections of 1,806 unique Starlink satellites — roughly 28% of every Starlink in orbit at the time of observing. This is by a wide margin the most comprehensive catalogue of low-frequency satellite radio emissions ever compiled.
• 76% of all v2-Mini Ku-band satellites and 71% of all v2-Mini Direct-to-Cell satellites were identified radiating UEMR — the second-generation hardware is being detected at rates approaching unity.
• In the worst-case datasets, up to 30% of sky images contained a detectable Starlink satellite — almost one image in three corrupted by a single constellation.
• 703 satellites were detected emitting between 150.05 and 153 MHz — inside the ITU primary radio astronomy band where all other emissions are prohibited. A further 13 satellites were identified between 73.00 and 74.60 MHz, another band protected for radio astronomy.
• Broadband Starlink UEMR detected from 70 MHz through 225 MHz, with no detections above ~320 MHz (consistent with reduced antenna sensitivity at the top of the band).
• Narrowband features at 137.05 MHz with a characteristic ~100-second pulsing signature — a distinctive “heartbeat” of onboard hardware.
• Terrestrial FM broadcast transmissions reflecting off Starlink satellites at 99.70 MHz were also detected — the spacecraft are now operating as unintended retroreflectors for ground-based emissions.
• Received UEMR flux density varies sinusoidally with satellite attitude, with orthogonal polarisations anti-correlated — a classical signature of a rotating field emanating from internal hardware rather than a deliberate transmission.

This matters directly to Australian operators because EDA2 shares a site with SKA-Low, the world’s most sensitive upcoming low-frequency instrument, built on the principle that the Murchison is one of the last genuinely radio-quiet places on the planet. Overflying satellites are — by the design of LEO — the one interference source that radio-quiet zones cannot shield against. Prof. Tingay’s blunt public assessment was that the impact will vary by science case “but has the potential to result in increasing data loss for projects.” The same physics that ruins a radio astronomer’s 29-day dataset will also raise the noise floor of the 6‑metre and 2‑metre operator tuning across the same sky.

2.4   NenuFAR, France — Below 100 MHz

The NenuFAR station in central France extended the observations below 100 MHz and reported broadband UEMR concentrated in the 54–66 MHz range at intensities above 500 Jy — again with significant polarisation and clear differences between v1 and v2 hardware. 54–66 MHz is, of course, the low-VHF television and weather-fax spectrum globally and, critically, the 50–54 MHz 6‑metre amateur allocation sits directly in the emission corridor.

3. Which Bands Are Affected?

The table below summarises the frequency ranges where Starlink UEMR has been directly measured in peer-reviewed work, mapped against amateur and broadcasting allocations. Note that peer-reviewed published measurements exist for VHF and very low VHF; confirmed in-band HF measurements of satellite UEMR are limited by the sensitivity of the instruments so far deployed to the problem, but the same switching-supply and digital-processor mechanisms that produce the VHF emissions are known to radiate throughout HF. Operators reporting an unexplained elevated noise floor at low-VHF and high-HF should document it rigorously — evidence gathered by the amateur community is going to matter.

Frequency Range	Impact	Services Affected
54–66 MHz	Broadband UEMR > 500 Jy (NenuFAR)	6 m amateur (50–54 MHz), legacy TV Ch.2–4, 60 MHz Es beacon window
70–88 MHz	Broadband UEMR (EDA2, LOFAR)	4 m amateur (UK/EU), 73.8‑74.6 MHz RAS band, FM broadcasting (US 88–108 Band II lower guard)
110–138 MHz	Broadband + narrowband (125, 135 MHz)	Aeronautical (118–137 MHz), weather satellites, radio astronomy 120 MHz line
143–146 MHz	Narrowband 143.05 MHz (GRAVES radar reflection component) + UEMR floor	2 m amateur (144–148 MHz), satellite telemetry
150.05–153 MHz	ITU-R RA.769-2 threshold exceeded by 10–30 dB per satellite	Primary radio astronomy band
159.4–188 MHz	Broadband + narrowband 161, 175 MHz	VHF marine, SKA-Low band 1 (50–350 MHz), DAB
Below 30 MHz (HF)	Theoretically expected from switching-supply harmonics and digital clocks. No peer-reviewed quantitative satellite-pass detections yet published — instruments of adequate sensitivity have not been looking.	All HF amateur bands, shortwave broadcasting, WSPR beacon networks, ionospheric research
Medium Wave (0.53–1.7 MHz)	Not directly measured from satellites. MW DX noise floor elevation reported anecdotally by DX-ers is more likely from terrestrial Starlink user terminals, adjacent Ethernet and switch-mode PSU gear than from the satellites themselves.	AM broadcasting, 160 m amateur, longwave nav beacons, NDBs

4. Why Mega-Constellations Change the Game

A single satellite radiating 49 dB μV m⁻¹ worth of UEMR 550 km overhead is a nuisance. A constellation of ten thousand satellites, with new launches every four to six days and permanent sky coverage, is a structural change to the background electromagnetic environment of planet Earth. Consider the scale:

• ~10,200 active Starlinks in orbit as of April 2026 (KeepTrack / McDowell telemetry).
• 11,700+ total launched since 2019 across 600+ missions; SpaceX continues to launch every three to six days.
• FCC January 2026 authorisation raised the approved ceiling to 15,000 satellites; SpaceX’s long-form filings target up to 42,000.
• 65% of all active satellites worldwide are Starlinks — SpaceX now operates more spacecraft than all other operators combined.
• Amazon Kuiper (3,200 planned), Eutelsat OneWeb (720 flying, 6,400 planned), Chinese Qianfan/G60 (~14,000 planned) and Guowang (~13,000 planned) are all building Gen-2 mega-constellations with similar underlying hardware philosophies. Starlink is the first and largest, but it is not the only one.
• Projections from the Satellite Industry Association put the global active-satellite count at 100,000 by 2030. Almost all will be in LEO. Almost all will radiate UEMR. And all of them will be overflying your station.

This is the aggregate interference problem that the ITU‑R spectrum regime was never designed for. Individual power-flux-density limits mean very little when the sky is filled with emitters. And the radio-quiet zone concept — the cornerstone of every major radio astronomy installation since Green Bank — cannot shield against something overhead.

5. What SpaceX Has Done — and Not Done

To SpaceX’s credit, the company has engaged meaningfully with the astronomy community. The 2019 NSF – SpaceX coordination agreement, extended in 2022, established a standing technical channel between SpaceX and the National Radio Astronomy Observatory (NRAO). Two concrete engineering outputs have emerged from that collaboration: the Operational Data Sharing (ODS) system, which feeds live telescope observing schedules to the Starlink fleet in real time, and the Starlink Telescope Boresight Avoidance algorithm, which uses that data to instruct individual satellites to redirect beams or mute electronics for a few seconds as they transit through a telescope’s main beam. SpaceX voluntarily leaves the lower two 240 MHz Ku-band downlink channels unoccupied to protect the 10.6–10.7 GHz radio astronomy band. It cooperated on the “DarkSat” and “VisorSat” programmes to reduce optical glare.

In June 2025 the collaboration widened when SpaceX and the SETI Institute announced a joint programme to reduce interference at the Allen Telescope Array (ATA) in Shasta County, California — the first and only observatory built specifically for the search for extraterrestrial intelligence. The ATA is now one of the first sites running the boresight-avoidance techniques operationally. Researchers at the ATA are also developing a concept called radio dynamic zones, a flexible spectrum-management idea aimed at letting scientific and commercial users coexist in the same bands through real-time coordination rather than rigid static allocation.

But the measured reality in 2024–2026 has been that each successive hardware generation has radiated more UEMR, not less — the 32× increase between v1 and v2-Mini is the most striking datum, and the direct-to-cell variants add further emissions in the 700 MHz and 1.9 GHz mobile bands. The mitigation agreements above all target intentional downlink emissions through ODS-style scheduling — they do nothing for UEMR, which the onboard hardware radiates continuously whether the satellite is above a telescope or not. And none of this voluntary engagement is binding on any other operator. Amazon Kuiper, Qianfan/G60, Guowang and OneWeb are under no obligation, regulatory or otherwise, to replicate SpaceX’s goodwill gestures.

6. How Amateur Radio Operators Can Voice Their Concerns

Regulatory outcomes follow the weight of the record. Below are the concrete avenues — national and international — where amateur operators, DX clubs, and RF engineers can add that weight. Treat this as a checklist; the more of these you engage with, the greater the chance that UEMR gets added to binding international regulation before the constellation count trebles again.

6.1   Australia — ACMA, WIA and the Murchison

The Australian Communications and Media Authority is the national spectrum regulator and is the correct first destination for any interference observation that you can document. ACMA administers the Radiocommunications Act 1992, including the foreign space object listing under which Starlink is authorised to operate in Australian bands.

• File an ACMA interference report — acma.gov.au/interference-radiocommunications-request-investigation-form. Be specific: frequency, date/time UTC, your location, the satellite NORAD ID if you can identify a specific pass (use Celestrak TLE data), the field strength if you have an S-meter reading, and a spectrogram capture if your SDR can produce one. ACMA has limited field enforcement capacity these days, but every logged complaint builds the permanent record that future rulemaking relies on.
• Write to the Wireless Institute of Australia (WIA) — wia.org.au, national office at Unit 20, 11‑13 Havelock Road, Bayswater VIC 3153, phone (03) 9729 0400, email [email protected]. The WIA is the IARU Region 3 national society and represents Australian amateurs before ACMA, the Department of Infrastructure / Communications, and in the CEPT/ITU process. Forward interference logs to the WIA Technical Advisory Committee (TAC), which collates member evidence and feeds it into the WIA’s contribution to the Australian delegation for WRC-27.
• Participate in ACMA public consultations — ACMA regularly releases discussion papers on spectrum management, class licensing, and foreign space-object authorisation. These consultation periods are the formal mechanism for the amateur community to remind the regulator that elevated aggregate noise floors from overhead constellations are an interference problem deserving rule-making attention, not a tolerance threshold to be quietly raised.
• Contact the Department of Infrastructure, Transport, Regional Development, Communications and the Arts — the Minister’s office accepts correspondence on spectrum policy matters, particularly concerning the Murchison Radio-astronomy Observatory and SKA-Low protection.
• Submit directly to SKA Observatory and CSIRO ATNF — both are actively building scientific UEMR data cases. Credible amateur observations with calibrated noise floors are useful contributions.

6.2   United States — FCC, NTIA and the ARRL

The Federal Communications Commission is SpaceX’s licensing authority and is therefore the most consequential venue in the world for this issue. The FCC’s Electronic Comment Filing System (ECFS) makes it straightforward for any citizen — including non-US operators — to submit comments to any open proceeding.

• File comments via ECFS — fcc.gov/ecfs. Relevant active dockets include the Space Bureau’s ongoing proceedings on non-geostationary satellite systems (search Docket 22-271, 25-201 and related), SpaceX Gen2 modifications, and the WRC-27 Advisory Committee Docket 24-30. Comments become part of the permanent record.
• Report interference via the FCC Consumer Inquiries and Complaints Center — consumercomplaints.fcc.gov — for specific documented interference events on licensed amateur allocations. Under FCC Part 15.5(c) even unintentional emitters must cease operation when notified they cause harmful interference to a licensed service.
• Engage with the ARRL — arrl.org/regulatory-rfi-information. The ARRL Regulatory Information Office and the Clean Signal Initiative have historically been effective in getting the FCC’s attention on aggregate RFI problems (BPL, grow-lights, LED ballasts). Member reports fuel the League’s docket filings.
• NTIA and Congressional offices — the National Telecommunications and Information Administration manages federal spectrum use and coordinates US positions for WRC. Constituent correspondence to House and Senate Commerce Committee offices creates political pressure that the FCC notices.
• The FCC WRC-27 Advisory Committee (Docket 24-30) develops US positions on the 2027 World Radiocommunication Conference agenda. Comments filed there directly influence what the US delegation argues at ITU.

6.3   International — ITU, IARU, CEPT and the IAU

UEMR from satellites is fundamentally an international problem because the satellite is not in any single administration’s territory. Several international venues are where binding rules could be written in the next decade.

• The International Amateur Radio Union — iaru.org — is the amateur service’s sole standing representative at ITU-R. Contact your regional member society (WIA in Region 3, ARRL in Region 2, RSGB in Region 1) with data and let them escalate.
• IARU Monitoring System (IARUMS) — iaru.org/spectrum/monitoring-system. The regional IARUMS teams maintain the formal intruder-watch process for non-amateur transmissions inside primary amateur allocations — which is relevant to the documented narrowband Starlink spurs that fall in amateur bands (for example, the 143.05 MHz feature near the 2‑metre allocation). Note that IARUMS’s terms of reference explicitly exclude RFI/EMC, so broadband UEMR is outside its remit — that half of the problem belongs with the IAU CPS and national regulators. Report narrowband spurious emissions to IARUMS; report broadband noise-floor elevation to ACMA, FCC, and the CPS.
• IAU Centre for the Protection of the Dark and Quiet Sky — cps.iau.org. The CPS SatHub collects citizen-scientist observations through the Satellite Constellation Observation Repository (SCORE). The CPS secured a five-year MoU extension to 2030 and has active working groups on policy, industry engagement, and technical measurement. Amateur observers are welcome; see the CE Hub guidance.
• World Radiocommunication Conference 2027 (WRC-27) — convening in Shanghai, China, from 18 October to 12 November 2027, with the Radiocommunication Assembly (RA-27) running 11–15 October beforehand. This is the first WRC ever hosted in Asia, and it will be the treaty venue where any new regulation of satellite UEMR either gets agreed or gets kicked down the road to WRC-31. The radio astronomy community has already placed UEMR-related items on the agenda. CEPT’s Conference Preparatory Group for WRC-27 (cept.org/ecc/groups/ecc/cpg/) is open to administration input — national administrations (ACMA, Ofcom, FCC) accept stakeholder submissions ahead of CPG sessions.
• UN Committee on the Peaceful Uses of Outer Space (COPUOS) — the IAU successfully got “dark and quiet skies” placed on the COPUOS agenda. National delegations to COPUOS include scientific observers; your national astronomy society can relay evidence.
• CISPR and IEC — the civilian EMC standards bodies could, in principle, extend CISPR‑32 style limits to spacecraft. Amateur and broadcast advocacy to national standards bodies matters here in the long run.

6.4   Building an Effective Observation Report

Regulators and standards bodies respond to data, not to opinion. An effective report from an amateur station looks like this:

1. Station identification: callsign, ITU zone, precise geographic coordinates, antenna type, feedline length, elevation above ground, receiver and preamp details.
2. Baseline noise floor characterisation: establish and document the pre-constellation noise floor in each band of interest. Historical spectrum captures from 2018–2020 are gold-standard evidence — if you have them, preserve them. WSPRnet and Reverse Beacon Network historical data is useful.
3. Time-stamped spectrogram captures during satellite passes. Use SDRs (KiwiSDR, RX-888/Web-888, Hermes Lite 2, Airspy HF+) with calibrated input and record in SigMF or cs16/cf32 raw format with time-stamp metadata. Software such as SatDump, GQRX and CyberSpectrum/SDR Console can cross-reference captures against TLE pass predictions.
4. Satellite identification. Cross-reference emission onset and fade timing against TLE passes from celestrak.org or n2yo.com. Report NORAD IDs where possible. Anti-correlated X/Y polarisation and sinusoidal fade envelope are positive indicators of spacecraft UEMR over terrestrial sources.
5. Rule out local sources — shut down your own switching supplies, LED lighting, PLC adapters and any Ethernet-over-power for the duration of the capture. Report what you tested.
6. Provide rather than conclude. Submit the raw data, the methodology, and the observation — let the engineers interpret. Speculation weakens reports; calibrated measurement strengthens them.

7. The Larger Fight

There is a respectable case for Starlink. Rural broadband in places where copper and fibre never reached, disaster-response connectivity after hurricanes and earthquakes, maritime and aviation internet at latencies legacy VSAT never delivered — these are genuine public goods. The amateur radio community is not, in general, opposed to satellite broadband. We are opposed to the proposition that ten thousand uncoordinated, unregulated RF emitters can be allowed to raise the global noise floor in bands we have used for a century, simply because those emissions are incidental rather than deliberate.

That proposition is the current regulatory default. It does not have to remain so. The ITU has a mature framework for protecting radio services from intentional emissions — power flux density limits, equivalent power flux density masks, coordination thresholds. What is missing is the same framework extended to unintended emissions for satellites that cross international borders thirteen times a day. Building that framework requires the record-building that amateur operators, radio astronomers and broadcasters do together.

The Collins engineers who designed the R-390A, the 75S-3 and the KWM-2 in the 1950s and 60s optimised their receivers for a noise environment that no longer exists. Galactic background and atmospheric noise set the floor at HF; at VHF the sky was genuinely quiet between transmitter passes. Preserving enough of that quiet for the next generation of radio operators, radio astronomers and weak-signal DX‑ers is — quite literally — a conservation fight. The only question is whether our community shows up with evidence.

References & Further Reading

1. Di Vruno F., Winkel B., Bassa C. G. et al. “Unintended electromagnetic radiation from Starlink satellites detected with LOFAR between 110 and 188 MHz.” Astronomy & Astrophysics, Vol. 676, A75 (2023). aanda.org/aa46374-23
2. Bassa C. G., Di Vruno F., Winkel B. et al. “Bright unintended electromagnetic radiation from second-generation Starlink satellites.” Astronomy & Astrophysics, A&A (2024). arxiv.org/abs/2409.11767
3. Grigg D., Tingay S. J., Sokolowski M. “The growing impact of unintended Starlink broadband emission on radio astronomy in the SKA-Low frequency range.” Astronomy & Astrophysics (July 2025). aanda.org/aa54787-25 · preprint arxiv.org/abs/2506.02831
4. ICRAR / Curtin University media release. “Interference to astronomy the unintended consequence of faster internet.” July 2025. icrar.org
5. SETI Institute & SpaceX. “SETI Institute and SpaceX Collaborate to Minimize Satellite Interference on Radio Astronomy.” June 2025. seti.org
6. NSF. “NSF and SpaceX Astronomy Coordination Agreement” (ODS / Telescope Boresight Avoidance). nsf.gov
7. ITU. “China to host ITU World Radiocommunication Conference 2027 in Shanghai.” December 2025. itu.int
8. Max Planck Institute for Radio Astronomy. “Starlink satellite electronics interfere with radio telescopes.” Press release, 2023. mpg.de/20610867
9. IAU Centre for the Protection of the Dark and Quiet Sky. cps.iau.org
10. ITU Recommendation ITU-R RA.769-2, “Protection criteria used for radio astronomical measurements.”
11. ITU WRC-27 Agenda. itu.int/wrc-27
12. Australian Communications and Media Authority — interference reporting portal. acma.gov.au
13. Federal Communications Commission — Electronic Comment Filing System (ECFS). fcc.gov/ecfs
14. ARRL Regulatory & RFI Information. arrl.org/regulatory-rfi-information
15. Wireless Institute of Australia. wia.org.au
16. McDowell, Jonathan — Starlink Launch Statistics. planet4589.org/space/con/star/stats.html
17. KeepTrack X Report — Starlink constellation operational status. keeptrack.space