Radio astronomy has published a cost for wireless power transfer. It is a separation

distance of about five kilometres, and it was computed for inductive chargers delivering

more than thirty watts below 30 MHz. For a space-to-Earth beam there is no number at all,

and this article works out roughly how large the missing one is.

The source is the Summer 2026 issue of CRAF News, the newsletter of the European Science

Foundation's Expert Committee on Radio Astronomy Frequencies, published in September 2026.

Its article "Wireless Power Transfer and Radio Astronomy: an emerging compatibility

challenge", by Fabio Giovanardi of the Arcetri Astrophysical Observatory at INAF in

Florence, is the first place this project has seen the radio astronomy community put a

quantified compatibility assessment of wireless power next to a statement about space-based

solar power.

It is worth being exact about what that is and is not. CRAF is a scientific expert committee.

It does not set the ITU agenda, it does not write the Radio Regulations, and a newsletter

article is not a work item. Nothing in it changes a pinned edition. What it does do is put a

method and a distance on the record, and the interesting part is where that method stops.

What CRAF actually published

The assessment behind the article is the Addendum to ECC Report 333, which the newsletter

describes as covering higher-power, non-beam inductive wireless power systems operating in

specific bands below 30 MHz and delivering more than 30 W. The bands at risk are given as

the radio astronomy allocations at 13.36 to 13.41 MHz and 25.55 to 25.67 MHz, and the

primary case study is LOFAR, which ASTRON's own telescope page puts at more than 70,000

antennae across Europe, with 38 stations in the Netherlands, six in Germany, three in

Poland, and one each in France, Ireland, Latvia, Sweden and the United Kingdom.

The compatibility analysis takes a single-interferer, ground-wave approach under

Recommendation ITU-R SM.2028, applies the protection criteria of Recommendation ITU-R

RA.769, and draws emission assumptions from ETSI EN 303 417, CISPR 11 and CISPR 14. The

result, in the newsletter's own words, is that "a separation distance of around 5 km from

the radio observatory would be necessary". Because the devices are portable, the article

says enforcing such a zone is a regulatory problem rather than a planning one.

The newsletter also sets out why the terrestrial problem is getting worse rather than

better. Electric vehicle charging, covered by ECC Report 289 and its own addendum, is put at

transferred power of up to 11 kW for light duty and up to 500 kW for heavy duty. CRAF's

point about these is not the single-unit figure but the aggregate: a distributed fleet of

chargers raises the noise floor in a way that is harder to mitigate than one high-power

transmitter, and low-frequency signals propagate well. Against that, the one consolation is

that vehicle chargers are fixed installations, so a separation distance is at least

enforceable. The portable devices in the 30 W class are the ones nobody can put a fence

around.

Then there is half a page on space-based solar power. CRAF's position is that the

interference problem does not stop at ground level, and that, quoting the Summer 2026

newsletter, "SBSP transmissions would be visible over continental scales and could affect

multiple radio astronomy observatories simultaneously". There is no distance, no power

density and no band in that paragraph. It is a flag, not a study, and the article says so by

calling for the question to be taken up within ITU frameworks rather than reporting an

answer.

The method stops at thirty megahertz, and it says so in its title

This is the part that matters for anyone planning a receiving site. The five kilometres

cannot be carried across.

Recommendation ITU-R SM.2028 is listed on the ITU's own editions page as SM.2028-0 (09/2012),

In force, under the title "Protection distance calculation between inductive systems and

radiocommunication services using frequencies below 30 MHz". Inductive systems. Below 30 MHz.

A ground-wave propagation model is the right tool there and the wrong tool for a microwave

beam arriving from orbit at a steep elevation angle.

Recommendation ITU-R RA.769 is more pointed. The in-force edition is RA.769-2 (05/2003),

listed by the ITU as incorporated by reference in the Radio Regulations, which gives its

tables real regulatory weight. Section 2 of its Annex 1 opens by saying that the threshold

levels in Tables 1 and 2 "are applicable to terrestrial sources of interfering signals". It

then handles geostationary satellites as a special case, and recommends that for

transmitters in the geostationary orbit the tabulated levels be adjusted by a further

15 dB, because a telescope pointed within about five degrees of such a transmitter sees it

through a much higher side lobe.

RA.769-2 also anticipated the problem in 2003. Among its considerings is the observation

that "some transmissions from spacecraft can introduce problems of interference to radio

astronomy and that these cannot be avoided by choice of site". Siting is the entire

mitigation available for terrestrial wireless power. It is the one mitigation that does not

work against a satellite.

The missing number is about 181 dB, and that is an estimate

Here is the arithmetic, labelled as an estimate because it is mine, derived from two

documented figures rather than taken from any study.

Table 1 of RA.769-2 gives, for continuum observations centred on 4 995 MHz over a 10 MHz

bandwidth, a detrimental threshold power flux density of −171 dB(W/m²). That is the 4 990 to

5 000 MHz radio astronomy band, and it is the nearest protected continuum entry in the table

to the band space solar power concepts most often name. The same table gives −201 dB(W/m²)

at 13.385 MHz and −199 dB(W/m²) at 25.610 MHz, which are the low-frequency bands CRAF is

worried about. Both figures assume a 2 000 s integration, a 0 dBi side lobe, and a single

antenna.

Take an upper bound on what a ground spot would be allowed to carry. ICNIRP's 2020

radiofrequency guidelines set a whole-body reference level of 10 W/m² above 2 GHz, and

IEEE C95.1-2019 is the companion standard a United States site would be assessed against.

A beam constrained to 10 W/m² at the spot sits about 181 dB above the RA.769 threshold for

the adjacent protected band. Apply RA.769's own 15 dB geostationary adjustment and the

figure is about 196 dB.

That is not a claim that a beam would cause 181 dB of interference. The beam is not in the

radio astronomy band, so what the number describes is the out-of-band suppression, in the

telescope's side lobes, that a space-to-Earth system would have to demonstrate in a

neighbouring protected allocation. It is a rough upper bound and it ignores the bandwidth

conversion between the table's 10 MHz reference and a real spurious emission mask. The

reason to state it anyway is that nothing in the literature this project can reach states

anything at all, and 181 dB is a different order of engineering problem from the 5 km

exclusion zone CRAF has costed. Note also that 5.8 GHz, the band most frequently discussed

for this, is not regulatorily available for space-to-Earth power transmission; there is no

allocation and the ITU's own Report SM.2392-2 (06/2026) is a report, which recommends

nothing and binds nothing.

The regulator has already set a limit above what its own studies asked for

The terrestrial precedent is worth reading closely, because it is the one that will be

reached for.

ECC Recommendation (25)01, "Emission limit requirements for Wireless Power Transmission

(WPT)", approved 14 February 2025, sets in its Table 4 an unwanted emission limit in the

spurious domain of −20 dBµA/m per 10 kHz at 10 m across 5 to 30 MHz, for non-beam generic

WPT. That is the figure CRAF cites, and reading the recommendation first-party confirms it.

Two other lines in the same document matter more. Its considering k) records "some study

results in ECC Report 333 found limits to protect certain radio communication services that

are lower than those in this Recommendation". The regulator knew the studies asked for

tighter limits and set looser ones. And its recommends 6 states "that WPT applications

permitted under this ECC Recommendation operate on a non-interference, non-protected basis".

For an operator that is the sentence to notice. Non-interference, non-protected means the

wireless power system carries the whole risk: it must not interfere, and it has no claim if

something interferes with it. If that framing is inherited by beamed systems, a ground

receiving site near a protected passive allocation is not a planning risk that can be bought

off. It is an unbounded liability with no standing.

What this changes for a ground site

Nothing yet, and that is the honest answer. No edition in this project's registry moved this

week. But it adds a column to the question of where a receiving site can go, and the column

is not one that appears in a land search. Distance to a radio astronomy facility is

currently nobody's site-selection criterion for beamed power. On the terrestrial evidence it

is already worth five kilometres for a thirty-watt charger, enforced through a device

standard rather than a licence, and the only body that has written the satellite case down

has written it down as a concern rather than a calculation.

The useful posture is to treat the distance as unknown and to say so, rather than to assume

that an allocation at 5.8 GHz would make the adjacent 4 990 to 5 000 MHz band somebody

else's problem.

What would change my mind

A published compatibility study for beamed space-to-Earth power against RA.769, by anyone,

using a propagation model valid above 30 MHz. If that study comes back with a required

suppression well below the 181 dB estimated here, because a real spurious mask and a real

antenna pattern are far better than an upper bound on main-beam density implies, then the

framing above is alarmist and the correct answer is that filtering solves it.

Equally, a first-party ITU-R work item that takes the space case under an existing agenda

item would change the shape of this. CRAF's article routes the question to WRC-31

preliminary agenda item 2.2 under Resolution 910 (WRC-23), which this project has already

examined. If instead a WRC-27 item absorbed it, the timeline moves forward by four years and

the argument that nobody is working on it fails.

One limitation is stated rather than hidden. The Addendum to ECC Report 333 itself was not

read for this article: docdb.cept.org failed TLS certificate verification on every attempt

from this run, so the Addendum's scope, its five-kilometre result and its ECC Report 289

companion for electric vehicle charging are reported as the CRAF newsletter states them, and

have not been independently verified against CEPT's own text. ECC Recommendation (25)01 was

read first-party, through a different host, in its 14 February 2025 edition; a later amended

edition is indexed and was not reachable.

Sources