Glonce

Starlink leakage floods SKA-Low's key radio frequencies, study finds

researchThe Glonce editorsHacker News🔥76Glonce rating76Impact72Novelty68Relevance68Surprise77Players50Credibility87
Starlink leakage floods SKA-Low's key radio frequencies, study finds

The Square Kilometre Array Low (SKA-Low) is designed to detect one of the faintest signals in astronomy: neutral hydrogen radiation from the cosmic dawn, the era roughly 13 billion years ago when the universe's first stars switched on. A team from Curtin University tested how well that plan is holding up using the Engineering Development Array 2 (EDA2), a prototype SKA-Low station in Australia. Over 29 days, the team analyzed about 76 million radio images and catalogued 112,534 individual radio emissions coming from 1,806 unique Starlink satellites, spread across 73 to 235 MHz, the exact frequency range SKA-Low needs. The results were published in the peer-reviewed journal Astronomy & Astrophysics.

The emissions are not the broadband beams Starlink deliberately transmits for internet service. They are accidental electronic noise leaking from onboard satellite hardware and radiating across frequencies nobody authorized, which means operators cannot schedule around it the way they could a planned downlink. Some satellites emit periodic 13-kHz tones at around 137 MHz every 100 seconds, and because the leakage is unpredictable, researchers say they cannot model it well enough to subtract it from their data. The scale is severe: leakage reaches as high as 10⁶ Jy/beam, while detecting the target cosmic-dawn hydrogen signal requires sensitivity near 10⁻⁵ Jy, meaning the Starlink leakage is reportedly around 10,000 times stronger than the signal SKA-Low is trying to hear. At some frequencies, up to 30% of the images the team captured contained Starlink interference, and emissions turned up inside two bands that the ITU protects specifically for radio astronomy, 73 to 74.6 MHz and 150.05 to 153 MHz, where no signals are supposed to exist at all. Steven Tingay of Curtin University described the leaked emissions as "comparable to the brightest natural radio sources in the sky."

The article frames the deeper problem as regulatory rather than technical. The ITU's rules protect designated radio astronomy bands only from intentional transmissions; unintended hardware leakage sits largely outside that framework, so it stays legally invisible even when it lands inside a protected band. In practice, nobody operating Starlink is breaking a rule by producing this leakage. SpaceX has previously worked with astronomers on related problems, fitting satellites with visors to cut optical brightness and reaching a coordination agreement with the NSF that addressed higher-frequency radio bands, but neither approach fixes low-frequency hardware leakage. The ITU is reportedly discussing the issue, though the article notes that discussion is not the same as regulation, and Starlink's constellation already exceeds 6,000 satellites. Researchers shared their findings with SpaceX, which is reportedly open to dialogue about future hardware changes. Algorithmic mitigation, filtering the leakage out in software, is being explored, but scientists describe the approach as "embryonic" and warn it could demand computing power rivaling the science processing itself.

The article's own framing treats a hardware-level engineering fix, similar to the visor redesign that addressed optical brightness, as the most realistic path forward, while leaving open when or whether that happens. It closes on an unresolved question: not just what the leakage is doing to astronomers now, but who gets to decide how much of the radio sky is worth protecting, and whether that decision arrives before the usable window closes.

Key facts

  • A Curtin University team analyzed about 76 million radio images over 29 days using the Engineering Development Array 2, a prototype SKA-Low station in Australia, and catalogued 112,534 emissions from 1,806 unique Starlink satellites across 73 to 235 MHz, the frequency range SKA-Low needs.
  • The leakage reaches up to 10⁶ Jy/beam, reportedly around 10,000 times stronger than the roughly 10⁻⁵ Jy sensitivity needed to detect cosmic-dawn hydrogen signals, and at some frequencies up to 30% of the images contained Starlink interference.
  • Some satellites emit periodic 13-kHz tones near 137 MHz every 100 seconds, unscheduled leakage researchers say they cannot model or subtract, and emissions were found inside two ITU-protected astronomy bands, 73 to 74.6 MHz and 150.05 to 153 MHz.
  • ITU rules protect astronomy bands only from intentional transmissions, so the unintended leakage sits in a regulatory gray zone even inside protected bands, and the Starlink constellation already exceeds 6,000 satellites.
  • Researchers shared their findings with SpaceX, which is reportedly open to dialogue on hardware changes; scientists call proposed algorithmic mitigation "embryonic" and say it could demand computing power rivaling the science processing itself.

Why it matters

SKA-Low exists to detect one of the faintest signals in astronomy, neutral hydrogen radiation from the cosmic dawn roughly 13 billion years ago, and that mission needs sensitivity down to about 10⁻⁵ Jy. This study puts real numbers on how much Starlink leakage sits inside the exact frequencies the telescope depends on: 112,534 emissions from 1,806 satellites, reaching up to 10⁶ Jy/beam, or roughly 10,000 times stronger than the target signal. Because the leakage is accidental hardware noise rather than a scheduled transmission, astronomers cannot simply plan around it or filter it out with existing techniques. At the frequencies most affected, up to 30% of images were contaminated, which the article says risks rendering entire frequency bands scientifically unusable.

Who it affects

Radio astronomers and the Curtin University team running the SKA-Low prototype station are the ones documenting and absorbing the interference. SpaceX and its Starlink satellites are the source of the leakage; the company has been in contact with the researchers and is reportedly open to discussing hardware changes. The ITU, the body that sets international rules for protected radio astronomy bands, is named as the standards authority whose current framework does not address unintended emissions, and the NSF is named as a precedent, having reached a prior coordination agreement with SpaceX covering higher-frequency radio bands.

How to use it

There is no fix available today. The article points to two precedents SpaceX has already delivered elsewhere: satellite visors that cut down optical brightness, and the NSF coordination agreement addressing higher-frequency radio bands; neither approach touches this low-frequency hardware leakage. Two paths are floated for the future: an engineering redesign of the hardware itself, which the article treats as the more realistic route, and algorithmic mitigation that would filter the leakage out in software, which scientists call "embryonic" and warn could require computing power rivaling the science processing itself. Researchers have shared their findings with SpaceX, and the ITU is reportedly discussing the issue, though the article is explicit that discussion is not regulation.

How solid is it

The central numbers come from a peer-reviewed paper in Astronomy & Astrophysics, based on a stated methodology: about 76 million radio images analyzed over 29 days on the EDA2 prototype station. The article names one researcher on the record, Steven Tingay of Curtin University, with a direct quote about the emissions' strength. It does not name the paper's other authors, give a publication date for the paper or the article, or specify the calendar dates of the 29-day observation window. It also does not say which or how many of the 1,806 satellites produce the periodic 13-kHz tones, describing that only as "some satellites."

Risks and caveats

The article does not describe any concrete hardware fix, only that engineering and design changes are the likely direction; SpaceX has not committed to a specific change, only reported openness to dialogue. No timeline or deadline is given for possible ITU action, or confirmation that regulation will happen at all, and the Starlink constellation already exceeds 6,000 satellites. The proposed software alternative, algorithmic mitigation, is explicitly called "embryonic" and flagged as potentially demanding computing power that rivals the astronomy processing itself, so neither path is close to solving the problem the study documents.

“Comparable to the brightest natural radio sources in the sky.”

— Steven Tingay, Curtin University