wrote:
But the
ones I have used seem plagued with near-signal desensitization, front
end overload, etc., and I suppose all that comes from putting the
selectivity so far downstream.

I'm almost tempted to get an old 75s4 and shut up, but I really don't
need another room heater, so, instead, I'm thinking of building my own
receiver along the lines laid down by the late Doug DeMaw in his _QRP
Notebook_. Single conversion 160m superhet with Collins mechanical
filters in the IF and a series of down-converters for the other bands.
Anybody got any experience with the DeMaw Design?

The DeMaw designs are very nice but with their band-wide converters
they also suffer from having selectivity downstream rather than at the
front. Preselection at the front helps with any of these.

The modern choice of a 45MHz or so first IF is really pretty nifty for
a general coverage receiver from the past couple of decades. Lack of
preselection can be largely cleared up by bolting something before the
front end, and indeed many of the receivers you see that cost as much
as a car have tracking preselectors.

But if you don't want to upconvert, as I see it you have two choices:

1. Don't even try bandswitching. Coil sets for each band. You end up
with the HBR-16, a very elegant and homebrewable receiver.

2. Do bandswitching but with elaborate tracking front end (and if
necessary a tracking IF, but with modern synthesizers you probably
wouldn't), probably with many switching sections to handle the required
octaves. You end up with a R-390A.

Tim.

"Tim Shoppa" ) writes:

1. Don't even try bandswitching. Coil sets for each band. You end up
with the HBR-16, a very elegant and homebrewable receiver.

One reason there was that phase with a separate converter per band ahead
of a receiver tuning a fixed band was to avoid switching tuned circuits.
By that time, the semiconductors cost so little that it was easier to
duplicate them for each band, and then the bandswitching becomes so much
easier. No fussing about getting close to the tuned circuits, you simply
switch the "B+" and the input and/or outputs. Or, make those
converters plug in, and then no switches required at all.

I sort of alluded to this in an earlier post. Make a good receiver,
minus the variable oscillator and the front end tuning. Maybe even
put the mixer in the "plugins". Then you end up with a good receiver
that is quite flexible, because the things that you may want to play
with and may give trouble are in a separate box or plugins.

YOu can even play with things like tuneable frontends versus something
that is broadband across a ham band. Some bands might interest the
builder more than others, so they could build a really good plugin
front end for that band, and lesser front ends for other bands, or
leave off the bands they aren't interested in (but those bands
can easily be added later, unlike a bandswitched rig).

IN the sixties, there was a guy who had a whole slew of receivers
described in CQ. Virtually all of them were single bands, and he
made the point that it left off bandswitching, and of course he
could choose an IF that better matches the tuning range. Good points,
but an awful lot of duplication. Build a good receiver first,
and then play with the frontends endlessly.

Michael VE2BVW