[HBR] Band imaging and upconversion (was 'Re: Capacitor question - terminology')

[Walt Hutchens]

Tim posted

> In many utilitiarian cases two different ham or SWL bands could be selected,
> with the same LO range, just by adjusting the preselector for either IF+LO or
> IF-LO sensitivity. This is intentional band-imaging.

Here's a specific example -- a receiver I built a few years ago.   The IF is
5450 kcs using a half lattice crystal filter.   The LO tunes 1450-1950 kcs
-- convenient because it allows use of a BC receiver tuning cap and it's
very simple to build an LO stable enough for SSB at this frequency.

IF - LO gives 4000kcs to 3500kcs, thus covering the 80M band

IF + LO gives 6900kcs to 7400kcs, covering 40M

You pick 80M or 40M by tuning one or more circuits between the antenna and
the mixer.   With this pair of bands you can do that with one coil and
capacitor covering 3500kcs to 7400kcs per tuned circuit; as usual two or
more tuned circuits are a good idea to prevent leakage of the image
frequency.

An added advantage of this scheme is that LSB 'on the air' is LSB for both
bands in the IF.   So you can use a single crystal at about 5450.75 kcs for
the BFO.  

Band imaging is a very handy way to get a two-band receiver without the
complications of a bandswitch, plug in coils, etc.   Even some multiband
sets do it for two bands, using conventional methods to get the rest of
them.  Back when crystals were expensive and crystal controlled converters
were becoming popular as the first section of a receiver, some sets did band
imaging for two of the bands to allow a saving of one crystal.

The other common band imaging combination is an IF of 9 mcs with an LO
tuning 5 mcs to 5.5 mcs.   This gives you 4 mcs to 3.5 mcs using IF - LO and
14.0 mcs to 14.5 mcs with IF + LO.

That one, I expect, is the reason that 9 mcs was for a time a very common
receiver/transceiver IF.

It's also possible to use LO + IF and LO - IF.   So, for 80M you can use an
LO tuning 5200-5700 and an IF of 1700 kcs to get 3500kcs -> 4000kcs and
6900kcs -> 7400kcs.   The LO is at a higher frequency so stability is a
bigger challenge but they tune the same direction: The higher received
frequencies are at the same end of the LO dial, which is more natural.   But
sidebands are reversed between the two bands (LSB on 80 will be USB in the
IF while LSB on 40 remains LSB) so you need a BFO that can tune both sides
of the filter center.

Finally, Tim mentioned 'up converting' receivers.   The problem with IFs of
just a few Mcs is that there are ham bands on both sides.   That's not a
problem for a ham-band only set: You just pick an IF that's between bands.
But how do you do continuous coverage?   Say 1.5mcs to 30mcs?

You can get tricky in a couple of ways:   You can leave a 'hole' for the IF
-- the receiver is continuous coverage, EXCEPT FOR ...   Or, you can use two
different IF's, using one of them for most bands and switching to the other
when the receiver is tuned to the band that includes the regular IF.

Both methods have been used but both, obviously, are somewhat ugly.   A
seemingly tidier answer is to put the IF above the highest band used:  Say
at 70mcs.   Then you just tune the LO to cover everything.   For 1.5 mcs the
LO is 68.5 mcs, for 30 mcs it's 40 mcs.   Since the LO is always outside the
tuning range this is pretty clean and no trickery is involved.

It's not hard to build a crystal filter for 70 mcs although you can't make
one that's of SSB bandwidth and has a good shape.   But you can do a fixed
down conversion to a better frequency -- even to (say) 455 kcs and put a
second (sharper) filter there.

A stable LO would seem to be impossible at frequencies of tens of mcs but
modern solid state receivers all use some form of phase locked loop (PLL) in
which a high frequency tunable oscillator is constantly adjusted to a
frequency determined by a very stable crystal oscillator plus a very stable
(low frequency!) tunable oscillator.

So why isn't this an ideal receiver?   Well for one thing a PLL is
inherently noisy.   It HAS to be, because it's being continually tweaked to
keep it on frequency and each tweak is noise.   That noise winds up back in
your receiver and more or less hurts sensitivity.   Beyond that, with an
upconversion then a down conversion to get decent shape and selectivity,
designing for a wide dynamic range (so very large off-channel signals don't
mess up the one you want to hear) becomes much harder.

Well designed modern (solid state PLL) receivers handle these problems
decently but the more decently they do it, the more expensive they are.  The
best 'obsolete' vacuum tube ham band only single conversion designs are
fully competitive in sensitivity and signal handling ability.

Walt 
KJ4KV