[Starlink] Starlink power use & satellite tracking

Ulrich Speidel u.speidel at auckland.ac.nz
Sat Feb 18 07:36:42 EST 2023


On 18/02/2023 11:52 pm, David Lang wrote:
> On Sat, 18 Feb 2023, Ulrich Speidel via Starlink wrote:
>
>> This requires the satellite to be in view of both terminal and 
>> gateway. There is, however, no reason why a user's packets could not 
>> travel via a diversity of satellites.
>
> In theory you are correct, but I don't think they are doing that yet. 
> We had a discussion of the signaling protocol and someone posted that 
> they had done research and found that there was only one satellite 
> illuminating a cell at a time. IIRC this was when the paper came out 
> on using starlink signals for positioning.

Given that the number of satellites has effectively doubled over the 
last year, does this still hold? For handover purposes, dishy must be 
tracking multiple satellites anyway, and exchanging successive packets 
with different satellites shouldn't be all that difficult then.

I've just run a series of 20 speedtest.net test and tried to spy the 
upload / download power usage half-way through each test. The result 
shows a significantly stronger correlation between power use and 
download final rate than between power use and upload final rate. Of 
course, it's a small sample, and I'll try to do a few more tomorrow.

Of course, if it receives from multiple birds at the same time, then 
that would keep its signal processor a little busier than if it only 
received from one. So maybe it's that. Or maybe we're looking at one 
bird with multiple channels/bursts, also imposing a higher compute load?

Also intriguing: In the past, we used to see typical power usage in the 
80-100 W range with the thing, maybe as low as 60 W if I caught it 
lucky. But now it's within the official specs of 40-75 W most of the 
time, and it's changed this week.

>
> As they launch more satellites and get the in-space routing up and 
> running, I expect that they will change that at some point.
>
> it is an interesting question.
>
> David Lang
>
> _______________________________________________
> Starlink mailing list
> Starlink at lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink
>
-- 
****************************************************************
Dr. Ulrich Speidel

School of Computer Science

Room 303S.594 (City Campus)

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





More information about the Starlink mailing list