[net.general] delays in phone propagation

davej@uw-vlsi.UUCP (10/13/83)

from Victor J. Slabinski, Astrodynamics Department, COMSAT:
	"I wish to inform [you] that AT&T often routes domestic long
	 distance phone calls via the COMSTAR communications satelite.
	 In that case the time signal travels approximately 38,000 km
	 from the earth to the satelite and another 38,000 km back to
	 earth with a resulting propagation path delay (uplink plus
	 downlink) of 250 milliseconds. ... Long-distance callers to
	 the Master Clock should be aware of this possible path delay
	 when setting their clocks."
If your company or institution subscribes to a non-AT&T WATS service,
the delays can be equivalent or longer.  The time delay is VERY evident
in coast-to-coast calls, and makes communications with the partry at the
other end difficult and frustrating!

@pesnta.UUCP (10/15/83)

Can anyone out there in netland tell me definitively how much delay
might be encountered when calling the recently announced 800 number for
WWV?  I've heard that there can be as much as (or more than!) several
seconds, depending on the routing of the call.  This would make the
accuracy of such a service extremely questionable at best.  Lauren, do
you have the answer?
-- 
	Marty Shannon
UUCP:	{alice,rabbit,research}!mjs
Phone:	201-582-3199

@pesnta.UUCP (10/15/83)

Regardless of the routing of your call (which may take the several seconds
you describe), the phone communication is approximately light-speed.  Yes,
there are finite delays involved in the switching circuitry, but remember
back to the most recent coast-to-coast phone call you made.  Did you notice
a big lag in talking to the other person?  I'll warrant not.  As far as I
know, the WWV dial-in is located in the continental US, (and probably some
place centrally located), so at least you will have a delay of

	2,000mi / 186,000mi/sec = .01 sec + electronics delay

(Or not very much delay at all).

	-Dan Klein, Mellon Institute, Pittsburgh

@pesnta.UUCP (10/15/83)

	Regardless of the routing of your call (which may  take  the
	several  seconds  you  describe), the phone communication is
	approximately light-speed.  Yes,  there  are  finite  delays
	involved  in  the  switching circuitry, but remember back to
	the most recent coast-to-coast phone call you made.  Did you
	notice  a  big  lag  in  talking  to the other person?  I'll
	warrant not.  As far as I know, the WWV dial-in  is  located
	in  the  continental  US, (and probably some place centrally
	located), so at least you will have a delay of

		2,000mi / 186,000mi/sec = .01 sec + electronics delay

	(Or not very much delay at all).

		-Dan Klein, Mellon Institute, Pittsburgh

This is all essentially true, at least from the viewpoint of physics
(ignoring the fact that the speed of propagation of a signal in wire
is closer to 120,000mi/sec).  The problem is that I think the telephone
companies have started using satellites, which I think packet switch,
which could lead to delays on the order of seconds.  I don't know how
many telephone companies use satellites for intracontinental traffic,
but I'm pretty sure that at least some do.  I would expect this to
become more widespread as the cost of satellites decreases with respect
to the cost of maintaining long land lines or microwave links.

Before anybody flames; yes, I know, the difference in delay caused by
propagation at c versus the real propagation rate is not that much, 35%
to be exact, since I used a velocity constant of .65, meaning that the
above .01 + K (K = electronic delay, above) delay becomes .017 + K.  It's
K, and the magnitude of K as caused by satellites that I'm commenting on.

Rick Wise
ucbvax!decvax!harpo!seismo!rlgvax!cal-unix!wise

donr@hp-pcd.UUCP (10/21/83)

#R:rabbit:-203200:hp-kirk:8300005:000:288
hp-kirk!donr    Oct 19 12:12:00 1983

As an altertnative to the phone in system:
  Certain (unnamed to follow net protacol) consumer electronics companies
sell radio clocks which can recieve WWV, or WWVH and provide interfaces
to computer equipment.  The propagation delay is then fixed by your distance
from the transmitter.

wm@tekchips.UUCP (Wm Leler) (10/24/83)

If you are concerned about what the exact time is, find a copy
of the new heathkit catalog.  They have a "most accuate clock"
which synchs itself to WWV (or WWVH), will compensate for
propagation delay, and can even display UTC1 time (which
compensates for the wobble in the earth!)  The clock also
has an RS-232C port option so you can connect it to your
computer.

			Wm Leler
			wm.Tektronix@Rand-relay