[ntp:questions] Questions about joining pool.ntp.org

unruh unruh at wormhole.physics.ubc.ca
Tue Aug 30 01:50:34 UTC 2011

On 2011-08-30, jimp at specsol.spam.sux.com <jimp at specsol.spam.sux.com> wrote:
> Chris Albertson <albertson.chris at gmail.com> wrote:
>> On Mon, Aug 29, 2011 at 1:46 PM, <jimp at specsol.spam.sux.com> wrote:
>>> Harlan Stenn <stenn at ntp.org> wrote:
>>> > GPS can be done very affordably and can offer great time.  There are
>>> > several *potential* pitfalls:
>>> >
>>> > - It is *possible* for the US Gov't to detune the GPS system (locally or
>>> >  in-general).  Since GPS is now increasingly used for "human safety"
>>> >  things, the costs/risks for doing this have gone up significantly so
>>> >  this risk may now be more of a threat than a reality
>>> GPS aviation navigation is now so deeply entrenched that if you live near
>>> a major airport the US Gov't "detunes" GPS, you will have much bigger
>>> things to worry about than the current time and should start filling
>>> things with water...
>> In the context of an NTP server any "detuning" of GPS will not matter.
>> It's called "Selective Availability" and the way it works (in short) is
>> they encrypt the lower few bits.   Or in English, they scramble least
>> significant decimal points so that GPS users who don't have the "secret key"
>> see much reduced precision.   But how does this effect TIME?   Remember that
>> light travels at about 1000 feet per microsecond.  So if you GPS is
>> "detuned" such that it is seeing a random 1,000 foot error then the time is
>> also "detuned" by about one microsecond.   NTP will hardly notice.
>> The reason this may seem counter intuitive is that it is hard to picture in
>> your mind the incredible time precision we get using GPS.  In normal use the
>> error is about one part in a hundred million.  If they "detune" it so as to
>> make the error 1000 times larger we still get less error than NTP can track.
> But then there goes the 10 ns accuracy.
> How could anyone survive with only microsecond accuracy?
>> The thing to worry about with GPS is if the system were totally shut down or
>> (more likely) if it was jammed by some local radio transmitter.   The common
>> way you address this (and all cell towers have to address this.  I think the
>> requirement they have is to handle a 24 hour loss of GPS) is to have a good
>> stable local oscillator.  They use either an ovenized crystal or
>> rubidium oscillator.  Sounds exotic but the price of good surplus is about
>> $100 to $150.     In simple terms the GPS keeps the oscillator running at
>> correct speed and then if GPS goes away the oscillator is stable enough that
>> it will serve as a clock until GPS comes back.  Cell towers need to track
>> time at the uS level and they all work like this and can work for a long
>> time with no GPS signal. (for more info google the term "gps holdover".
>> That is the term for keeping time when GPS is down)
>> Chris Albertson
>> Redondo Beach, California
> It would be more realistic to worry about a swarm of locusts and a plague
> than GPS going away for 24 hours absent WWIII, in which case you will have
> much bigger things to worry about than the current time at any accuracy.
> As far as being jammed goes, the FCC is rather sensitive and very aggressive
> about the subject as LightSquared has discovered.

Have they discovered that? I thought it was still up in the air.
And we keep hearing about the UK jamming GPS for hours at a time in
regions of the UK. 


More information about the questions mailing list