[Starlink] Ka vs Ku Band, Signal Angle, and Weather Impact

[Starlink] Ka vs Ku Band, Signal Angle, and Weather Impact

[Colin_Higbie]

Ulrich,

I know we see significant rain fade with geostationary satellites, which I have long assumed is at least in part because from our latitude (around 44 degrees north), the angle to a geostationary satellite is so small that it's going nearly horizontally through hundreds of miles of clouds. In contrast, Starlink satellites, nearly overhead, punch almost vertically straight through the clouds. This means it has far fewer water droplets and clouds to pass through to reach a satellite in the same weather. I assume this is at least partially responsible for why Starlink is VASTLY more reliable at holding a connection in bad weather than geostationary links. 

First, is that correct?

Second, does this have any bearing on your point about Ka not being good in the rain? I.e., maybe it's not as good in the rain as Ku, but because of the angle of communication and therefore reduced signal attenuation, it can still get through typical cloud cover and moderate rain, still "good enough"?

While obviously none of us want to lose connectivity or signal in the rain, I'd rather drop to some fractional capacity (whatever fits on Ku) during occasional bad storms, if it meant that the rest of the time, I could have much more bandwidth with added Ka support. But I acknowledge that I don't know how all these factors interact for final results.

Thanks for any info/education on this,
Colin


Date: Sun, 24 Nov 2024 16:39:28 +1300
From: Ulrich Speidel <u.speidel@auckland.ac.nz>
To: Dave Taht <dave.taht@gmail.com>, Michael Richardson
	<mcr+ietf@sandelman.ca>
Cc: starlink@lists.bufferbloat.net
Subject: Re: [Starlink] Starlink Internet Speeds Could Skyrocket to 2
	Gigabits Per Second, SpaceX President Says
Message-ID: <78ba68e7-11a7-4355-9264-b7ff16afab84@auckland.ac.nz>
Content-Type: text/plain; charset=UTF-8; format=flowed

On 24/11/2024 9:05 am, Dave Taht wrote:
>> 3. Elon will own the FCC.
> This FCC promises to open up more spectrum. Where?

I think it's not so much a matter of opening up spectrum. SpaceX's 
spectrum "reserve" at the moment is Ka-band. That's hitherto been used 
for gateways only, but with the lasers, that's becoming less of a 
bottleneck because you can in principle uplink from and downlink to a 
remote gateway or set of remote gateways (unlike with the end user - I 
mean Elon can't simply provide the service Dave's paying for over, say, 
New York, rather than where Dave's boat is). So that frees up a bit of 
resource to deploy directly to users in tight spots. Except that Dishy 
doesn't talk on Ka.

The trouble with Ka is that it's a bit more susceptible to rain fade 
than Ku, and also that the amount of spectrum available there isn't 
exactly an order of magnitude larger either. To the contrary - the three 
Ka-bands in which SpaceX can downlink data have a combined 1.8 GHz of 
spectrum - less than for Ku. The pros of Ka are that the shorter 
wavelength makes for sharper beams and a nicer link budget with the same 
size antennas. As long as you don't get rained on. So you better throw 
in some extra antenna gain, which is what they're doing with the 1.85 m 
dishes (not Dishys) at their community gateways. But that's an entirely 
different business model.

And yes I'm aware that there's some labelling issues as to what counts 
as Ka and what doesn't, internationally, and that the FCC's notion of Ka 
starts at lower frequencies than some other peoples'.

-- 
****************************************************************
Dr. Ulrich Speidel

School of Computer Science

Room 303S.594 (City Campus)

The University of Auckland
u.speidel@auckland.ac.nz
http://www.cs.auckland.ac.nz/~ulrich/
****************************************************************

Re: [Starlink] Ka vs Ku Band, Signal Angle, and Weather Impact

[Ulrich Speidel]

[-- Attachment #1: Type: text/plain, Size: 4683 bytes --]

There's a bit to unpack here, so see below.

On 25/11/2024 6:59 am, Colin_Higbie wrote:
> I know we see significant rain fade with geostationary satellites, which I have long assumed is at least in part because from our latitude (around 44 degrees north), the angle to a geostationary satellite is so small that it's going nearly horizontally through hundreds of miles of clouds. In contrast, Starlink satellites, nearly overhead, punch almost vertically straight through the clouds. This means it has far fewer water droplets and clouds to pass through to reach a satellite in the same weather. I assume this is at least partially responsible for why Starlink is VASTLY more reliable at holding a connection in bad weather than geostationary links.
>
> First, is that correct?

Partly. Clouds don't reach more than about 10 km in normal 
circumstances, with thunderstorms sometimes going up to about 20 km. At 
44 deg north, you're just over 3000 km above the equatorial plane 
(ballpark), and the geostationary orbit sits around 39,000 km away from 
the point where your location projects onto the equatorial plane. Assume 
a GEO sat on the same longitude as you for a start. Do an arctan on 
3,000/39,000 and you get the angle between a geostationary sat's line to 
you and the equatorial plane, about 4.4 deg. Deduct this from 90 deg and 
you get the angle between the line between you and the sat and the 
projection from your location to the equatorial plane. So about 85.6 
deg. The angle between the tangential surface plane at your location and 
the line of projection onto the equatorial plane is 90 deg minus your 
latitude, so 46 deg. Subtract that from the 85.6 deg and you get the 
highest elevation of the geostationary arc as viewed from your location. 
Makes 39.6 deg.

Assume we're working with a 20 km path through a thunderstorm cloud 
straight up. Now going at an elevation of 39.6 deg instead of 90 deg 
through a 20 km cloud layer results in a path of under 32 km. Not 
hundreds of miles. Now that path stretches a bit obviously if you are 
aiming your antenna at satellites that are not on the same longitude as 
you, and that's where you could get into the hundreds of miles 
potentially if they're at a longitude that's far off yours.

Noting here that Starlink talks upwards of an elevation of 25 degrees - 
so there's at least some overlap between LEO and GEO domain here in 
terms of potential path lengths through rain clouds.

The reason why GEO connections struggle compared to LEO is the path loss 
over the distance - a factor of about 1000 in terms of power, order of 
magnitude, over Starlink. This is why GEO either needs very powerful 
transmitters in space (sat TV) or big antennas with high gain on the 
ground (everything else pretty much). High gain antennas also tend to be 
pretty sensitive in terms of directionality. If you can't track 
electronically as Dishys do, then anything (wind, atmospheric 
refraction) that knocks the beam off course means lower signal at the 
receiving end and more trouble keeping connected.

Basically, rain fade gets worse the higher you go in frequency - so Ka 
is worse than Ku, no matter whether it's a GEO or a LEO link. That's why 
C band was more popular than Ku until it filled up. And now you have Ku 
being more popular than Ka but it's also filling up.

> Second, does this have any bearing on your point about Ka not being good in the rain? I.e., maybe it's not as good in the rain as Ku, but because of the angle of communication and therefore reduced signal attenuation, it can still get through typical cloud cover and moderate rain, still "good enough"?

The standard technical response to rain fade (regardless of band) is to 
change modulation when it gets bad. So you trade bit rate against 
robustness by downgrading from, say 64QAM to 16QAM or somesuch.

Or you can up transmit power.

> While obviously none of us want to lose connectivity or signal in the rain, I'd rather drop to some fractional capacity (whatever fits on Ku) during occasional bad storms, if it meant that the rest of the time, I could have much more bandwidth with added Ka support. But I acknowledge that I don't know how all these factors interact for final results.

The main issue with "added Ka support" is that Dishy isn't dual-band, 
and I'm not sure if or when we'll ever see a Ka-band Dishy. Period.

-- 
****************************************************************
Dr. Ulrich Speidel

School of Computer Science

Room 303S.594 (City Campus)

The University of Auckland
u.speidel@auckland.ac.nz 
http://www.cs.auckland.ac.nz/~ulrich/
****************************************************************



[-- Attachment #2: Type: text/html, Size: 6434 bytes --]

Re: [Starlink] Ka vs Ku Band, Signal Angle, and Weather Impact

[Michael Richardson]

[-- Attachment #1: Type: text/plain, Size: 519 bytes --]


Colin_Higbie via Starlink <starlink@lists.bufferbloat.net> wrote:
    > While obviously none of us want to lose connectivity or signal in the
    > rain, I'd rather drop to some fractional capacity (whatever fits on Ku)
    > during occasional bad storms, if it meant that the rest of the time, I
    > could have much more bandwidth with added Ka support. But I acknowledge
    > that I don't know how all these factors interact for final results.

I feel that there is a Jurassic Park-ish disaster movie plot here.


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 511 bytes --]