[time-nuts] FTS1200, OSA8600, OSA8601 phase noise and ADEV measures

[Magnus Danielson]

Fellow time-nuts,

I have borrowed Björn's FTS1200 and OSA8600 and that complements my own 
OSA8600 and OSA8601.

I use a HP5087A and one of the OSA8600 as references, such that I can 
cross-correlate out the reference oscillators. I'm still not quite 
satisfied just yet, so I may need to work some more to perfect this 
technique.

http://rubidium.dyndns.org/~magnus/time/timelab/FTS1200_OSA8600_OSA8601_20120802_1.png

Here you see the phase-noise of the four oscillators.
It is distinct how the white noise floor of the FTS1200 is around -141 
dBc/Hz when the OSA8600s reach for -155 to -156 dBc/Hz.

The spurs on the FTS1200 may trace back to the fact that the FTS1200 is 
-1.5 Hz of the references, but I like Johns comments on that.

You can see that the older OSA8601 has better 1/f noise and that Björn's 
OSA8600 SN711 has higher 1/f^3 noise than any of the other three which 
seems comparable. The OSA8600 SN711 and SN817 has comparable 1/f noise.

http://rubidium.dyndns.org/~magnus/time/timelab/FTS1200_OSA8600_OSA8601_20120802_2.png

The ADEV plot of the oscillators.
The noise on the FTS1200 dominates, it's inital response, but it is more 
clearly seen that the flat section is better and the OSAs.

http://rubidium.dyndns.org/~magnus/time/timelab/FTS1200_OSA8600_OSA8601_20120802_3.png
The Hadamard Deviation plot of the oscillators.
The linear drift of the FTS1200 and OSA8601 is seen.

http://rubidium.dyndns.org/~magnus/time/timelab/FTS1200_OSA8600_OSA8601_20120802_4.png

The phase difference plot of the oscillators:
The linear drift noticed from the ADEV plot and mostly removed in the 
Hadamard plot is seen more clearly in the phase plot as quadratic 
phase-shifts, a slot tendency on the other oscillatos can also be seen.

http://rubidium.dyndns.org/~magnus/time/timelab/FTS1200_OSA8600_OSA8601_20120802_5.png

The Frequency difference of the oscillators.
The noise of the FTS1200 is visible. I haven't looked into it.

As you may have noted, the measurement length varies, I haven't looked 
into it.

My OSAs has been heated for a fairly long time, even if they had a short 
dip of voltage. Björns FTS1200 and OSA has been heated for almost 2 
weeks now.

Measurement files:
http://rubidium.dyndns.org/~magnus/time/timelab/FTS1200_HP5065A_OSA86000_20120802_1.tim
http://rubidium.dyndns.org/~magnus/time/timelab/OSA8600SN711_HP5065A_OSA86000_20120802_1.tim
http://rubidium.dyndns.org/~magnus/time/timelab/OSA8600SN817_HP5065A_OSA86000_20120802_1.tim
http://rubidium.dyndns.org/~magnus/time/timelab/OSA8601SN1314_HP5065A_OSA86000_20120802_1.tim

Comments welcome. I will spend some more quality time on this now that 
I've made a first stab at it.

Got to locate a cable so I can toss a well-heated 10811A into the battle.

Cheers,
Magnus