[evlatests] 3-bit update

[Vivek Dhawan]

I managed to squeeze 30min of work into 3 hours.

The 3-bit samplers work. I did not get as far as setting the equalizer
slopes, or the post-RQ gains. I poked at the GUI that makes state-count 
histograms, but it did not listen, so I can't say yet if they look better
Also, need to do a on/off source to get the noise on a blank field.

But I did figure out that the sysPower data are actually present for the
whole timerange (The partial dropouts are an AIPS feature in SNPLT, Eric
knows what it is) and the Tsys were reasonable, so there is optimism even
in the un-optimized state.


ant  Tsys 3bit           Tsys 8bit
      R3 L3 4880MHz       R8 L8 4872

ea12 33 34               30 28
ea15 32 34               30 32
ea22 33 45               31 31
ea28 42 34               31 32


Vivek Dhawan wrote:
> A follow-up on switched power:
> 
> Last time we did 3-bit tests, in September 2010, there was no switched
> power data. Now there is.
> 
> The bad news: Tsys for the 3-bit antennas was higher (80-180K) than when
> the 8-bit path is enabled (~25-30K).
> 
> The good news is that as levels and gain slopes were adjusted toward
> optimum rms, Tsys improved dramatically.
> 
> The other bad news: in every single one of the files we have run so far,
> the switched power data disappear after 5-10 minutes. All antennas drop
> at the same instant. Bruce did not see anything on the station board
> outputs to indicate anything amiss, but will look into it more tomorrow.
> 
> onward.
> 
> 
> Michael Rupen wrote:
>> We spent most of the day on 3-bit testing.
>>
>