Active8 wrote:
PIN diodes have an intermodulation problem when
the forward bias is right at the threshold of
conduction. Ordinary diodes are much worse. At
higher/lower forward bias levels this effect
quickly goes away.

In order to assure that the 20 dBm third-order
intercept point of the receiver is not degraded by
this effect at the critical diode bias point, I
use three 1N4007s in series at each of the three
locations (verified by measurement).


you mean by doing that, the chances of all the diodes switching at the
threshold is reduced?

The diodes are pretty much uniform, as compared to
many years ago. The reason for three diodes in
series is to divide the RF/IF signal voltage into
three approximately equal segments. This greatly
reduces the intermodulation in each diode,
therefore in the set of three. I verified this
with a spectrum analyzer plus tracking generator
equipment. Also verified many (33) years ago.

Bill W0IYH