[evlatests] 2GHz-wide 3-bit wideband data

[Rick Perley]

Michael/Ken took 40 minutes of full wide-band data with the 3-bit 
samplers.  Data quality looks very nice, with uniform delays, continuous 
amplitudes/gains across all 16 subbands. 

But the rms noise is far higher than it should be.  In subband 12 (6928 
MHz -- same as the simple test done by Ken a little earlier), we have 
for our 'noise matrix':

             12            15              22               28
------------------------------------------------------------
12         XX          .41            .75             .51
15         .32           XX           .74             .50
22         .38           .45            XX            .99
28         .43           .49            .57               XX

    The matrix entries (says here) should be 0.15!  So we are a factor 
of 2 to 6 too high.
    Perhaps more interesting is that the entries for the baselines 
amongst 15,22,28 are three times higher than they were for the simple 
1-subband test done by Ken a little earlier.  It's as if all noises have 
been multiplied by three. 



Michael Rupen wrote:
> The file name should be C3tune_3bit2GHz_000.
>
>                   Michael
>
>> We have a full 2 GHz of 3 bit sampelr data with four antennas:
>> 12, 15, 22, and 28.
>>
>> 2132 UT  Start script
>> 2135     finish adjustments
>> 2207     finish script
>>
>>
>> Ken
>>