One of the greatest problem for radio astronomy receiving systems is gain and baseline
drift in the presence of ambient temperature changes. Amateur installations are even more
subject to this problem, single antenna systems in particular.
The usual solution to this is " Dicke switching" . The input of the first LNA is being switched
continuously between the antenna and a termination resistance (which is normally cooled to
a few deg. K in liquid helium). The switch can in principle be electro-mechanical, but is
usually electronic. The switching device adds losses to the LNA input, which may be tolerable
for a system working with a 60m dish, but amateur systems typically use dishes <6m and can
therefore ill afford to throw away input sensitivity.
With this in mind I started to explore the possibility to modify the Dicke switching concept
to make it a bit more "user-friendly" for single antenna amateur systems . There is not much which can be done to avoid losing 50% of the signal ,because we have to spend 50% of the
total time on the "reference termination", just like in the normal Dicke system.
But for the remaining 50% of the time when the LNA is connected to the antenna, we want to
get all we can , without suffering additional losses in the switching components.
One possible way out of this dilemma is to develop a different type of LNA. In this project it was
decided to build a dual input stage unit, the switching is done AFTER the LNA ,but before any
further amplification. Both LNA input channels have to be exactly identical, or at least as far as
this can be achieved. The rationale is that any drifts will be experienced by both channels to the
same degree. In practical terms this means we should get drift cancellation from the whole LNA
assembly, because any gain changes by the "antenna channel" should also be happening in
the "reference channel".
Whilst designing the receiving system it occured to me to also implement automatic gain
control at the same time ? (I guess this will make some of the ARA die-hards cringe.....).
The AGC loop is set up to derive a correction voltage from the detector signal level during the time the input switch is on the 50 Ohm termination. This is compared to a detector setpoint
level, the error signal is amplified and filtered and drives the dc injection to a DBM mixer module
at the I.F. input . The mixer is actually operated " upside down" , working as a continuously
variable attenuator. At least 20dB variation can be extracted from this set up, much more than
is actually required. The entire system (minus the initial LNA stage), from the subsequent LNA
stages all the way to the detector output stage, is encompassed by this AGC loop.
Surprisingly, the additional circuitry required to implement the AGC turned out to be almost
trivial. I'm not going to claim that no temperature control needs to be installed for this
receiver, but the system gave a very good account of itself without any temperature stabi -
lization whatsoever.
