US spy satellites built by SpaceX send signals in the “wrong direction”

It seems US didn't coordinate Starshield's unusual spectrum use with other countries.

See full article...

 

[Wickwick]

Thoughts as to whether these satellites would have enough receive gain to be using these bands for Doppler radar?

 

[Wickwick]

The US once again using the diplomatic methods of: "Guess what? Screw you, that's what!"

Strictly speaking, they're allowed to operate in a non-interfering manner (which they're apparently doing) in any band without consultation. It's only after a country complains that there's interference that we can just the US's diplomatic method.

 

[Wickwick]

Why not? Satellite-sourced SAR is a pretty well-understood mechanism , and the band they're using would seem to be good for medium-resolution imaging.

https://www.earthdata.nasa.gov/learn/earth-observation-data-basics/sar

Doppler radar would be specifically for moving targets if it's more focused toward the horizon than down-looking. Full global awareness of all moving craft seems like a military advantage.

 

[Wickwick]

Doing a bit of surmising here, but I don't think Doppler would be performant in all axes of movement, no matter the power/gain? I'd envision OpenCV-style motion detection analysis as a better situation solution.

Doppler radar gives you the velocity along the transceiver direction. So a down-looking radar would see a velocity of zero for an aircraft at level flight. Sure, you can use image processing to give you velocities, but that takes processing.

Doppler radar gives you the velocity on the return signal directly. There's no more work than the heterodyning you have to do on a radar signal. If you're looking at the horizon, you're getting the component of velocity to/from the satellite. You can then filter for velocity ranges such that hypersonic vehicles would stick out like a sore thumb.

 

[Wickwick]

What you're describing is the mentality of "it's not wrong if we get away with it". If radio frequency spectrum was reserved for specific purposes, it should only be used for those purposes.

Right there in TFA is a description that off-nominal uses are specifically allowed on a non-interfering basis. The "reservations" are for potentially interfering use cases.

 

[Wickwick]

The signals are there, which is apparent.

But are they just like a navigation beacon - broadcasting a signal at a specific frequency? Or are they coherent signals containing data?

These are SPY SATELLITES. Transmitting a signal means waving a big-assed flag saying, "I'm here!" They announce when they're up there and WHERE THEY ARE if someone wants to passively home in on them and do mean things to them.

So, to ME at least, that's a BIG fucking problem for spy satellites.

Other kinds, probably not so much.

Any satellite that's sending information to the ground is waving a massive beacon as you describe. The entire point of Starshield is that there are too many to bother to home in on individually. It's more expensive to shoot one down than to have launched it in the first place.

 

[Wickwick]

To be fair though, Tilley strikes me as somebody with a lot more expertise and investment than just an 'amateur home brewer'. Like, widespread deployment of Starshield began in 2023? 2024? So far, the only "public" response to this has originated with him.

Discussions of ethics of disclosure aside, if this really was causing issues, you would figure somebody actively using the spectrum outside the U.S. would have screamed by now.

I think the point is that Tilley's disclosure didn't come as a surprise to any nation state interested in US capabilities in space.

 

[Wickwick]

[my emphasis] To be pedantic, in special relativity, you do get a Doppler effect on EM bounced off objects moving perpendicular to the direction of view. I have no idea whether modern SAR can take advantage of that. At LEO speeds, the effect would be quite small.

I think the frequency offset from the relativistic Doppler shift would be within the emission bandwidth of the emitter for objects traveling below a few km/s.

 

[Wickwick]

If the target is flying at a vector perpendicular to the beam, with doppler you're going to get a very different velocity than if it were flying head-on to the beam. Might be okay for detection, but to develop an intercept I'd think you 'd need a much better velocity vector. Now, if you have multiple radars painting the same bogey, fusion might make doppler sources work, but that's processing.

Right. You get the component of velocity along the beam. So a craft flying perpendicular to the satellites view would not yield an interesting result. You can tell it's there, but you can tell that without using Doppler.

But if I have an overlapping constellation of satellites looking at overlapping regions of space from multiple directions, it's impossible for a craft to be moving perpendicular to all viewers. I don't need to correlate between satellites. I need only check that nothing is exceeding my speed limits for each satellite. Not to mention that the view angle of an object is always changing for a given satellite.

Edit: And I'm not suggesting that these systems would be accurate enough to resolve an intercept. More powerful assets can be brought to bear for that. But those more powerful assets tend to be tightly-focused devices. They aren't good at large-area coverage.

 

[Wickwick]

The high signal-to-noise ratio is a dead giveaway. You wouldn't get that from an accidentally scattered transmission, it requires a high-gain antenna pointed in the right direction, i.e., towards the ground.

Unless we're seeing the side lobes of a massively high-powered transmission. And sending that much power between satellites would almost certainly be interfering with ground-to-space operations for any strays caught in the transmission.

 

[Wickwick]

As an amateur radio operator and amateur astrophotographer, I thoroughly enjoyed this article.

Interestingly, the signal powers that Tilley describes are likely underestimates. You would assume the satellites are using directional antennas, and Tilley's antennas are almost certainly not directly in the target receiving location. 15-20dB above the noise from a random location likely means a quite powerful transmission at the actual intended receiver location.

Unless it's an omnidirectional (or wide-angle at least) emission.

 

[Wickwick]

Not high power - a few watts into an omni is easily detectable above the noise floor from LEO with even a modest antenna. All those cubesats using S-band downlinks have omni antennas.

At 20dB?

 

[Wickwick]

Not really relevant to your main point, but 2 GHz is low enough that modern digital receivers can sample it directly. No mixer required.

A very good reminder, thanks! All my background in heterodyning is at optical wavelengths. We're a wee bit away from sampling that directly!