[time-nuts] Quick Allan deviation measures

[Magnus Danielson]

Hi!

Just to make some comparision I have made a number Allan deviation measures
using the borrowed CNT-90, which has been clocked with my Cs-beam and then
having measured a number of sources at various taus. All measures have been
performed by taking 100 frequency measurements, which is clearly on the cheap
end but I was not attempting perfection, the choice of taus where 1 ms, 10 ms,
100 ms and 1 s. As the subjectline says, quick includes cheating out on 10 s
and 100 s.

The sources was:

Rb: My R&S Rubidium clock, a classic Rb with sine modulation tracking,
    observed at the 5 MHz output of the actual Rb reference, not the frequency
    synthesis unit.
Cx1: The reference clock unit as taken out of a functional Ericsson AXE-10
     station. Observed at its 5 MHz (as being modded in, normally output 8kHz).
Cx2: The option 001 on a HP4195A, basically a HP10811. Observed at its 10 MHz
     which is directly available as it needs to be linked into the instrument.
GPS: The HP Z3801A which I have been talking about.
Cs1: The 5 MHz output of the Cesium beam. Used to establish noise floor.
     Measured while being locked to the 10 MHz.
Cs2: The 10 MHZ output from the counter, while being locked to the Cs 10 MHz.

The results are:
        1 ms     10 ms     100 ms    1 s
Rb	6,5E-8   7,0E-9    5,7E-10   1,94E-11
Cx1	6,5E-8   1,68E-8   1,62E-9   3,08E-11
Cx2	3,8E-8	 4,8E-9	   4,8E-10   2,47E-11
GPS	3,9E-8	 5,2E-9	   5,6E-10   7,83E-11
Cs1	5,7E-8	 5,5E-9	   5,7E-10   8,2E-12
Cs2	7,0E-8	 6,6E-9	   5,3E-10   1,128E-11

See how the GPS diciplined 10811 with extra oven keeps it track above that of
a normal 10811 sitting in a unit. Notice also that the Cx1 had been powered
only for an hour instead of over 24 hours.

Infact, the GPS is worst of all these sources!!!

I am quite sure that a 10 s measure would should the GPS even more
infavourable.

Hopefully these measures will help you to grasp what I mean about "walking
around". Unfortunatly I don't have a complete solution for getting digital
readings out of any of my instruments at this moment (and no, tossing me links
to yeat another Windows tool will NOT help, the laptop don't have GPIB and I
will not get one for it since it is not my private laptop, I have a propper
NI PCI card sitting in one of my Linux boxes, that will have to suffice).

I could of course grab my digital camera, but it is such a ugly way of
documenting things.

I hope these measures have shed some initial light over things in case you
where wondering.

Cheers,
Magnus