[sdiy] A Frequency Standard for Poor People?

[ASSI]

On Monday 15 December 2003 14:32, Glen wrote:
> In the past, I've seen DIY articles about building inexpensive,
> precision frequency references. Unfortunately, I don't have any of
> these articles handy at the moment. In concept, I think they all
> relied on extracting precision reference signals from "someone
> else's" existing reference signal.
[...]

If you have access to STM telephone lines you have a frequency source 
referred to an atomic clock normal (typically not farther than five 
hops away). Next best thing is IIRC GPS receivers (most easily 
accessible as well in the form of GPS modules), then certain video sync 
signals (this is not always the case anymore it appears, so you've got 
which channels still do - also when a channel is switching content 
between networks there might be rather large jumps). Broadcast time 
signals are subject to severe jitter from the transmission channel, but 
are long-term stable. The NTP pages at University of Wisconsin discuss 
some of these issues and the theory behind recovering the "true" time 
(or frequency) in detail and NTP would deliver a way to sync a VCXO or 
something like that to a time normal. A good VCXO is not exactly cheap, 
but the short-term jitter is superior to almost anything else and the 
long-term drift can be taken care of by the NTP slewing. A good time 
server setup consists of a VCXO, a GPS receiver (fixed antenna!), a 
radio receiver (WWV or DCF77) and of course a network connection. You 
need at least two of these components for a frequency source. 
Convergence of NTP is rather good even with just sporadic connection to 
a time normal when the calibration of the local clock is almost perfect 
already (which would be the case for a VCXO).


Achim.
-- 
+<[Q+ * Matrix-12 * WAVE#46 * microQkb/Omega * XTk/30 * sonic heaven]>+

SD adaptation for Waldorf microQ V2.22R1:
http://homepages.compuserve.de/Stromeko#WaldorfSDada