[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Time codes for PCs (fromn German Banking)



At 11:42 PM 1/29/96 +0100, [email protected] wrote:
>From:	SMTP%"[email protected]" 27-JAN-1996 03:43:05.83
>
>>A peripheral I've long wanted to see, commonly available:  ACCURATE  time, 
>>broadcast to the millisecond/microsecond/nanosecond, available from sources 
>>as varied as TV VIR's, FM subcarriers, and other sources, available as an 
>>easy input (via a peripheral card) to a computer.
>>
>	Yup! Do you think it is really possible? If I remember well
>speed of light is 300.000 Km/s. That means that light takes around
>1 ms. to cover 300 Km. If you use a satellite, antenna, whatever
>to broadcast a timing signal, the accuracy will depend on when
>do you receive it, and that in turn on your distance from the
>source.

<sigh!>  I am well aware of the facts you indicate, and many more of which
you aren't even aware.  But one of the functions (in fact, the basic,
intended one) of GPS is to locate, as precisely as possible, the exact
location of the receiver.  Thus, time delays can be compensated to the
accuracy of the location fix, at the very least.  If we assume 100 meters
error, max, that's about 300 nsec error.  (GPS receivers AUTOMATICALLY
compensate for such delays!)


>>I have a 12-year-old Heathkit "Most Accurate Clock" that I assembled myself, 
>>and had the foresight to install it with its computer  interface option. 
>>(receives 5, 10, or 15 MHz signals broadcast from Boulder, Colorado, 
>>containing "exact" time.) 
>>
>	Just remember that the best you can get would be microseconds if
>you're in a 300 meter radius, or milliseconds on a 300 Km. And possibly
>nanoseconds at 0.3 m.

Well, this clock doesn't pretend to be better than about 5-10 milliseconds
even in signal-locked condition.  However, there is a dipswitch on the
bottom of the unit, settable in 500 mile increments, from Boulder, Colorado.
500 miles corresponds to 3 milliseconds.  Clearly this was a good device
when made, but has obviously been supplanted by at least a factor of 1000 by
GPS.


>	Then remains the cypherpunk part on all this: how can you
>trust the *signal* your receptor receives? How do you know no one is
>interferring it or sending an inaccurate or false one?

Simple answer:  You don't.  More complicated answer:  Most such devices
don't merely input the time signal, but they use an accurate internal clock
to maintain good time when signals go away.  In fact, the best units
"discipline" the local oscillator, either actually changing its frequency or
at least following its errors over time.  The result is that sudden errors
would be noticed.  A good TCXO is stable to well better than 1 ppm.


>	And that on a broadcast system. A system owned by someone who
>you may not trust (say a private TV channel, radio or satellite). So
>you may want to have several sources, and to be able to verify that
>the signals you receive all come from their respective sources.

Well, since "everybody" in one locale can receive a particular (local) TV
channel, one solution might be to compare time with respect to the beginning
of a particular scan line..  You may not trust the signal, but you know its
a signal that "everybody" receives.  There may be no other provision for
adding time to it, but a relatively low accuracy crystal oscillator could
identify particular frames, etc.


>	Yum! a nice problem to think about. One factor is that you
>wouldn't expect changes in public sources used by sensible systems
>since those could not pass unnoticed and might raise big protests.
>But you still have the MITM attack to consider...