[evlatests] Phase Jumps and the Flukes

[Barry Clark]

Rick has extended my hypothesis.  In the form I gave it, I claimed that 
return-to-phase was guaranteed only if the Flukes themselves were set to
a multiple of 10 MHz.  I now see that if they are not, but if they are
commanded to a frequency differing from what they are set to by a multiple
of 10 MHz, the internal divider chains of the Flukes are not reset, and 
phase will return if the original frequency is reset.  This behavior, 
which surprised me, is encouraging evidence that the Flukes not being 
disturbed by transients involved with resetting their frequencies.  (I 
don't see it as being paricularly useful otherwise.)

> From evlatests-bounces at donar.cv.nrao.edu  Mon Mar  5 19:45:18 2007
> Date: Mon, 05 Mar 2007 19:44:46 -0700
> From: Rick Perley <rperley at nrao.edu>
> User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US;
> 	rv:1.8.0.9) Gecko/20061215 Red Hat/1.0.7-0.1.el4 SeaMonkey/1.0.7
> MIME-Version: 1.0
> To: evlatests at aoc.nrao.edu
> 
>     A test was run today to test Barry's hypothesis that changing 
> frequencies by a value selected so that the Fluke synthesizes are 
> incremented by a multiple of 10 MHz will result in no phase jumps.
> 
>     Barry is correct. 
> 
>     Here are the details.
> 
>     The first 30 minutes of the 60 minute test alternated between two 
> frqeuencies, 1465 and 1445 MHz for IF #1, and 1385 and 1405 MHz, for IF 
> #2.  These values were selected so that the frequency change could be 
> set by the Flukes alone (i.e., the L6 settings were made identical).  
> Specifically:
>    
>        1465/1385:      L6:  3640 and 3560,    Flukes:   121.975 and 221.775
>        1445/1405:      L6:  3640 and 3560,    Flukes:   101.975 and 241.775.
> 
>     (Ken reminded me that it is insufficient to change the frequency by 
> 10 MHz, as the Flukes running the B and C IFs increment by half that set 
> for A and D). 
> 
>     During this first 30 minutes, I alternated back and forth, with a 
> one minute integration. 
> 
>     The second 30 minutes were spent in the same way, but with a pair of 
> frequencies that were NOT different in frequency by a multiple of 20 
> MHz, but which also used the same L6 setting.  Specifically
> 
>        1465/1385      L6:  3650 and 3560      Flukes:  121.975 and 221.775
>        1450/1390      L6:  3650 and 3540      Flukes:  106.975 and 226.775
> 
>     For all observations, the BW was 6.25 MHz. 
> 
>     Results:
> 
>     For the first pair of frequencies (which differed by 20 MHz), all 
> phases are rock steady.  There are no jumps. 
> 
>     For the second pair of frequencies, there are many 'global' phase 
> jumps -- all of them of the exactly 180 degree variety.  The 
> characteristics of these 180-degree jumps are identical to those 
> reported earlier (in the five or seven L-band frequency tests:)
>        All VLA antennas jump identically. 
>        The RCP and LCP sides, for a given IF, jump identically.
>        The two IFs do not jump in synchronism.  For example, for IF1 at 
> 1465 MHz, the VLA was in the 'up' phase statesfor eight of the 13 
> observations, and in the down (by 180 degrees) state for five.  But IF 
> #2, at 1385 MHz, the 'up' state was seen in four scans, the 'down' in 
> nine.  Meanwhile at the other frequency pair (1385 and 1390 MHz for IF 1 
> and 2, respectively), the patterns were quite different:  IF1:  7 ups 
> and 6 downs;  IF2:  5 ups and 8 downs.  There seems to be no correlation 
> between changes between up and down between IF pairs or frequencies. 
> 
>       It is interesting to note that the 72-degree regular phase 
> advancement (or retardation) noted in my multiple, cyclic frequency 
> tests two weeks ago, is entirely absent.  The only jumps in this 
> database are by 180 degrees.  Given the manner in which the Flukes were 
> commanded, it seems pretty clear that the 180 degree family of phase 
> phenomena originate with the Fluke synthesizers. 
> 
> 
> _______________________________________________
> evlatests mailing list
> evlatests at listmgr.cv.nrao.edu
> http://listmgr.cv.nrao.edu/mailman/listinfo/evlatests
>