[HBR] The long, SLOW HBR project

[Walt Hutchens]

> Go Walt!!

Thank you!  I'll keep chugging along as long as possible.

> I bought the 19JN8's a few years ago.

I'd suggest looking for the 19J6 as well; there's a young guy out there
Hoovering 'em up for a project he makes and sells, so vintage constructors
have competition for that one and it may take a bit of time.

The 12AT7 might be a substitute but I haven't tried it in this set.  A VHF
power tube (19J6) is a better choice as an HF mixer ...

Also chase down whatever you plan to use as an IF filter.  I had two
crystals that I had custom made about 15-20 years ago for a previous
project: 1700kcs + and - 750 cps, for a half-lattice filter but a real 1650
kcs (whatever) crystal filter would be even better.

And the usual "can't start without it" items -- the dial and tuning
capacitor.

The rest can be bought (or made -- the coil forms) as needed.  The modern
small parts are so far superior to what we had in the 1960 time frame that I
wouldn't consider going vintage underneath.

Current progress:  The grounding in the RF stage has been redone.   The
leads in a plug in coil set can't be as short as you'd like but I've got
them down to the minimum and untangled (I think!) the possible ground loops
between RF stage input and output.

Strange problem found and fixed: Intermittent loss of audio, caused by loose
pins on the octal socket for the 117N7.  Top quality military socket; don't
know what that was about, but it responded to tightening the pins by
carefully forcing a jewler's screwdriver down each hole in the bakelite base
to collapse the contact.

I'm waiting to catch the VHF oscillation in the mixer (if that is what it
is) in the act.   Right now it shows up only as a brief burst of noise when
the set starts to warm up, followed by a few loud clicks.   Then the audio
is muddy for five minutes or so.   I need to try to get it to act up with
the chassis upside down: If touching a mixer grid with a lead pencil makes
the problem go away then the diagnosis is made and I can throw the soldering
iron at it.

I have done some work on the frequency shift of the LO coming from AGC
action.   The shift of frequency by 20 cps (on 80M) comes from a 10 volt
change of osc. high voltage as the AGC voltage goes from zero to cut off.
Going back to the other side of the filter circuits (rectifier cathode) the
voltage change is only 4 volts so I could use a separate filter to get that
much reduction.   Since the oscillator draws only 1 mA that's not a
difficult filter.

But a 60% reduction probably isn't going to be enough for 20M and higher.

It is sometimes possible to change the various resistor values in a pentode
oscillator (with screen serving as plate for the oscillator and the rest of
the tube as a buffer) so frequency is nearly immune to HV changes; I will
look at that.

Beyond that the only option I can think of is a gas regulator.  Most of them
have current ranges 5 mA to 20-25-30 mA depending on the tube.   That's
quite a bit of extra power, especially when you consider that they can't be
operated too near the minimum because if some operating conditions take the
tube lower, the discharge becomes unstable with large current (hence
voltage) excursions.

There is a subminiature gas regulator, the 5644, that would be useful in
some applications but it is in that same power range.

You might think of using a neon bulb as a regulator but the ones designed as
indicators are too unstable in the best of circumstances.  The 991 (used I
think in the LM frequency meter) might be an option but is a hen's tooth
item.

Then I remembered the 5651 voltage reference tube.   This is a neon gas
regulator intended as a reference in high power high accuracy regulators.
Typically it was used to control a small pentode driving a 6080 (low mu
triode) in a series regulator circuit to supply A LOT of HV current to
equipment requiring stable voltage.  One application was in shipboard radar
displays.

The 5651 regulates at about 86 volts, current range 1.5 to 3.5 mA, so it
should work perfectly for an oscillator at 1 mA in a conventional shunt
regulator circuit.  Striking voltage is a bit over 100 volts, easily
available in a line operated set.

In order to get reliable operation, such circuits have to be designed so
that with the minimum output from the rectifier and the maximum load from
the set, there's enough to strike the tube.   You can't rely on the HV surge
you often get at turn on because if the set is snapped off and then back on
before the tubes cool, near full current may be drawn before the rectifier
reaches full output, so you get a gradual rise to final voltage, with no
surge.

Anyway, there are now a couple of ways to go.

Walt
KJ4KV