08 FEB 2008 - Doug explained some of the IF circuit (sheet 3/3) to me

The networks of resistors and PIN diodes between U1 and U2, and between
U2 and U3, serve as an AGC.  Basically these circuits are operating as
voltage controlled attenuators, which Doug says is better than actually
changing the gain of an amplifier, which screws up the waveform.

CR5 is the detector diode.  Detector output voltage is amplified and
buffered by U16.  Output of U16 controls the loss in the AGC networks.

Examine the network between U1 and U2.  Think of the PIN diodes as
variable resistors at RF, controlled by the DC voltages across them.
If the IF level out is near zero, CR1 is forward biased, and has
low resistance at RF, so output of U1 is connected directly (except
for capacitors) to input of U2, skirting the attenuator formed by
R4, R5, and CR2 (which has ~zero DC bias across it).  As IF level
increases, forward bias across CR1 drops and forward bias across
CR2 increases; ultimately reistance of CR1 is high and that of CR2
becomes small, so most RF travels through attenuator formed by
R4, R5, CR2.  Doug says dynamic range of each AGC is ~30 dB.

Note that the IF Pwr line from page 2 is strictly and on/off thing,
to turn off IF when locked on wrong sideband.