[HBR] A Homebrew design that is rarely mentioned

[Walt Hutchens]

Ian posted:

> I compared the mixing schemes with the SX-146, which has the following specs
> for "in-band tweets": less than 0.25uV equivalent CW except for 1uV at
> 21.33Mc.
> 
> The SX-146 also uses a 5-5.5Mc VFO and 9MHz IF. In contrast to the W5OMX
> design, it does no premixing on 80m and 20m. For 15m and 10m the mixing
> scheme used by both receivers is the same. For 40m the SX-146 uses a 21.5Mc
> xtal while the W5OMX uses an 11Mc xtal.

This is another interesting receiver I didn't know about.   There's also the
predecessor SX-133.   I haven't homed in on a schematic yet.

Most sets that use the 5-5.5 Mcs VFO/9 Mcs IF scheme will tune 80 and 20M
without using a crystal at all.   In this case there shouldn't be detectable
birdies (i.e., they'll be of such high order that the receiver won't hear
them) on these bands.  This is the 80/20 'band imaging' receiver design.

The W5OMX design, OTOH, uses premixing with a crystal on these bands too.
As an example of how that can produce spurious signals, on 80M the crystal
is 7.5 Mcs.   7.5 x 2 - 5.5 x 2 = 4 Mcs.   As it happens, the receiver tunes
4 Mcs with the LO at 5.5 Mcs so that's a very fast moving (reverse direction
from tuning) tweet right at 4 Mcs.  Also, 3 x 5.5 - 7.5 = 9 Mcs; that's a
fast moving SAME direction tweet, also at 4 Mcs.

The more frequencies you generate in a receiver, the more potential for
spurious responses and the more care is needed with mixer design, shielding,
and signal levels.  There's no substitute for a careful study of the numbers
before starting, and thinking how the mixing can best be done to minimize
the problems.

The advantage of multiple conversions (instead of using the same frequencies
by premixing and a single conversion) is that when you add new frequencies
one at a time, there are far fewer products to think about.

> I haven't worked out the combinations but this means that the SX-146 tunes
> some bands backwards and has LSB/USB switched between 80m and 40m.
> But that's probably not a bad price to pay if the birdies are minimized as
> a result.

This depends on your use for the receiver and how much simplification you
get in exchange for the birdies.   In a commercial design, pretty much any
detectable birdie is a no-no.   But I have a homebrew band imaging 80/40 M
receiver with a strong birdie in each band that I consider fully
satisfactory:  It has adequate sensitivity with very good strong signal
handling, very stable, excellent AGC and is VERY simple:  7 tubes including
rectifier, no power transformer ...

Generally as  you raise the complexity/cost of a receiver you want to
improve its performance more or less uniformly across all measures:  The
W5OMX is about the most costly/complex receiver a ham would be likely to
build and should have had few to no birdies.   Indeed, as the author
suspects, the basic design would have allowed fewer.

One very simple improvement would have been the use of a double tuned
circuit (per band) in the plate of the premixer.   Use two separate tuned
circuits located near enough each other to get some coupling, peak them both
near the center of the band, then add capacitive coupling to get a double
peak encompassing the middle half or 2/3 of the band.   This will give both
more uniform drive across the band, and reduced spurs.

This isn't really a design for a beginner to build; that level of
sophistication would  have been okay.

Tim commented:

> 7360 Mixer beam deflection voltage is limited by simple capacitive coupling
> from the Hetrodyne mixer. The 10pF series cap has the virtue that it is
> probably a good and simple way to keep LO level constant as frequency goes up.
> It's just that the level is always too low by your or my standards :-).
> 
> Running enough power through the heterodyne mixer to give 10V at the
> deflection plates of the 7360 would probably make even more spurs - not made
> in the 7360 mixer, but from the heterodyne mixer. The heterodyne mixer only
> has a LC tank to tame the spurs.

All true, however had both the premixing and the signal mixing been done
with balanced circuits that could have been much reduced.   20 dB rejection
can be obtained without any balance adjustment and a one time all band
adjustment ought to get most of another 10 dB.

> Applying HBR-XX style plug in coil per band (would take more than a 5 pin base
> though) might be a good way to get balance and more drive voltage to the 7360
> deflection plates.
> 
> In terms of plug in coil sets if we went this hypothetical way... we would get
> rid of RF amp. Have one plug in coil/cap set (double tuned? I'd aim for that)
> for front end coupling mixer to antenna; another one for heterodyne oscillator
> (make it singly balanced to tame spurs even more); a third for coupling
> heterodyne mixer to 7360. That just might work.

Indeed it might.  If you do go with premixing you also need to swap
crystals.   Another way to do it is to use an LO just as for the HBR-series
but perhaps make it push pull?   Not easy, but not impossible, either.

It's also worth looking at slight tweaks of the numbers when you find spurs
that are on a band edge.  50 kcs one way or the other?   Of course such
adjustments need to be checked for all bands: Sometimes they'll fix problems
on two bands and deliver something really nasty on two others!

Ian again:
 
> Is the tuning drive used any more available than that in the HBR?

The 6:36/1 ball drive was used in a number of medium grade commercial sets
and is accordingly not too hard to find in used condition.   Used ones
generally do have backlash although it seems to me that if the grease is
washed out and new grease worked in, then the 'fingers' squeezed ever so
slightly, it ought to be possible to bring these units back to 'like new'
operation.  

I haven't seen a new one recently but Google and eBay might yield something.

Finally, those who want to really dig into dynamic range testing can hardly
do better than the thoughts of John Thorpe of AOR.   The notes at the link
below aren't pitched for beginners but if you have something like an ARRL
Handbook from 1995 or later to get you over the concepts and terminology
then this will take you on from there.   This testing can be done at home
and it will improve your equipment: You just have to know the pitfalls.

http://www.ab4oj.com/test/ar7030comments.html

Walt 
KJ4KV