a handful of us got together this afternoon, and here is what was decided.
1. we will take one more "standard" 3-bit test dataset, then will
be done with testing for a short period. this will give us four
such datasets, three of which remain to be reduced/analyzed.
that number should allow us to make some statements on either
how stable the system is or how uncertain our measurement of
the quantities is. i believe ken and vivek are taking this data
right now (or will be shortly). we also have the ~8 hour test
dataset taken last monday evening. vivek has had a very
preliminary look at this, and the data look good, except for a
single polarization on one antenna, which wasn't fringing. he
will continue to work on that dataset first (before the
remaining three standard ones).
2. doug and his group will go out and remove the 3-bit digitizer
units either wednesday or thursday and bring them back to the lab.
while here, a number of tests will be done (why antenna 22L is so
bad, why needed input power levels are so variable, etc.). mike
will also install the new firmware that will allow access to bits
2,3,4 in addition to 1,2,3. (as an aside, mike, how do we actually
switch between the two?)
3. those 3-bit units will be replaced with standard 8-bit units,
but those replacement units will have the two attenuators on them
that should remove or at least reduce the rolloff at the top
end of the bandpass of the 8-bit units (1 GHz). we will need
somebody to do a test once those units are out on the array, to
see if the rolloff is actually improved or removed (rick?).
4. mucking around with the 3-bit units here in the lab might take
a week or more. when they are ready, they will go back out onto
the array for repeats of the tests (and some new ones - one
that was considered key was re-routing B1 into C1 and seeing if
there is any crosstalk). when they go out, they might
the extra attenuators on them, depending on the results of the
test in item 3 above.