[HBR] What would W6TC do?

[Walt Hutchens]

Peter said:
> TC would probably abandon the dual conversion design and either go
> to a color burst, or uP crystal ladder filters in the 3.5 to 10 MHz
> range. This could yield good selectivity, and open up the
> possibility of multiple filters for both SSB and CW. Ladder filters
> are pretty easy to make.

And Tim said privately (but gave permission for me to answer on the
list):

> I disagree about keeping the IF at 2 MHz... move it to circa 9 MHz.

My first thought was that 9 Mcs is an obvious winner: You get 80 and
20 with a single oscillator coil ... right? (5-5.5 Mcs LO frequency.)
But no ... that only works if you uncouple the tuning of the RF stage
from the LO, because if you do band imaging this way the bands tune in
opposite directions. Single knob tuning is an important feature of the
HBR series.

To get them to tune in the same direction, the oscillator has to be on
the same side of the IF for both bands which is only possible for
ABOVE: Say 12.5-13 Mcs and 23-23.5 Mcs for a 9 Mcs IF.

If you use an IF of 5.25 Mcs, however, an oscillator tuning 8.75-9.25
Mcs will work for both bands with the added advantage that the normal
SSB transmission modes on these two bands will both appear as USB in
the IF. One set of dial markings will work on both bands, although it
will still be necessary to re-trim the calibration when switching.

The 40 meter oscillator will tune 12.25-12.55 Mcs and 10 will be
covered with 22.75-24.75 Mcs. The 80M oscillator frequency is higher
than for a lower IF, but the others are lower -- probably a good
trade-off.

Doing it that way there's a savings of one coil -- the same oscillator
coil covers 80 and 20 -- and using single crystal (which would have to
be a custom item) for the BFO gets the normal SSB transmission mode on
all the common bands.

Or you could build a tunable BFO. Might possibly do it with a tunable
crystal oscillator.

A more important advantage (compared to an IF around 2 Mcs as I
originally suggested) is that I believe a ladder crystal filter of SSB
bandwidth can be built with 5.25 Mcs crystals.  And the crystals seem
to exist: They aren't common but this company claims to have them:

http://www.opamp-electronics.com/catalog/525-mhz-at49-thruhole-crystal-pack-p-729.html

> I suspect the filter would be based around cheap readily available
> microprocessor crystals. Remember W6TC's ubiquitous goal of using
> commonly available parts.

Agreed. These crystals are $0.15/each. And pretty nearly any ham who
can build a receiver could follow the directions to build a tolerable
ladder filter.

I wouldn't be surprised if one or more of the ham sets used an IF in
the 5.25 Mcs range; if so, salvage would also be an option.

There is still one serious issue: Any superhetrodyne receiver has to
deal with images. And some of those from the 41 M (nominally 7200-7450
Mcs) shortwave BC band are monsters on the east coast. When conditions
are right you can hear World Christian Radio (WWCR and three
affiliates, all with 100 kW, Nashville TN, 7465 and 7490) on a rusty
frying pan here in Virginia; I doubt that an RF section with two tuned
circuits is going to stop such signals. I think that in this part of
the country one would be better off to choose a frequency that puts
the 80M image band somewhere else. WIND separate coils for every band.

If you use separate antennas for each band this may not be such a
problem but I use loops.

6.144 Mcs with an LO tuning 9.644-10.144 for 80M, 13.144-13.644 for
40, and 7.856-8.356 for 20 might be a possibility.  The IF is in the
upper end of the 49M SW band but I've never heard anything close to
that frequency and images from the 19M band appearing on 80M shouldn't
be very strong.

I think this is an area where the best design compromise is unclear.
Using an IF around 2 Mcs with a single half-lattice filter will almost
certainly lead to problems with 41 M stations via filter blow-by
unless excellent IFTs (better than what I'm using now!) are used for
back up.

> It's even likely that instead of trying to get tracking in the front
> end, we would just have pluggable bandpass filters.

I don't think this really works for an HBR. Bandpass filters are
tougher than tracking for a beginner when they actually have to be
'bandpass' rather than just a single tuned circuit of low Q. AND a
tunable front end will definitely help crossmodulation and strong
interfering signal performance on 80 and 40.

> Many (most?) ham band receivers by the 1960's had abandoned tracking
> front ends.

That's true -- they went to a tunable IF with a separate tuning knob
for the RF stage. This preserved performance, allowed conversion
schemes in which the bands didn't all tune the same direction, and
saved money. But in a simple tunable 1st oscillator-fixed 1st IF set,
tracking the front end isn't difficult when you have parts values and
coil winding instructions.

> Tracking front ends and single-knob tuning were a bit of a holy
> grail in the ham homebrew community for those who wore their "my
> homebrew receiver is as good as or better than the commercial
> receivers" chip but this led to creeping featurititis. The best
> homebrew ham receivers should be ham receivers... not copies of
> every undesirable or unneeded feature in commercial receivers.

I agree with the philosophical point that simpler is better as long as
everything YOU need is there, but that doesn't yield a tight
specification. Consider:

>> 2. WARC and SWL band coils would be part of the plan for those that
>> want them. The use of an AM-FM three gang capacitor with jumpers in
>> the coils to connect the larger sections would be an option for those
>> wanting general coverage -- say 2-6-18-30 Mcs.
> 
> I disagree here. I would make the radio would even more ham band centric.
> Eliminate AM detector and make it be strictly SSB and CW. If anything
> the tuning ranges would be narrowed, not broadened. And yes, additional
> coil sets for anyone who wants WARC or SWL bands, but they would all
> be specific to each band and not general coverage.

More ham band centric makes sense, although general coverage by the
route described costs only time for coil winding and perhaps some
additional dial scales. And some hams do find other coverage very
useful; in particular, it's almost essential to have some form of
general coverage receiver around for testing other gear, digital
tuning isn't always best for that, and a receiver that is vacuum-tube
robust is a useful addition to one's emergency preparedness.

AM capability is only 'featuritis' if you don't regularly work AM. I'm
betting that most people building a vacuum tube receiver today will
want to use it at least some of the time to work AM stations.

But this kind of tweaking of the design -- adding or removing features
that alter nothing basic, according to the needs of the builder -- is
not much of a compromise. I usually put a lighted dial on my receivers
because I like to operate in a dimly lit room but that hardly requires
giving up anything else. Some people paint their gear, some use
natural aluminum, some label the controls ...

> I disagree with the unavailability of [power] transformers. Right
> off the top of my head, Hammond 263, 269 series transformers fit the
> bill and are readily available. If you think the plate voltage is
> too high, that's because you're using a bridge rectifier instead of
> noticing that the center tap is still there :-).

I hadn't noticed these transformers although I knew about the Hammond
line. Neither one is perfect but the 269, rated 125-0-125 @ 100mA and
6.3 @ 2A would be usable, maybe with a separate 6.3 @ 2A filament
transformer for help. An 8-tube receiver (with mostly dual section
tubes) using solid state rectifiers and low-power audio stage will
draw closer to 60 mA plate current than 100 but might go high enough
on filament current to push the 25 watt rating of this unit.

> Cheap commercial SW receivers did use series filament strings and
> hot or isolated chassis... and I think at least a couple Knight-kit
> radios did too...

The S-38, right? And others based on the 'all-American 5' design, plus
a tube or two. These were certainly minimal performers ... but the
same number of stages with a transformer would have performed no
better and cost half again as much.

I don't see a logical reason that much more complex sets would not
have worked every bit as well with series filaments and
transformerless HV.
> 
> ... correct me if I'm wrong, the ARRL/QST never published designs
> for radios that didn't have a filament transformer or that ran the
> plate without a transformer. (I know of a few non-radio accessories,
> like CPO's, that they showed without transformers.)

If they did, I don't remember seeing it.

But the times were different. Few houses had a ground wire electric
system (maybe they just had a few three-wire plugs with the ground pin
not connected?) and who knew if the neutral and hot wires were correct
in every outlet? In fact, they WERE NOT always right. There were far
too many ways for a two wire radio with no transformer to put 110 VAC
on the chassis or anything coming out of the set that might be
connected to the chassis or have leakage to the same. It was possible
to do anything you wanted in perfect safety with transformerless
construction but it involved getting a number of details right and I
wouldn't bet that every ham constructor would have done so. I can't
say that I blame the ARRL.

OR (for that matter) the commercial ham gear builders of the day. By
the time TV sets were all series string, NOBODY was designing ham sets
with vacuum tubes.

Today you use a three wire cord with the green wire to the chassis,
you follow the proper neutral/hot convention in wiring and check with
a voltmeter to see that you did it right, and just to be sure, you use
an outlet checker. Everything that comes out of the chassis is at
chassis potential. What's the problem?

(One wrinkle: If your antenna is grounded when in use then it needs to
enter the receiver via a floating link to avoid a ground loop that
will introduce much trash.)

I'd far rather work on a transformerless set than a linear.

But I still wouldn't recommend a transformerless RX design unless you
want to do it for the challenge. I have one that works well (the band
imaging set) and I'm working on an HBR; once that works perfectly
perhaps I'll feel differently but it's still a learning process.

Walt
KJ4KV