[time-nuts] ARRL 2006 Frequency Measuring Test

[John Ackermann N8UR]

Hal Murray wrote:
>> You might also be interested in the 2005 analysis page (linked from
>> the  URL above) -- some statistical analysis indicates that W1AW was
>> actually  about 0.4 Hz off in the frequency they reported on 160M. 
> 
> What are the theoretical limits?
> 
> If I only have N seconds of a signal, that sets a lower limit on the 
> bandwidth as the signal has to turn on and back off.  Is there a simple 
> rule-of-thumb?  Is it seconds or cycles that matters?  (Does it scale by the 
> frequency?)
> 
> How stable in the atmosphere?  If the propagation path changes during a 
> measurement the receiver will see a Doppler shifted version of the signal.
> 

The atmosphere is the biggest limiting factor; you can often see the 
signal wobble around as the ionosphere moves up and down.

The old rule of thumb is that it's pretty hard to measure to much better 
than 1 Hz over a skywave path, but I think the FMTs have proven that you 
can do quite a bit better than that.

I just looked at my results over the 4 years the test has been run, and 
I had five measurements that were less than 0.1 Hz off the ARRL measured 
frequency, and four more that were less than 0.5 Hz off (using 
fractional frequency notation doesn't make much sense here because at 
these frequencies the reference error is negligible compared to the 
atmospheric errors).

The only measurement that was more than 0.5 Hz off was due to an 
operator error.  And, I should point out that some of these measurements 
were made on signals that were barely discernable by ear; only an FFT 
over a minute or more of signal made the reading possible.

John