10 ms·

Ehm. You need a precise clock for any form of GNSS[0] to function. AFAIR StarLink satellites doesn't have precise clock on the board. And anyway, modern GNSS c

Ehm. You need a precise clock for any form of GNSS[0] to function. AFAIR StarLink satellites doesn't have precise clock on the board. And anyway, modern GNSS chips are.. $1? Cheaper? No way you wouldn't use GNSS already deployed and working (and you are not paying for it) in a commercial product. [0] GPS is GNSS, it's just in America GPS became the synonym for GNSS

cduzz 3y ago
So -- for the purposes of locating a base station that's connected to the starlink network -- the starlink network would rely on the base station reporting it's location? Hopefully not... It certainly doesn't need to just trust it when it has direct observation available too. I'm not suggesting that whatever GPSish thing you'd be able to get out of a pure starlink protocol communication would give you super resolution, but it'd certainly be enough to identify "Crimea" vs "Indian Ocean". And as far as base station -> constellation -- putting a GPS receiver in it is of course smart to speed that up, but I'm not sure even that's necessary or just helpful.
skissane 3y ago
> You need a precise clock for any form of GNSS[0] to function How precise? I assume that inaccuracy in the clock produces imprecision in location - how much time inaccuracy produces how much location inaccuracy? > And anyway, modern GNSS chips are.. $1? Cheaper? No way you wouldn't use GNSS already deployed and working (and you are not paying for it) in a commercial product. Rather than using the Starlink satellites as an alternative to GPS - get a location fix using both Starlink and GPS, and compare them. Even given the former is going to be significantly less accurate - suppose GPS is accurate to within 5 metres and Starlink is accurate to within 5000 metres - well, if the two differ by 50,000 metres, you know something is going wrong, and the terminal might decide to disable itself in such a situation.
- justsomehnguy 3y ago
  > How precise GNSS satellites orbits are between 20000km and 30000km. 20kkm is 0.06671s of c, 30kkm is 0.10007s, 5km is 0.000016678. So a very precise, that's why they have an atomic clock on board instead of cheap quartz one. > if the two differ by 50,000 metres, you know something is going wrong Like what your imprecise methods give you a very rough estimate where is the terminal and your legitimate customers in the nearby areas (not even doing anything for or against) are saying thanks for their shut off terminals. https://upload.wikimedia.org/wikipedia/commons/b/b4/Comparison_satellite_navigation_orbits.svg https://upload.wikimedia.org/wikipedia/commons/b/b4/Comparis... https://www.omnicalculator.com/physics/light-year https://www.omnicalculator.com/physics/light-year

cduzz 3y ago

So -- for the purposes of locating a base station that's connected to the starlink network -- the starlink network would rely on the base station reporting it's location? Hopefully not... It certainly doesn't need to just trust it when it has direct observation available too. I'm not suggesting that whatever GPSish thing you'd be able to get out of a pure starlink protocol communication would give you super resolution, but it'd certainly be enough to identify "Crimea" vs "Indian Ocean". And as far as base station -> constellation -- putting a GPS receiver in it is of course smart to speed that up, but I'm not sure even that's necessary or just helpful.

skissane 3y ago

> You need a precise clock for any form of GNSS[0] to function How precise? I assume that inaccuracy in the clock produces imprecision in location - how much time inaccuracy produces how much location inaccuracy? > And anyway, modern GNSS chips are.. $1? Cheaper? No way you wouldn't use GNSS already deployed and working (and you are not paying for it) in a commercial product. Rather than using the Starlink satellites as an alternative to GPS - get a location fix using both Starlink and GPS, and compare them. Even given the former is going to be significantly less accurate - suppose GPS is accurate to within 5 metres and Starlink is accurate to within 5000 metres - well, if the two differ by 50,000 metres, you know something is going wrong, and the terminal might decide to disable itself in such a situation.

justsomehnguy 3y ago
> How precise GNSS satellites orbits are between 20000km and 30000km. 20kkm is 0.06671s of c, 30kkm is 0.10007s, 5km is 0.000016678. So a very precise, that's why they have an atomic clock on board instead of cheap quartz one. > if the two differ by 50,000 metres, you know something is going wrong Like what your imprecise methods give you a very rough estimate where is the terminal and your legitimate customers in the nearby areas (not even doing anything for or against) are saying thanks for their shut off terminals. https://upload.wikimedia.org/wikipedia/commons/b/b4/Comparison_satellite_navigation_orbits.svg https://upload.wikimedia.org/wikipedia/commons/b/b4/Comparis... https://www.omnicalculator.com/physics/light-year https://www.omnicalculator.com/physics/light-year