[evlatests] Stability Tests, on Friday afternoon

[Jim Jackson]

We'll talk to Matt and Pete on Monday about the L302.  The unit installed 
today in 13 BD is the one used to develop and test the new firmware so, of 
all of them, it is the one that should have worked.  On the bench, the 
longest out of lock time Matt or I measured in these synthesizers was 
10-20us - orders of magnitude shorter than the older ones.  The only 
exception is on coarse frequency changes which cause out of lock conditions 
for many 10's of milliseconds but that should not be happening every 10 sec 
(unless for some reason you are trying to calibrate or change bands that 
often).

We also need to discuss the status of this module and it's effect on the 
schedule.  Due to these problems, we have not ordered any additional 
motherboards, YIG driver boards and DDS boards for this thing. We will have 
only enough L302's for antenna 26. After that we will be outfitting 
antennas with no L302's.

Jim


At 03:41 PM 6/23/2006, Rick Perley wrote:
>     Ken and I ran a 1-hour test early this afternoon to test for amp &
>phase stability,
>and to see if phases return correctly after having changed bands.
>
>     Antennas:  13, 14, 16, and 18.    Antenna 24 was not available.
>
>     Bands:  X and C, alternating on a one-minute time scale.
>
>     Integration:  0.41 seconds.
>
>     Correlator:  Continuum.
>
>     Weather:  Not great -- thunderstorms in all directions.  Phase
>stability was
>pretty bad, with radian-size excursions on a few minutes of timescale.
>
>     Results:
>
>     A)  Phase.
>
>     Antenna 13 phase connects correctly on all four IFs.  There are no
>odd phases
>visible.
>     Antennas 14 and 16 do not connect.  For each, A and C have identical
>non-connections,
>B and D also (but different than A and C).
>
>     Antenna 18:  A and C do not connect.  B and D do connect correctly.
>
>     From this, it seems the L302s in antenna 13, and in 18 B and D,
>behave correctly.
>
>     B) Amplitude.
>
>     The situation here is not so good.
>
>     Antennas 13 and 18 B and D show the return of the fast dropout
>problem.  The
>characteristics are the same as what we had many months ago:
>
>     - Amplitude drops are large -- the most common are by 2/3 and 1/4.
>The dropped
>amplitudes are not random in amplitude -- only certain values are seen.
>     - The drops only occur 'on the 1s' -- i.e. at 11, 21, 31, 41, or 51
>seconds in time.
>     - Not every time 'on the ones' has a dropped amplitude.  Most do.
>     - Drops are usually a single 410 ms record, but some are double.
>When double,
>they are of different amplitude.
>     - There is no phase perturbation visible -- this is purely an
>amplitude effect.
>
>     Antenna 18, IFs A and C (at X-band only -- the c-band system is not
>working) is
>a mess, with fast (but not random) variations in amplitude.
>
>     Antenna 14, IFs A, B, and C, is fine.
>     Antenna 14, IF 'D' shows a smooth change of gain at the beginning of
>every
>scan.  The gain change timescale is 30 seconds, and the gain change scale is
>large -- by 30% at X-band, and by 10% at C-band.  It sounds like the system
>recovering from a significant change in power level -- except that for both
>X and C bands, the gain change is in the same sense (the amplitude is
>too high
>at the beginning of the scan, and decreases to the stable level after ~
>30 seconds).
>If due to different power levels associated with X and C bands, i'd
>expect the
>change to be positive for one band, and negative for the other.  Read on ...
>
>     Antenna 16 is very good on all IFs, except for a small gain
>recovery, lasting
>about 10 or 20 seconds, for all four IFs.  In this case, the amplitude
>is the same
>for both bands (~3%), and the slope is oppostive -- fringe amplitude
>rises at the
>beginning of an X-band scan, and drops at the beginning of a C-band scan.
>
>
>
>_______________________________________________
>evlatests mailing list
>evlatests at listmgr.cv.nrao.edu
>http://listmgr.cv.nrao.edu/mailman/listinfo/evlatests