[time-nuts] synchronizing a large number of weaklycoupled oscillators

[Eugen Leitl]

On Wed, May 12, 2010 at 12:14:50PM -0400, Bob Camp wrote:

> If the oscillators all lock to each other, then multiple oscillators don't
> have any particular advantage. 

They're not strongly coupled, they're weakly coupled. There's no
master-slave relationship, but a slow synchronization to a 
resonant oscillation pattern of the assembly as a whole.
Because of that they tend to average out individual inaccuracies,
but as other have observed you're running into diminishing returns
after some 10-100 oscillators assembly size.

Apart from enancing accuracy (e.g. integrating a planar array of
some 10^2..10^3 locally-coupled freerunning oscillators on a single
piece of silicon) each of them being very inaccurate due to geometry 
differences from fabrication and very high oscillation speed (GHz to 
THz) that could be interesting for synchronizing very large, distributed 
assemblies (global and larger) where relativistic delays figure 
prominently. 
 
> Let's assume you can isolate them so they don't lock to each other. If they
> are all of similar construction in a similar environment, there is a very
> real limit to the advantage you would obtain, no matter how many you have.
> It turns out that's true at one level for simple crystal clocks, and at a
> very different level for atomic clocks. 

-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
______________________________________________________________
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE