[evlatests] Results from T304 Attenuator and Requantizer Tests

[Michael Rupen]

     Michael:

    It appears that I used the AC for the most part.  I do know (from earlier 
tests) that the BD IFs share the same ills, but less critically -- Cyg A has a 
steep spectrum, so the problems recede a bit at 1850 MHz...
     I checked the Stokes 'I' residuals for all tests -- it's easy to do. Two
     things improved with these tests:

    1) The egregious closures are gone with proper power levels.  I bet the 
worst closure problems were due to overflows in the correlator accumulators -- 
it was always on short spacings (highest cross-power), and only at the lowest 
frequency.    2) The PDif compression was reduced -- but only by a modest 
amount.  We'll have to look elsewhere to explain this annoying problem...

    Rick

Michael Rupen wrote:
>  Hi Rick --
>
>    are you using both IF pairs (AC, BD) or just one?  and do the results
>  depend on the IF pair?
>
>    The non-closing residuals could just correspond to those antennas
>  for which the attenuators were set improperly or inadequately (i.e.,
>  couldn't knock down the power enough to achieve linearity).  This can be
>  IF-dependent, hence my question.  Keith knows how all the attenuators
>  were set, as well as the resulting power levels, so we should be able
>  to check.
>
>    For which test did you check the Stokes I residuals?  That is,
>  were the results given true for *all* tests, or just for #2, #4, or what?
>
>    Your note gives the impression that these tests did improve (if
>  not eliminate) the closure problems.  Is that true?  Were specific
>  baselines "cured", and if so, which ones?  We could compare those with
>  Keith's information to see whether there's a correlation.
>
>                     Michael