[time-nuts] synchronizing a large number of weakly coupled oscillators
alan.melia at btinternet.com
Wed May 12 23:41:36 UTC 2010
Well I guess "no" because accuracy is the deviation from a known standard (I
think) Stability repeatability might be better but you need to consider what
the variables might be. Variations in thickness (basically frequency), cut
angle (temp coeficient and maybe others), crystal purity (aging, ESR ?).
If you average many randomly selected samples you might reduce the level of
variablilty of these aspects but would that make them "more accurate"? I
doubt that maybe more capable of staying within a given accuracy once
I still think the cost effective way is get one "good" one, the best you can
afford, characterise it and the make adjustments either calculated or by
disciplining. Even an ordinary crystal can be made to perform quite well by
adjusting it to track it to something better. Many LF BC stations in Europe
are much better than a cheapy (computer grade) crystal and Droitwich and
Allouis are locked to a Rb standard and regularly measure against the
national standards. A few part in 10^11 costs a couple of hundred
dollars.This is 5 orders better than a cheapy crystal. My back of envelope
calculation suggest you might need about 100,000 oscillators to achieve this
level (ok tell me I'm wrong with the calculation and, as the exam script
says, "show you working" .....the red wine was very nice.... I can take it
!! :-)) )
----- Original Message -----
From: "Hal Murray" <hmurray at megapathdsl.net>
To: "Discussion of precise time and frequency measurement"
<time-nuts at febo.com>
Sent: Wednesday, May 12, 2010 11:31 PM
Subject: Re: [time-nuts] synchronizing a large number of weakly coupled
> > Presumably, a large population of cheap coupled oscillators could be
> > accurate collectively.
> There are 2 main sources of error in inexpensive crystal oscillators.
> The first is the initial manufacturing error. I'd expect crystals made
> the same batch to have similar errors. If you want a large population,
> are going to get most/many of them from similar batches.
> The other is temperature. I'd expect that oscillators of a specific
> to have similar temperature dependencies. Some vendors even include a
> in their app-notes.
> It might be interesting to collect oscillators from different vendors and
> batches and see what sort of spread you end up with.
> These are my opinions, not necessarily my employer's. I hate spam.
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to
> and follow the instructions there.
More information about the time-nuts