21oct13
 Folks,

Mike Nolan is asking that we extend the sliding shutter
tests to encompass a significant range of frequencies as
well as to cover the range of zenith angle settings, in
order to be sure that there are no significant gravitational
effects on the shutter's isolation, due to air gap or other
mechanical variations.

I'm in full agreement with doing this, with the further
objective of checking that no mistakes were made in the
simpler set of measurements we performed earlier.

The part about covering a reasonable range of frequencies will
require a moderately different measurement setup and method
if the measurement time is not to explode into something wholly
untenable.  But even with the new method, I've concluded that
doing the setup and all the measurements will require quite a
bit more time than can be accommodated in a single day.  One
issue is that keeping track of TX power during the tests will
be significantly more complicated.

Therefore, we will need to arrange the setup so that certain
major components can be left in place for several days without
interfering with operation of the telescope. 

In particular, the 15W power amplifier (with its power supply)
and the coupler will need to be mounted near the JHU/APL coaxial
hybrid well enough that they will stay out of trouble in the
course of the usual telescope movements, floor rotations, etc.

There will need to be two coaxial cables from the vicinity of
ALFA to the TX floor, and these should be left in place between
measurement sessions.  I think this can be managed OK by planning
to disconnect them from the 15W PA and the coupler when not in
use, and coiling them up and cable-tying them to the rotary drum
(probably on top).  I do not want to run these cables through the
normal cable wrap because the lengths and hence the loss would be
intolerably high.

There will be no need for the Agilent power meter in this setup-
the pulsed nature of the VNA's output would prevent proper power
measurement by ordinary means and have forced me to use a different
method.  Also we would not need the CW signal generator.  So the
equipment setup near ALFA will reduce to the VNA plus a manual
step attenuator.  The price paid will be the additional calibration
procedures to relate TX power to step attenuator settings and to
use the VNA for absolute RF level measurements, as well as some
additional risk of mistakes in doing all this stuff correctly.

See the attachment.

I will need support in:
> Lugging the R&S VNA and its CAL kit back up to the dome.  Also
  the 15W PA and its power supply if they're not already still
  in the dome.
> Lashing down the 15W PA, its power supply, and the coupler very
  near the JHU/APL coaxial quadrature hybrid.
> Stringing the two coaxial cables.  BTW, we may need to make one
  or both of these cables anew- I don't know if either of the 30ft
  cables we once made for this general situation still survive.
> General help in setting up and calibrating, as well as meeting
  the "2-man minimum" safety rule.

Note that the above assumes using the highly-portable R&S VNA for
these tests.  Keeping track of the TX power could be simplified
if we were to use the Agilent PNA instead, since it uses a CW (not
pulsed) RF output, permitting going back to using a regular power
meter.  However, this instrument is decidedly not portable, and
would require lifting to the dome in a bucket as well as a lot of
unpleasant manhandling inside the dome.

We will need freedom to move to arbitrary zenith angles throughout,
both for dome access and for "gravitationally stressing" the sliding
shutter system during testing.  So we will need to coordinate with
the platform crew to avoid conflicts.

Also I think we should be thinking about priorities between this
activity and trying to chase down the reported instabilities in
ALFA, which are apparently interfering with its observational use.

Dana