Starlink : New FCC and ITU Filings

[RedLineTrain]

Pending further review and coordination with Federal users, we defer acting on SpaceX’s request to use the E-Band as well as any action regarding SpaceX’s described use of the emergency beacons.

In my opinion, this is a very heavy blow to SpaceX. Now the main problem of space X is the lack of satellite bandwidth, which is determined by the fact that the transmission of information to the satellite is limited to the 2000 MHz in  Ka band (total 4000 Mhz in 2 polarisation)  Therefore, the most important change in Gen2 was the addition of the E band where 5000 MHz can be used, while in the SpaceX application the frequency channels in the E band were repeated in such a way that the total band available for transmission from the Earth to the satellite was almost  8 times more than that of the first generation satellite .
And precisely because the same frequency range had to be used several times (as far as I remember up to 4 ) - that is, the satellite had to have 4 antennas aimed at different gateways, satellites of the second generation and had to be much heavier and larger in sizes than generation 1. Of course, now they can be designed to use only the Ka band, and then only 4 x 2000 x 2 polarizations of 8000 MHz will be available to StarLink. instead of the planned 5000  x 4 + 2000 x 2 x 4 = 28000 MHz

that is, less than 30% of the planned, respectively, and the Ku band antennas (from Sat to User terminal) can only be used by 30%, and therefore you don’t need so much energy for satellite, and such large solar panels and batteries are not needed on board too...

This is a temporary setback, considering that V-band frequencies will be requested for Gen2 shortly.  I suspect that SpaceX wasn't ready with equipment for the E-band frequencies quite yet anyway.

To recap, the V-band frequencies were 5 GHz up and 5 GHz down, both for user terminals and gateways.

[gongora]

I'd guess V-band would just be used for gateway comms on these sats (unless maybe they make a V-band aero terminal eventually).

[russianhalo117]

Pending further review and coordination with Federal users, we defer acting on SpaceX’s request to use the E-Band as well as any action regarding SpaceX’s described use of the emergency beacons.

In my opinion, this is a very heavy blow to SpaceX. Now the main problem of space X is the lack of satellite bandwidth, which is determined by the fact that the transmission of information to the satellite is limited to the 2000 MHz in  Ka band (total 4000 Mhz in 2 polarisation)  Therefore, the most important change in Gen2 was the addition of the E band where 5000 MHz can be used, while in the SpaceX application the frequency channels in the E band were repeated in such a way that the total band available for transmission from the Earth to the satellite was almost  8 times more than that of the first generation satellite .
And precisely because the same frequency range had to be used several times (as far as I remember up to 4 ) - that is, the satellite had to have 4 antennas aimed at different gateways, satellites of the second generation and had to be much heavier and larger in sizes than generation 1. Of course, now they can be designed to use only the Ka band, and then only 4 x 2000 x 2 polarizations of 8000 MHz will be available to StarLink. instead of the planned 5000  x 4 + 2000 x 2 x 4 = 28000 MHz

that is, less than 30% of the planned, respectively, and the Ku band antennas (from Sat to User terminal) can only be used by 30%, and therefore you don’t need so much energy for satellite, and such large solar panels and batteries are not needed on board too...

The deferment was without prejudice which means that they can resubmit at a later date. This has happened before for many companies so not a big deal. Just a temporary setback since V-Band rollout is their priority anyways at the moment.

[gongora]

I haven't really seen anything saying V-band is a priority

[waveney]

I am an amateur astronomer and student of astronomy. Astronomy can deal with this.

These actions by the regulators are about monkey wrenching SpaceX while paying lip service to astronomers and fans of the night sky. US over regulation in this area will cede the industry to other nations.

Here Here.

I too am an amateur astronomer.  (With my name on several Galaxy Zoo papers)

I don't think any Astronomers use the old film and plates with long exposures.  Everyone and everything these days stacks many short exposure images in some way or another  (even Hubble and JWST) If you provide accurate info on when and when they will be observed then the images can be processed to take the effects of a passing sat.  There are far worse problems to worry about, loss of radio spectrum, loss of dark skies for example.

[TheRadicalModerate]

I am an amateur astronomer and student of astronomy. Astronomy can deal with this.

These actions by the regulators are about monkey wrenching SpaceX while paying lip service to astronomers and fans of the night sky. US over regulation in this area will cede the industry to other nations.

Here Here.

I too am an amateur astronomer.  (With my name on several Galaxy Zoo papers)

I don't think any Astronomers use the old film and plates with long exposures.  Everyone and everything these days stacks many short exposure images in some way or another  (even Hubble and JWST) If you provide accurate info on when and when they will be observed then the images can be processed to take the effects of a passing sat.  There are far worse problems to worry about, loss of radio spectrum, loss of dark skies for example.

I am no longer an amateur astronomer, but I got interested in this problem enough to learn about CCDs.  Bottom line: you can't just stack exposures on top of each other for photometry applications.

CCDs are not bit-addressable.  You expose the CCD array to your target light source, and every photon causes the pixel it strikes to accumulate a bit more charge.  Then, when the exposure is over, you shift each pixel in the array, line-by-line and then pixel-by-pixel, through the same A/D converter to develop the digitized image.  It's a bucket brigade architecture.  The pixels in the last line in the array are shifted through all the pixels at the same address in all of the preceding lines.

The problem is that the CCD is calibrated to accumulate a certain number of maximum photons.  If you exceed that maximum, three things happen:

1) You no longer have a good photometrically-calibrated value for that pixel, because it's now essentially clipping.

2) The pixel overflows into surrounding pixels, invalidating them as well.

3) As you're shifting the pixel's line through the bucket brigade, over to the edge of the device, for reading through the ADC, the overflow leaves a trail of charge behind it, which will contaminate all the pixels at the same position, in all the succeeding lines, as they shift through the trail.

The problem with bright satellites is that they cause overflows, and usually they cause overflows in at least a couple of pixels in most lines, which means that your photometry across a large amount of the image is extremely noisy.  If you add together a bunch of images like that post-exposure, your photometry is all messed up.

One solution that you might think would work is simply to stop the exposure before the satellite goes over, read out the image, and then re-expose after the satellite has left the field of view.  But photometric-grade CCDs need to be cleared and re-calibrated for noise between exposures.  How this usually works is that the mechanical shutter closes, the CCD is read, then cleared (how it's cleared I don't understand), and then the blank CCD usually takes a picture of the closed shutter to get a noise calibration.  Only then is it ready for a new live exposure with the shutter open.

That takes... a while.  It's not minutes, but it's probably seconds.  And in a dense enough satellite environment, with a wide enough field of view (like Vera Rubin, which is the poster child for this problem), there are enough satellites crossing the FoV that you're consuming a sizable fraction of observing time.

What you need is a pixel that has all the nice noise properties of a CCD, but which is individually addressable, as well as readable and clearable on the fly, during an exposure.  Then you read it before the satellite crosses it, clear it after it's gone, and continue with the exposure.  You've messed up your photometry a little bit, but you never clipped the signal, you never overflowed into other pixels, and the noise you've added is less than you'd have when a wisp of cloud went over part of the FoV.

CMOS can do that.  However:

1) CMOS is a lot noisier than CCD.  It's getting better, but it's not there yet.

2) If you read a CMOS pixel, you have to read it into its own ADC (or an ADC assigned to a small group of pixels), and it turns out that calibrating a bunch of ADCs so they all consistently convert the same amount of charge to the same digital value is extremely difficult.

So, right now, there isn't a detector technology that works.  However, I suspect the answer is to work on some pixel-addressable technology until the noise and ADC calibration problems are no longer an issue.

So it's a real problem, but it's a problem that likely has a technological workaround if you throw enough semiconductor geeks at it.  Then you still have the problem that good wide-field photometric detectors are heavily integrated into their telescopes, and swapping one out isn't easy or cheap.  But, again, that's a technological and financial problem, with a much cheaper solution than "don't put up any more satellites".

[gongora]

The astronomy conversations need to move to the astronomy & megaconstellations thread.

[RedLineTrain]

Here's vaporcobra's article regarding this partial approval.  He is downbeat on the conditions that the FCC imposed.  I share this view at least with regard to the FCC's intent, but believe that over time SpaceX could turn this around using an incremental approach.

In any event, it will take a while to ramp Starship launches and to integrate the V-band frequencies into Gen2.  And we will see fairly shortly how well the dialectic stickers work.  If those need some iteration, this partial approval gives SpaceX time to do that.  Lastly, there could be some benefit to SpaceX in that the partial approval is a bit more lawsuit-proof.

https://www.teslarati.com/spacex-starlink-gen2-constellation-fcc-license-partial-grant/

[Reynold]

There is also the interesting point that SpaceX may be able to meet some of these conditions more easily than their competitors, who may eventually regret asking for such restrictions when they have to meet them too. 

[deadman1204]

There is also the interesting point that SpaceX may be able to meet some of these conditions more easily than their competitors, who may eventually regret asking for such restrictions when they have to meet them too.

Companies generally ask for restrictions that they know they can already meet.

[gongora]

https://twitter.com/thesheetztweetz/status/1600595289995345920

https://fcc.report/IBFS/SAT-PPL-20221206-00170
SAT-PPL-20221206-00170

Description: SpaceX requests U.S. market access for its German-licensed direct-to-cellular payload to communicate on an unprotected, non-interference basis in the 1910-1915 MHz uplink E-s and 1990-1995 MHz downlink s-E bands PCS G Block with off-the-shelf cellular mob

[TheRadicalModerate]

And we will see fairly shortly how well the dialectic stickers work.

The typo in this sentence makes me imagine a Starlink bird covered in Stalinist-era Soviet propaganda posters.

[su27k]

There is also the interesting point that SpaceX may be able to meet some of these conditions more easily than their competitors, who may eventually regret asking for such restrictions when they have to meet them too.

That would require FCC applies the same conditions on their competitors, so far we're not seeing it. For example the object-year thing should be established via rule making instead of being applied to Starlink specifically.

Also a lot of Starlink's competitors registered their constellation in foreign countries, so they don't need FCC authorization, just need market access/landing right which is much more lax and doesn't have the same conditions.

[Robotbeat]

I missed this filing from October with the proposed Gen 2 satellite dimensions for Starship and F9 versions.
https://licensing.fcc.gov/myibfs/download.do?attachment_key=17429628

Did anybody actually download this and attach it to a post? I can't get it to load.

[Robotbeat]

There is also the interesting point that SpaceX may be able to meet some of these conditions more easily than their competitors, who may eventually regret asking for such restrictions when they have to meet them too.

That would require FCC applies the same conditions on their competitors, so far we're not seeing it. For example the object-year thing should be established via rule making instead of being applied to Starlink specifically.
...

FCC could be sued if they don't apply the conditions in a fair manner. So I'm fully confident this will happen.

[gongora]

I missed this filing from October with the proposed Gen 2 satellite dimensions for Starship and F9 versions.
https://licensing.fcc.gov/myibfs/download.do?attachment_key=17429628

Did anybody actually download this and attach it to a post? I can't get it to load.

You may have to click it twice.  It's a feature of the FCC web site.

[su27k]

Swarm filed for STA and modifications so that their earth stations at Guam, Hawaii, Utah and New Jersey can communicate with Gen2's VHF beacon:

https://fcc.report/IBFS/SES-STA-20221215-01361
https://fcc.report/IBFS/SES-STA-20221215-01372
https://fcc.report/IBFS/SES-STA-20221215-01374
https://fcc.report/IBFS/SES-STA-20221215-01375
https://fcc.report/IBFS/SES-MOD-20221216-01364
https://fcc.report/IBFS/SES-MOD-20221216-01365
https://fcc.report/IBFS/SES-MOD-20221216-01367
https://fcc.report/IBFS/SES-MOD-20221216-01370

[raptorx2]

I missed this filing from October with the proposed Gen 2 satellite dimensions for Starship and F9 versions.
https://licensing.fcc.gov/myibfs/download.do?attachment_key=17429628

Did anybody actually download this and attach it to a post? I can't get it to load.

You may have to click it twice.  It's a feature of the FCC web site.

Or you can find here.  https://www.dropbox.com/s/dhjubtpzkl03z39/SpaceX.pdf?dl=0

[raptorx2]

I missed this filing from October with the proposed Gen 2 satellite dimensions for Starship and F9 versions.
https://licensing.fcc.gov/myibfs/download.do?attachment_key=17429628

Did anybody actually download this and attach it to a post? I can't get it to load.

There has also been speculation that SpaceX could implement a VHF GNSS service through these beacons

[Barley]

There has also been speculation that SpaceX could implement a VHF GNSS service through these beacons

Where is the market, when the US gives GPS, WAAS etc away for free?