[HBR] Re: HBR -- Part 4

[Walt Hutchens]

In part 3, the two audio stages got wired and working and I figured
out that the parts set from which the Eddystone dial and tuning cap
that I'm using came, had a Vackar local oscillator.

The audio stages have essentially no hum.

Yesterday I wired the audio detector -- a plate detector using the
triode half of the 1st audio 19JN8. At full volume there was enough
hum to be audible in the next room.

I fixed one obvious wiring mistake by moving this tube to the next to
lowest voltage point on the series filament string. That definitely
improved things. Then I did what I should have done first: replaced
that tube. Bingo! Hum all gone. Although the tube was 'new,' it had
excessive heater-cathode leakage. It may be satisfactory in
another stage.

One of the issues in designing with series filaments is that the
higher the voltage between the filament and cathode, the more leakage
current there will be. This usually isn't a problem if the cathode is
at signal ground for that stage, but there are a few stages in this
design for which that's not true. The audio plate detector is one of them:
for SSB and CW the BFO is coupled to the cathode and any AC current
flowing between the filament and cathode will also be injected there.

Interestingly (at least if your interests are a bit warped!) the local
oscillator stage has the same issue, even though the cathode is
bypassed to ground. That's because the LO circuit is so sensitive to
disturbance that even having heater-cathode leakage shunting the lead
from the cathode inside the tube to the point where it is (RF)
grounded to the tuning cap frame causes a significant change in tube
characteristics. In most oscillator circuits there will be enough
change to cause a few CPS FM on the oscillator, when operating on the
higher bands. This will be noticed as a warble when receiving CW and
as distortion on SSB voice signals.

Because of this issue the LO stage has to be at the cold end of the
series string and when the tube has two sections with filaments in
series (as in the 19JN8) the oscillator section should be the cold
one.

The Vackar circuit allows full control of coupling between the tube
and tank circuit, but because the ECO circuit doesn't have a very high
oscillator gain, I doubt that very much decoupling will be possible.
However, this is an issue that W6TC dealt with -- it's a major reason
a high Q oscillator coil and high-C design were chosen. It may be
okay.

(The higher the tank Q, the less is the effect of a change in tube
characteristics.)

The tickler feedback oscillator of the W6TC designs allows control of
coupling at the plate, but the tank itself is straight across the
grid.

A better design would use a high-gain separate triode oscillator --
one of the UHF TV oscillator triodes like the 6CW4 (Nuvistor) would
probably be excellent. Best of all, of course, is a crystal controlled
1st converter to a much lower tunable IF as used in advanced
communications receivers of the 1960's: crystals have Q's in the
thousands, so significant changes in tube characteristics are
tolerable.  But of course these would be more complicated designs.

When the set was switched on, the 117Z6 rectifier would flash one of
the two filament sections to near white brilliance.  NOT GOOD for the
life of the tube. Since our line voltage is around 122, I stuck a 100
ohm 1 watt metal film resistor in series: that reduced the problem by
half AND equalized the brilliance of the two halves.  Still not
great, but certainly much better.

With the resistor the (hot) filament voltage is reduced to 114. Plenty
high enough.

The only thing I can figure is that the strands of filament that
actually touch the cathode sleeve don't heat up as quickly and thus
stay at low resistance longer, leaving the center strands to absorb
most of the voltage and overheat, until the cathode sleeve begins to
get hot. Any difference at all between the two (series) filaments
would cause the same to occur in one section relative to the other. By
limiting the inrush current (the cold resistance of the tube is about
245 ohms) the series resistor slows the heating enough to allow better
equalization.

This problem is probably much more serious with some makes of tubes
than others but I only have two and they're of the same make.

Because the mechanics of posting with a mailer that isn't the one I
normally use are time consuming I'd rather write fewer longer posts.
However, since I also have trouble with getting tagged as spam for
content (the binary search and multiple posts of the other night) I'll
try to post each day that something actually gets done.

Hopefully that was a one-time event.

Next: The AGC circuits and winding of the 3rd IFT.

Walt
KJ4KV