[evlatests] BD T304 power meters and attenuators

[Michael Rupen]

> B/D attenuator settings being fixed works well if you use the same LO
> freqs. If different parts of the band are used (different LO tunings)
> the LUT tends to cause problems with incorrect level settings. Just my
> observation.
>
> Rob

Of course one could store settings per LO tuning rather than simply per
band, for those bands for which it matters.  Do the attenuator
settings change significantly* with position on the sky, or only with tuning?
The latter could be stored as a (big) table, but the former probably 
requires real-time setting.**

              Michael

* "significantly" of course depends on the samplers being used.
   I cling to the fond hope that we can tablify (?) the 8-bit case,
   though probably not the 3-bit one.

** The fact that we'll have to go to real-time setting at least for the
   3-bit case says to me that we have to come to grips with the effects
   of changing attenuation within a single observation.  Given that,
   in the long run, I'd be happy with ALCing all the time...but I'd prefer
   to understand the effect this has on the data, and how to account for it,
   before making this the default for scientific observations.