) writes:
I don't much like the receivers I've used in contemporary tranceivers
-- the general coverage synthisized open front end ones. (I hasten to
add I haven't used any of the $4000 rigs; can't afford them). 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?

Jim, K5YUT

But then you're just recreating something that is open to problems.

That sort of scheme was used to get a constant tuning range over each band,
and because synthesis wasn't easily applied.

But on anything other than the core band, you've got double conversion that
moves the selectiving past two mixers. And unlike current up-conversion
schemes, there is not even a crystal filter of some sort at the output of
the first mixer.

Switch to an IF in the HF range, and you immediately eliminate most image
problems. No fussing over the front end about that, and of course, you
won't have to have the front end tracking the tuning oscillator.

Remember, a whole wave of amateur transceivers and receivers went
to that sort of scheme. And unlike general coverage receivers, you
don't have to worry about any problems due to the IF being in the tuning
range, and you can think up various schemes to do the tuning since you
don't have to cover the 30MHz.

YOu can stick with the modular theme, and thus build only for the
bands you want or even build a band at a time. But instead of a whole
converter, you'd have the preselector circuitry (and maybe an RF stage) and
the variable oscillator for each band. Or build a good receiver up till
the input of the mixer, and then figure out what comes next.

IN the sixties, that wave of single conversion to IFs in the MHz range
used various schemes to deal with the local oscillator. Obviously some
bands needed a frequency range that could easily be supplied by a variable
oscillator without drift (and some of the rigs took that to the extreme
and used the variable oscillator directly on 10metres). The problem
with bandswitching the oscillator would of course be the issue of
getting it to tune only 500KHz or so on each band (and any stability
issues caused by the switching of the LC circuits).

Other rigs used pre-mixing, so the variable oscillator would always
tune a fixed range, but it would be converted to the needed frequency
with a mixer and crystal oscillator. One does have to watch out
for spurs on the output, but it gets the extra mixer out of the signal
path, and given a relative handful of 500KHz ham bands in the shortwave
region, the cost of the crystals wasn't out of range (though maybe
today..).

Then later, some rigs used PLLs. I can't remember if the Signal One
used one, but certainly in the seventies they came along. Same basic
idea as the premixer, but the PLL was the filter so the VCO directly
fed the receiver's first mixer.

Rigs like the TS830S used a PLL for that same purpose, though they
came up with a pretty fancy scheme to limit the number of crystals
needed.

One of the things about receiver design is that the trends have
often reflected limitations of the times. There may be a good
reasons for doing things a certain way (such as adding a third
conversion to a receiver so the BFO is on a different frequency
from the one where the main gain is), but it may also mean
they couldn't do anything better at the time.

Michael VE2BVW