[HBR] HR-10 to HBR project

[Walt Hutchens]

> With nothing ahead of the 2nd IF, the receiver can hear a GDO (tuned through
> the 1681 kcs IF) nearby on the table and  holding the GDO near the grid
> circuit kicks up the S-meter.   That's good enough for now:  I'll go on
> wiring up earlier stages in backward order.  <SNIPPED>
> 
> Maybe another week to have it all hooked up?  IF there are no big surprises.

It's wired and it works.   Actually, it works very well.   The three 6EH7s
give plenty of gain and they're perfectly stable.

The mixer had to be shielded as it was picking up way too much noise at the
IF.  

There was no problem at all with the AGC plate detector sharing an envelope
with the 100 kcs calibrator.  (Triode and pentode sections respectively.)

The AGC circuit was stolen from the (Yaesu) Tempo ONE.   In the original,
half of a 12AX7 (amplification factor 100) is used as the AGC tube and the
last IFT is tapped so 3/4 of IF output voltage goes to AGC but only 1/4 to
audio output.   In previous copies of that circuit I've done some equivalent
but 'simpler is better.' and the Heath IF transformers are not over-friendly
to modifications.  In this project I took the AGC voltage from the plate of
the 2nd IF and the audio from the transformer output.

 I first tried the 6EA8 that came with the HR-10 (as AGC detector) but that
triode has an amplification factor of 40.   The AGC gain was insufficient as
shown by low S-meter deflection and distortion on strong signals due to
overloading the product detector.    I increased the AGC load resistor (it's
the plate resistor of the AGC plate detector) from 1 meg to 2.2 megs: That
helped.   Adding a 100 mf bypass on the detector cathode resistor gave a big
boost to transient response.   Studying the RCA tube manual turned up the
6JW8, same base as the 6EA8, same 0.45 amp filament, similar pentode
characteristics but the triode has an amplification factor of 70.   With
these changes the AGC works fine.

An even better choice in this socket might be the 6CM8 which has a triode
with an amplification factor of 100 but that one would have required a
slight rewiring.   Depending on availability -- I don't like to use rare
tubes -- I may try it later.

(Don't you wish you had the freedom of an RCA engineer?  Imagine a tube for
a new design -- say half of a 12AX7 combined with a 6BH6 giving you a
triode-pentode that would do jobs like this for only 0.3 filament amps and
call up the tube designers and ask for it?   Some very handy tubes just
never got made.)

With bypassing of the diode to get rid of the buzz in the audio, the 6AM8
works fine as a combined BFO and AGC bias rectifier.

No trace of hum or warble; the BFO doesn't lock with strong CW signals
meaning that (because it isn't pulled) there won't be distortion from that
source on voice.   

The 1st IFT (drives the crystal filter) had a frozen slug and one that made
no difference when twisted.   I opened the can, freed the slug (threaded
slug in lacquered paper form), and reduced the primary capacitance from 62
mmf to 47 to allow resonating with just a bit more inductance in spite of
the higher capacitance of the 12AT7 mixer.   The peak is VERY broad so
changing the cap may not have really improved things.

More gain could be had with a better match -- say by connecting the mixer to
a tap on the coil and this could be done by turning the transformer around
and replacing the tap for the two crystals with two capacitors.   However
there doesn't seem to be a need for that.

I've  had the receiver on for a few hours over the last day, listening to
80M (only coils I've so far wound for this set) and it works well.   Very
little drift, the transformer gets only slightly warm to the touch --
105-110 degrees?  -- so no problem there in spite of increasing the number
of tubes by three. 

Voice signals were just as clear during the activity peak last night (in
spite of the contest in progress and lots of static crashes) so the dynamic
range is okay.   (Quick test of any receiver: Listen on 80 during such
conditions and also at 3:00 AM -- if signals are mushy, hard to understand
when the band is crowded with strong sigs but crisp when there are just a
few, that is an overloaded (distorting) receiver, most likely the RF stage
and/or 1st mixer.)

I did the antenna and mixer coils without the APC trimmers of the classic
HBRs;  the 80M mixer coil works fine with a properly chosen fixed cap and
the antenna coil is taken care of with the front panel Ant Adj control.
The oscillator coil needs a trimmer to allow precise tracking to the dial.

There is a just detectable change of LO frequency with high AGC voltages;
this would be due to changes in the (unregulated) B+.  That calls for a
design tweak to the LO plate supply and maybe the oscillator itself.   Quick
check for this:  Tune in a steady carrier and snap off the power switch:
How much does the note change before it fades away?

It's my impression that the BFO drifts a bit.   This needs to be checked out
-- it could be just the slug moving in the paper coil form after it is
adjusted.  

I will work on 40M coils later today.   The crystal filter load resistor and
tuning may be tweakable for a flatter passband.  It needs a standby switch
and muting arrangement.  I want to add an audio filter for CW.   I'm
thinking about a cabinet design -- not easy, considering the needs for
ventilation, access to coils, and protection.

After a few more days of use I'll try some performance measurements.   A
receiver of this type will never equal either the stability or selectivity
of (say) a good R-390A, let alone the quality of the dial mechanism but
there's no reason it shouldn't copy all the same signals on the bands it
covers.   And for under $100 (maybe a bit more if you have to buy all the
tubes) it's shaping up as a great homebrew project.

Walt
KJ4KV