[evlatests] L352 round-trip phase meters.

[Vivek Dhawan]

Updated versions of the L352's are in production, and are being compared
in pairs to check for liars. The first pair have been examined in gory
detail for a week, and meet or exceed the original stringent specs.

A number of fixes (some quite ingenious) were devised by Keith and Matt
Luce, to correct flaws in hardware and/or reference signals. These will
be enshrined in a forthcoming memo.

One blemish is much attenuated but detectable: there are small dead zones
of <1 milliturn every 250 milliturns. In use on the array, the resultant
phase error is <3 deg at X band, scales with frequency, and would occurr a
few times a week on each meter, independent of the others. These errors are
(a) Small, within spec; (b) Slower and smaller than almost all tropospheric
errors; (c) Less than other errors still in play on the array.

So I propose to declare the meters done as regards design. Evaluation will
continue as units are fabricated and deployed.

Vivek.