[HBR] Another one-week (okay, probably not!) HBR -- Part 1

[Walt Hutchens]

Long time listmates will remember when I tried to build an HBR
generally along W6TC lines in a week.  It actually took about five
weeks and even that didn't include making the wooden cabinet.

Ever since then I've wanted to try again. Obviously one motivation was
to 'get it right, this time,' but besides that, I figure that every
ham ought to have a receiver with plug in coils for each stage and one
using an Eddystone 898 dial. I had the dial; when I discovered that
1-1/4" low pressure PVC pipe fits perfectly on the bakelite base of a
metal octal tube, the building bug bit.

I don't really trust solid state devices, so I'm not going to use
them. (He said, typing on his laptop ...) Oh yes, this process is
entirely rational. I'm not sure why you question that?

Furthermore, I've always liked 'transformerless' designs. With modern
three-wire cords (chassis grounded) and high-quality parts they're as
safe as the conventional approach and there's a big saving in weight
as well as some in complexity and heat generation.  There are design
issues (RF grounds return to the chassis but DC and audio grounds go
to the B- buss, heater cathode leakage must be managed by the layout
of the filament string) but once solved, they're over.

Untold jillions of transformerless vacuum tube AM/FM radios and TVs
were built.  An all-band ham set isn't THAT difficult.

I considered the W6TC double conversion scheme but that would mean
another mixer and oscillator: more parts and (because the sharp filter
is farther from the antenna) more difficult overload issues. However
with a high (single) IF, a crystal filter is necessary. After a couple
of hours of junkbox diving, I found a 1700 kcs half-lattice pair of
HC-6/U crystals left over from another project.

With this IF, the oscillator tunes 5200-5700 kcs for 80, 5200-5700 for
6.9-7.4 Mcs coverage (same oscillator coil can be used on both bands),
6100-6350 for 13.9-14.4 Mcs, and so on, differing only in the details
from the W6TC scheme.

(The W6TC choice of a double conversion scheme probably was driven by
the fact that the higher single IF requires much greater care to keep
the BFO signal away from the IF input. Furthermore, it requires a very
carefully designed and built LC oscillator for the BFO, plus
adjustment of a crystal filter to get best results. Neither of these
circuits can be 'wire and go' like a double conversion IF. The
modestly experienced builder, however, need not be bound by these
considerations.)

The 'second draft' tube lineup was:

19JN8 (pentode-triode) RF + 1/2 mixer
19JN8 1st IF + 1/2 mixer
(Mixer is a push-push double triode design.)
12AT7 LO + AGC clamp diode
19JN8 2nd IF + AGC plate detector
19JN8 1st Audio + Audio detector (BFO injected for SSB/CW)
12AU7 2nd Audio + BFO
117Z6 (dual rectifer diode) plate and bias rectifier.

The main filament string adds to just 100 volts; that will require a
series resistor. However the plate currents total over 100 mA and
since the 117Z6 is rated at 60 mA/section, both of them will be
needed. In addition, (oops!) the mixer plate and 1st IF plate are at
the same frequency: They cannot be in the same envelope.

The 'third draft' lineup is:

19JN8 (pentode-triode) RF + 3.5 Mcs band edge marker
12AT7 Mixer
(Mixer is a push-push double triode design.)
19JN8 1st IF + LO
19JN8 2nd IF + AGC plate detector
12AT7 AGC clamp diode + BFO
19JN8 1st audio + audio detector (BFO injected for SSB/CW)
12AU7 2nd Audio + bias rectifier
117Z6 (dual rectifier diode) plate supply rectifier.

That's one more tube, solves the problems noted above, gets W6TC's
band edge marker in there, has just a couple of watts more power
dissipation, and a tidier front end layout. With filaments totalling
112 volts a series resistor of 50 ohms, 5 watts, or so will be
adequate.

The use of sharp cutoff tubes in the RF and IF stages is questionable.
Unfortunately there are no remote or semi-remote cutoff tubes (that
I've been able to find) in a tetrode/triode or pentode/triode
combination with 150 mA heater. With low IF signal levels and a
crystal filter in the first mixer plate circuit the potential overload
problem is confined to the RF stage where it can be managed by
reducing antenna coupling.

The Eddystone 898 is a good looking, very smooth, high ratio dial --
about 55 turns of the knob to rotate the capacitor 1/2 turn, making it
near ideal for tuning in SSB signals across a 500 kcs band and the
flywheel makes it possible to QSY quickly. The knob drive is of the
pinch type, so there's no precise ratio but the slide-rule type dial
is gear/dial cord driven with about 7" tuning range. There's a logging
scale with a calibrated dial having 500 divisions, end to end. The
dial scale is sub-mounted about 1/2", allowing edge lighting. This
can't have been a terribly expensive dial to build, when it was new --
maybe 3x the price of the National ACN and similar designs?

That's the end of the good news. The more obvious news is that this is
a remarkably ugly dial with which to build a radio. The knob shaft is
2-3/8" below the output shaft which puts it too close to the table if
the tuning cap is mounted on the top of a 3" chassis. But you can't
put it there anyway, unless you cut away most of the front of the
chassis to allow for the flywheel on the knob shaft AND the sub
mounting of the dial scale. Alternatively, you can space the chassis
back from the front panel and deal with the issues that creates for
the mounting of the controls.

By raising the capacitor shaft about 1-3/4" above the chassis, the
problem is reduced to cutting a 3" wide x 1-1/2" deep x 1-3/4" high
notch in the upper front edge of the chassis to clear the flywheel.
That's the approach I'll take.

That's not all the ugliness of this design, though. The flywheel must
be removed in order to install one of the mounting bolts for the dial
assembly, so an inch or more of clearance is required behind the
flywheel under the chassis. The dimensions tend to be 'about ...'
making layout more interesting.

I missed the MG and Jaguar experience; building with an Eddystone dial
will help me feel better about that.  Perhaps what I really mean is
"remember WHY I missed the MG/Jag experience."  I don't wonder that
W6TC designed around the American made planetary drive dials even
though the tuning rate on the single-speed ones is much too fast.

The no-name (possibly Electroniques?) three gang tuning cap that came
with is of very good quality and about the right size, so I'll use it.

I have some fairly small IFTs that will work and only three are
needed. Time to lay out the chassis. 8 tubes and three coils should
fit on 10" x 12", right? I need to find a 3" speaker and a REALLY
small meter. With a 12" chassis, the panel will be 14" wide by say 8"
high ... hummm ...

And finally, a question: There's no important handedness in the parts,
the most often used control on a receiver is the tuning knob, and most
of us are right handed. Yet all the HBRs I looked at on the web site
(WOW, are there some BEAUTIFUL examples there!) have the dial on the
left, making it necessary to reach across in front of everything else
to tune. Am I missing something?

Walt
KJ4KV