[GPS_Standard] Question about MonTrol statistics calculation

[Chris Howard w0ep]

On 1/19/2013 6:10 PM, Chris Howard w0ep wrote:
> 
> I'm trying to write a perl script which
> monitors the output of my GPSDO and calculates

I now think I have my statistics working OK.
The numbers coming out are similar to the numbers
that MonTrol gives.  I suspect the difference is
in data representation.

I also decided to print using decimal format instead of exponential.
I have the luxury of only having 10 MHz as the
base in my software (as opposed to MonTrol which
accommodates 5 MHz systems),
so I print "0.13"  instead of "1.3e-08".  And that
gets me down to individual Hertz.

I'm also trying to get my head wrapped around what
the statistics are telling me.

Avg. Freq. Offset: I believe it is the mean of the offsets
for all of the samples in the set so far.  I'm usually
seeing a number like 2.3e-10 here, something like that.
That would be 0.0023 Hz on a 10MHz signal.

Standard Deviation: is usually a number like 1.5e-08
(0.15 Hz).

I think this is telling me that for the current group of
samples the mean of the samples is 0.0023 Hertz
from the signal being put out by my OXCO.  And the
standard deviation means that the biggest chunk
of those samples is within a window +/- 0.15 Hertz
of the signal being put out by my OXCO.

The mythical centerline of the GPS samples is my
true standard.  So I'm not sure if I would say
my OXCO is 0.0023 from "right" or if it would be
more proper to say something about the SD.

My GPS is the Trimble Resolution T.  This has its own
software and there are options to set regarding the PPS
signal.  I wonder if any of those has meaning in the
current context, but not smart enough to know which ones.

Learning a lot, and enjoying myself.  I don't really
_Neeeed_ a  +/- .1 Hz 10 MHz frequency standard, but
I've enjoyed building this and trying to make it work.
(and attempting to understand whether it is working or not!)



Chris
w0ep