[evlatests] Hysteresis in Tuning failures

[Ken Sowinski]

On Thu, 17 Jun 2010, Claire Chandler wrote:

> Does "fail to tune" mean no fringes (noise) or is there an actual M&C flag
> that says IFs A+C failed to tune?

There are two classes of synthesizer failure (as you might expect).
Those discernible through monitor points and false locks which can
only be determined through lack of fringes or a frequency meter.
In the first case the useful monitor points ar M_LOCK, C_LOCK, and
STANDBY.  The first two generate alerts if set, but not the third.

A broken synthesizer will affect one or more basebands.  It is not
obvious to me that mcaf has all the information needed to make that
association.  I offer below a table for future reference that
purports to map synthsizers to basebands which can be used as a
guide when generating flag tables.  Someone should look at it with
care to see if I have made any mistakes.  My impression is that
most synthesizer lock failures are for the L302s, not the L301s,
so that we might gain a lot by implementing only the simpler L302
flagging matrix.


Associating a baseband with a broken synthesizer is a difficult
business because of the flexibility of signal switching at the
antenna.  The simple rule is that you need to know which samplers
are used to map L302 failures, and which band is in use to map
L301 failures.


The L302s are easy.  First column for 8 bit samplers, second
column for 3 bit samplers.

syn     8bit 3bit
===     ==== ====
L302-1   AC   AC1
L302-2   BD   BD1
L302-3   --   AC2
L302-4   --   BD2


Determining which basebands are affected by a broken L301 is much
more complicated.  'AC' and 'BD' denote all possible baseband pairs
regardless of which samplers are used.  'BD-T303' denotes BD is bad
if it goes through the T303 conversion path.

band       L301-1   L301-2
====       ======   ======
P,L,S,C      AC       BD
X            --       --
Ku           AC       BD
K,Ka,Q      all      BD-T303