Author Topic: Starlink : New FCC and ITU Filings  (Read 296737 times)

Offline OceanCat

  • Full Member
  • *
  • Posts: 138
  • Liked: 260
  • Likes Given: 217
Re: Starlink : New FCC and ITU Filings
« Reply #580 on: 02/22/2023 07:28 pm »
The FCC granted E-band temporary license with conditions (besides what looks like general restrictions for the band):

* SpaceX may not operate more than 1,500 satellites in the E-band in the year following grant of this STA, subject to any updates in a future authorization.
* Maximum of 51 total fixed gateway sites
* Maximum of 19 fixed gateway sites within any square on the Earth that covers two million km2
* Maximum of 32 active uplink beams per fixed gateway site (I think that matches the long term license request)

Sounds reasonable to me to get the ball rolling.

Offline kevin-rf

  • Elite Veteran
  • Senior Member
  • *****
  • Posts: 8823
  • Overlooking the path Mary's little Lamb took..
  • Liked: 1318
  • Likes Given: 306
Re: Starlink : New FCC and ITU Filings
« Reply #581 on: 02/22/2023 10:08 pm »
Quote
Maximum of 51 total fixed gateway sites within the continental US (CONUS)

« Last Edit: 02/22/2023 10:08 pm by kevin-rf »
If you're happy and you know it,
It's your med's!

Offline M.E.T.

  • Senior Member
  • *****
  • Posts: 2335
  • Liked: 2933
  • Likes Given: 510
Re: Starlink : New FCC and ITU Filings
« Reply #582 on: 02/23/2023 05:17 am »
The FCC granted E-band temporary license with conditions (besides what looks like general restrictions for the band):

* SpaceX may not operate more than 1,500 satellites in the E-band in the year following grant of this STA, subject to any updates in a future authorization.
* Maximum of 51 total fixed gateway sites
* Maximum of 19 fixed gateway sites within any square on the Earth that covers two million km2
* Maximum of 32 active uplink beams per fixed gateway site (I think that matches the long term license request)

Sounds reasonable to me to get the ball rolling.

What does that mean? Good news for higher bandwidth, or bad?

Offline kevin-rf

  • Elite Veteran
  • Senior Member
  • *****
  • Posts: 8823
  • Overlooking the path Mary's little Lamb took..
  • Liked: 1318
  • Likes Given: 306
Re: Starlink : New FCC and ITU Filings
« Reply #583 on: 02/23/2023 05:26 pm »
Well, 2 million km^2 is a radius of 800km, so not more than 19 ground stations in a circle of 1600km diameter?

The distance from Boston to Orlando is about 2000km. So it would limit them to about 19 ground stations on each coast. Guess it keeps them from concentrating all 51 in any single geographic location.

My question, like the limit of 51 US ground stations, does the 2 million km only apply to the US?
If you're happy and you know it,
It's your med's!

Offline Kiwi53

  • Full Member
  • *
  • Posts: 151
  • New Zealand
  • Liked: 153
  • Likes Given: 225
Re: Starlink : New FCC and ITU Filings
« Reply #584 on: 02/23/2023 09:50 pm »
My question, like the limit of 51 US ground stations, does the 2 million km only apply to the US?
This is a US FCC ruling, so yes, it would apply only in the USA (incl Hawai'i) and presumably also US overseas territories

Offline kevin-rf

  • Elite Veteran
  • Senior Member
  • *****
  • Posts: 8823
  • Overlooking the path Mary's little Lamb took..
  • Liked: 1318
  • Likes Given: 306
Re: Starlink : New FCC and ITU Filings
« Reply #585 on: 02/24/2023 02:28 pm »
So, no more than 19 ground stations on Guam it is ;-)
If you're happy and you know it,
It's your med's!

Online JayWee

  • Full Member
  • ****
  • Posts: 1005
  • Liked: 1002
  • Likes Given: 1896
Re: Starlink : New FCC and ITU Filings
« Reply #586 on: 02/24/2023 05:05 pm »
Well, 2 million km^2 is a radius of 800km, so not more than 19 ground stations in a circle of 1600km diameter?

The distance from Boston to Orlando is about 2000km. So it would limit them to about 19 ground stations on each coast. Guess it keeps them from concentrating all 51 in any single geographic location.
How much bandwidth is that? (please verify)

51 gateways * 32 beams per gateway * 10GHz * 64QAM FEC 8/9 (from v1), ie that's 5.11 b/Hz, I get 83 TBps over CONUS.

CONUS has area of 8 Gm², each cell 379 km², that's 21320 cells. So ~ 4Gbps per cell. + V1 Ku/Ka uplink.
« Last Edit: 02/24/2023 06:38 pm by JayWee »

Offline oldAtlas_Eguy

  • Senior Member
  • *****
  • Posts: 5307
  • Florida
  • Liked: 5009
  • Likes Given: 1473
Re: Starlink : New FCC and ITU Filings
« Reply #587 on: 02/24/2023 05:27 pm »
Well, 2 million km^2 is a radius of 800km, so not more than 19 ground stations in a circle of 1600km diameter?

The distance from Boston to Orlando is about 2000km. So it would limit them to about 19 ground stations on each coast. Guess it keeps them from concentrating all 51 in any single geographic location.
How much bandwidth is that? (please verify)

51 gateways * 32 beams per gateway * 10GHz * 64QAM FEC 8/9 (from v1), ie that's 5.11 b/Hz, I get 83 TBps over CONUS.

CONUS has area of 8 Gm², each cell 379 km², that's 21320 cells. So ~ 4Gbps per cell. + V1 Ku/Ka uplink.
Per 250Mbps customers that 83Tbps is 332,000 simultaneous downloading customers at full bandwidth. The total practical is more like a factor 20 more customers easily or >6M customers supported at download bit rates of 250Mbps. SpaceX could up that in a couple of years (by 2025) for the consumer terminals to do 500Mbps by being able to put in more ground stations as in 100 vs the current authorized in this first year or two of operations of 51.

Online JayWee

  • Full Member
  • ****
  • Posts: 1005
  • Liked: 1002
  • Likes Given: 1896
Re: Starlink : New FCC and ITU Filings
« Reply #588 on: 02/24/2023 06:32 pm »
Well, 2 million km^2 is a radius of 800km, so not more than 19 ground stations in a circle of 1600km diameter?

The distance from Boston to Orlando is about 2000km. So it would limit them to about 19 ground stations on each coast. Guess it keeps them from concentrating all 51 in any single geographic location.
How much bandwidth is that? (please verify)

51 gateways * 32 beams per gateway * 10GHz * 64QAM FEC 8/9 (from v1), ie that's 5.11 b/Hz, I get 83 TBps over CONUS.

CONUS has area of 8 Gm², each cell 379 km², that's 21320 cells. So ~ 4Gbps per cell. + V1 Ku/Ka uplink.
Per 250Mbps customers that 83Tbps is 332,000 simultaneous downloading customers at full bandwidth. The total practical is more like a factor 20 more customers easily or >6M customers supported at download bit rates of 250Mbps. SpaceX could up that in a couple of years (by 2025) for the consumer terminals to do 500Mbps by being able to put in more ground stations as in 100 vs the current authorized in this first year or two of operations of 51.
332000 * 20 * $100/month * 12 months = $8B/year
Verizon (142M users) had revenue of $132B in 2022 (that's $77/user).
« Last Edit: 02/24/2023 06:38 pm by JayWee »

Offline kevin-rf

  • Elite Veteran
  • Senior Member
  • *****
  • Posts: 8823
  • Overlooking the path Mary's little Lamb took..
  • Liked: 1318
  • Likes Given: 306
Re: Starlink : New FCC and ITU Filings
« Reply #589 on: 02/24/2023 09:37 pm »
Considering the FCC is limiting uplinks to a min elevation of 25 degrees. Even with a full buildout would, you ever have 32 satellites in the sky at the same time? The cone ignoring earths curvature is ~2000 miles in diameter... but then again you will have 19 ground stations with overlapping coverage.

Though, during severe weather it would allow them to route the E-Band data to different ground stations outside of the storm without resorting to ISL.

If you're happy and you know it,
It's your med's!

Online JayWee

  • Full Member
  • ****
  • Posts: 1005
  • Liked: 1002
  • Likes Given: 1896
Re: Starlink : New FCC and ITU Filings
« Reply #590 on: 02/24/2023 10:30 pm »
Considering the FCC is limiting uplinks to a min elevation of 25 degrees. Even with a full buildout would, you ever have 32 satellites in the sky at the same time?

One article mentioned they are using both polarities (at least for V1). So 32 beams can mean 16 uplinks.
« Last Edit: 02/24/2023 10:31 pm by JayWee »

Offline Roga

  • Member
  • Posts: 34
  • Liked: 48
  • Likes Given: 14
Re: Starlink : New FCC and ITU Filings
« Reply #591 on: 02/25/2023 04:03 am »
I looked into this a bit and there are some corner cases that have a point. For dark sky astronomy it is largely overblown. The satellites are visible but they are dim enough that they form a point on CMOS and CCD image sensors and the streaks can be removed in post processing.

However there are some projects that rely on low- elevation angle observations around twilight. The obvious example is NEO surveys. Most asteroids that cross Earth orbit are by definition much more likely to be observed at twilight. The sensors are also tuned to low contrast - dim streaks against a relatively bright sky. This means that a satellite that hasn't crossed the terminator can shine bright enough to saturate sensor pixels and capacitive couple into adjacent pixels causing fringing. This kind of effect is really hard to post process out - maybe you erase the first couple fringes,  but at some point you are left with a fringe that looks a lot like a NEO.

I get the argument that Starlink/ship will eventually enable space observatories that render Earth based observatories obsolete. But I also understand that that is maybe half a typical astronomer's career away, and the affected people have a legitimate gripe.

Offline vsatman

Re: Starlink : New FCC and ITU Filings
« Reply #592 on: 02/27/2023 07:16 pm »
Does that take into account both polarization? Based on the Gateway V3 filings, they hit 8 Bits/hz using both polarizations (16Gbps for 2Ghz), or 6.4 Bit/hz typical. There might be losses due to self interference and atmospheric attenuation, so maybe knock off 25-50% for a more realistic number... If you work that out:
51 Gateways * 32 Beams = 1,632 Beams
1632 * 5Ghz = 8,160Ghz
8,160Ghz  * 8 Bits/s / Hz best case = 65.3 Tbps
8,160Ghz  * 6.4 Bits/s / Hz typical = 51 Tbps

Add in the already approved Ka antennas:
888 Beams * 2Ghz = 1,776 Ghz
1,776Ghz * 8 Bits/Hz best case = 14.2Tbps
1,776Ghz  * 6.4 Bits/s / Hz typical = 11.4Tbps

Total of 62.4 to 79.5Tbps for the CONUS, or 2.9Gbps to 3.75Gbps per cell downlink, plus a bit less uplink.

In total, with V-Band they applied for 10.8Ghz Sat to GW, and 10.1Ghz GW to Sat (plus another 500Mhz in each direction the FCC deferred on); 8.35Ghz Sat to UT, and 5.25Ghz UT to Sat.

Unfortunately, your calculations are incorrect, the E and Ka bands are satellite-GW links, and nothing changes on the Sat-UT link, as well as the Ku band and channel has  240 MHz. Considering that the terminal works only in 1 polarization, the maximum that is possible is 8 channels at different frequencies per 1 cell. But the greater the throughput of the GW-Sat link, the more cells can be served by one satellite - this is undoubtedly a plus!

Offline Robotbeat

  • Senior Member
  • *****
  • Posts: 39286
  • Minnesota
  • Liked: 25283
  • Likes Given: 12125
Re: Starlink : New FCC and ITU Filings
« Reply #593 on: 02/27/2023 11:56 pm »
Wtf is “NCo=2”
Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

Offline russianhalo117

  • Global Moderator
  • Senior Member
  • *****
  • Posts: 8790
  • Liked: 4709
  • Likes Given: 768
Re: Starlink : New FCC and ITU Filings
« Reply #594 on: 02/28/2023 12:04 am »
Wtf is “NCo=2”
All i know is that it is in economics known as the Net Capital Outflow formula NCo=N which breaks down into two sub formulas for import and export (NCo=NI and NCo=NX).
« Last Edit: 02/28/2023 12:08 am by russianhalo117 »

Offline Robotbeat

  • Senior Member
  • *****
  • Posts: 39286
  • Minnesota
  • Liked: 25283
  • Likes Given: 12125
Re: Starlink : New FCC and ITU Filings
« Reply #595 on: 02/28/2023 01:04 am »
Wtf is “NCo=2”

NCo is the “Number of co-frequency signals” (aka beams), and the number refers to the limit of these that can be placed on a single spot on the ground. Gen1 is approved for NCo=1, meaning they can’t have 2 signals of the same frequency on the same spots on the ground (even if it’s below the ePFD limits). Gen2 is also approved for NCo=1 in Ku band, and the FCC has acknowledged the combined Gen1+ Gen2 constellations reach an effective NCo=2.

Given SpaceX has approval to use ~2Ghz of Ku spectrum for user downlink beams, that works out to 8 separate 240Mhz chunks (with 10Mhz guard bands). 1 of these is largely unusable due to radio astronomy protection, so 7 useable channels. Technically a sat from Gen1 and a sat from Gen2 could each beam 7x240Mhz onto a single cell, for a total of 14 beams. There’s also a whole polarization discussion, but I’m not well-versed enough to go into that.
Ah, okay. Makes sense.

It's my understanding that, because Starlink is using phased arrays on both sides, they could actually have NCo>>1, i.e. using spatial multiplexing to have basically as many beams on the same spot on the ground as there are satellites (to first order) as you can use the phased array to pick out whichever beam you want for amplification (the phased array elements are phased such that, say, they point to 45deg, 45 deg (+/-5 deg), avoiding the signal from 45deg 30deg--to the limits of the angular resolution of the phased array). And you're telling me that currently they only have approval for only up to NCo = 2.

Is this correct? Obviously, trying to do spatial multiplexing like this increases the noise, but overall should allow a lot more bandwidth to places that otherwise would be saturated. I did my physics senior thesis project on something like this, but I don’t understand the regulations around it.

(And polarization allows up to a doubling of bandwidth as you could just pick out one of two polarizations.)
« Last Edit: 02/28/2023 01:17 am by Robotbeat »
Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

Online DanClemmensen

  • Senior Member
  • *****
  • Posts: 5705
  • Earth (currently)
  • Liked: 4465
  • Likes Given: 1865
Re: Starlink : New FCC and ITU Filings
« Reply #596 on: 02/28/2023 01:42 am »
Wtf is “NCo=2”

NCo is the “Number of co-frequency signals” (aka beams), and the number refers to the limit of these that can be placed on a single spot on the ground. Gen1 is approved for NCo=1, meaning they can’t have 2 signals of the same frequency on the same spots on the ground (even if it’s below the ePFD limits). Gen2 is also approved for NCo=1 in Ku band, and the FCC has acknowledged the combined Gen1+ Gen2 constellations reach an effective NCo=2.

Given SpaceX has approval to use ~2Ghz of Ku spectrum for user downlink beams, that works out to 8 separate 240Mhz chunks (with 10Mhz guard bands). 1 of these is largely unusable due to radio astronomy protection, so 7 useable channels. Technically a sat from Gen1 and a sat from Gen2 could each beam 7x240Mhz onto a single cell, for a total of 14 beams. There’s also a whole polarization discussion, but I’m not well-versed enough to go into that.
Ah, okay. Makes sense.

It's my understanding that, because Starlink is using phased arrays on both sides, they could actually have NCo>>1, i.e. using spatial multiplexing to have basically as many beams on the same spot on the ground as there are satellites (to first order) as you can use the phased array to pick out whichever beam you want for amplification (the phased array elements are phased such that, say, they point to 45deg, 45 deg (+/-5 deg), avoiding the signal from 45deg 30deg--to the limits of the angular resolution of the phased array). And you're telling me that currently they only have approval for only up to NCo = 2.

Is this correct? Obviously, trying to do spatial multiplexing like this increases the noise, but overall should allow a lot more bandwidth to places that otherwise would be saturated. I did my physics senior thesis project on something like this, but I don’t understand the regulations around it.

(And polarization allows up to a doubling of bandwidth as you could just pick out one of two polarizations.)
In a reasonable system SpaceX will only hit any one spot on the Earth's surface using the same frequency with multiple satellites only if they are far enough apart in angular distance. Ten degrees of angular separation will almost certainly provide more than 20 dB of discrimination for a consumer-grade phased array. However, an antenna will have less gain for a satellite that is farther away from zenith, so in general they will use the "best" satellite for a spot until that spot congests that beam before using the less-well-positioned satellites. This best satellite will permit more efficient use of the bandwidth (more bits/Hz). You only use the less-well-positioned satellites when the "best" satellite cannot satisfy the demand in the spot. (This is theory. I have no information about the actual Starlink architecture.)

Offline gongora

  • Global Moderator
  • Senior Member
  • *****
  • Posts: 10309
  • US
  • Liked: 14121
  • Likes Given: 6029
Re: Starlink : New FCC and ITU Filings
« Reply #597 on: 02/28/2023 11:01 pm »
SpaceX filed for another generation of their Ku-band user terminals (they seem to refresh them fairly often).  Filed for normal (.0725 sq m) and high performance (.2052 sq m) versions of both fixed and ESIM.


Offline vsatman

Re: Starlink : New FCC and ITU Filings
« Reply #598 on: 03/01/2023 07:19 pm »

Given SpaceX has approval to use ~2Ghz of Ku spectrum for user downlink beams, that works out to 8 separate 240Mhz chunks (with 10Mhz guard bands). 1 of these is largely unusable due to radio astronomy protection, so 7 useable channels. Technically a sat from Gen1 and a sat from Gen2 could each beam 7x240Mhz onto a single cell, for a total of 14 beams. There’s also a whole polarization discussion, but I’m not well-versed enough to go into that.
I understand a little about polarization and most importantly, according to the documents sent by SpaceX to the FSS, the Starlink user terminal only works in one (right) polarization.
There is one more problem - that the shape of the practical cell is always not round with a diameter of 15 miles, but an ellipse with a long side of up to 25 miles, and it also covers neighboring cells, which must be taken into account when planning the distribution of links from satellites to cells, so as not to violate the rule N=1.

Offline vsatman

Re: Starlink : New FCC and ITU Filings
« Reply #599 on: 03/01/2023 07:32 pm »
The issue of frequency reuse has long been investigated and resolved both in cellular and satellite communications - this is the 4 color rule. It says that if you have circles of 4 different colors, then you can cover the surface of any size with them, and circles of the same color will never be neighbors. Based on the fact that Space X has 8 beams of 512 MHz in the allowed Ku band, the maximum number of beams that theoretically can serve 1 cell with continuous coverage of any territory on Earth is 2.

Tags:
 

Advertisement NovaTech
Advertisement Northrop Grumman
Advertisement
Advertisement Margaritaville Beach Resort South Padre Island
Advertisement Brady Kenniston
Advertisement NextSpaceflight
Advertisement Nathan Barker Photography
1