There's no question that the 1N400X series is a poor switching diode
when it comes to speed and reverse recovery characteristics, however,
for switching in and out bandpass filters I think the reverse bias
junction capacitance is the major issue to prevent unwanted signal
coupling. I would definitely look for true PIN diodes if I were
upgrading the filter switching diodes rather than use 1N4007s. It's
just interesting that there have been articles on the fact that they
appear to work reasonably well in some cases. I know the Kenwood
R-2000 I had used the equivalent of 1N914s for these diodes in the
front end and they were horrible with all kinds of cross talk from
stronger, out of band signals.

Frank

Jake Brodsky wrote:
wrote:
Now that I think of it Michael is correct. It was the 1N4007, the 1 kV
version of this slow diode that was recommended. It would tend to have
lower junction capacitance because of the wider depletion region due to
the heavier doping for the high voltage standoff. Once the diode is
biased on dc wise, the forward drop is not significant because it
doesn't "switch" with the signal current. I think the trend to
Schottky diodes was because of the lower intrinsic capacitance.

There are better switching diodes than the 1N4007, though they are more
expensive and less commonly available. It's not a bad switching diode
for RF if you don't have anything else handy. If I'm not mistaken,
Ulrich Rhode wrote an article in QST some years back about the use of
such diodes in high performance receiver designs. His opinion was that
although cheap, the 1N4007 didn't have a place in higher performance
designs.

DE AB3A