[time-nuts] Best reason

Jim Lux jimlux at earthlink.net
Wed Mar 28 13:30:19 UTC 2012


On 3/28/12 1:28 AM, Jim Palfreyman wrote:
> So when a member of the general public says:
>
> Why do we need really accurate clocks?
>

Not necessarily at "time-nuts" performance, but here's some clock and 
clock distribution applications

There are a lot of systems out there that timestamp something going in, 
and somewhere else, it gets timestamped going out, and the difference 
determines how much you pay.

Parking lots are a good example. You need to have all the timeclocks 
synchronized, so that people are charged the right amount

Automated speeding ticket generators on tollways compare timestamps of 
passage through the toll station with measured distance to determine 
average speed.


===

What's interesting is that a lot of things inherently depend on ppm 
kinds of timing accuracy that you wouldn't expect.. 10ppm is 1 second 
per day drift, and most inexpensive quartz oscillators aren't that good. 
  50-100 ppm is more typical.  While  1 second isn't a big deal, but 30 
seconds over a month might be, and 5 minutes in a month is a problem 
(you'd hate it if your DVR started recording your program 30 seconds 
after it started)





More information about the Time-nuts_lists.febo.com mailing list