[HBR] HR-10 to HBR project

[Walt Hutchens]

And the learning continues ... unfortunately.

The 40M oscillator coil was wound without incident.   Fortunately the HR-10
follows the pattern I prefer, operating the LO above the channel frequency
on 40M.   Arguably this gives a bit less stability but it keeps the
sidebands in the same places as on 80M meaning you don't have to retune the
BFO which changing bands.

This local oscillator is plenty stable, probably because (A) the chassis is
steel and thus doesn't expand and warp very much, and (B) the coil is out of
the most heated area and gets plenty of ventilation -- I used the plug in
calibrator socket for the coil.

Things got a little trickier with the antenna and mixer coils.

Since the HR-10 coils are on 5/16" paper forms and I'm using 1-1/4" OD PVC
pipe I used the HBR coil specifications for 80M.  No problem there --
everything came close to working as wound.

I tried the same with 40M and wow was THAT wrong.   Stick a 40M mixer coil
per HBR specs in with the HR-10 fixed and variable caps and it tunes a chunk
of 80M.   I'm ashamed to say that not only did this happen but it took me
half an hour of head scratching to figure it out.

A basic fact:  With a given tuning cap and range (in this case 7.0-7.3 Mcs)
the fixed capacitance is ... fixed.   And once you've got those parts hooked
in parallel, the coil is also fixed.   But what I found was that the HBR
coil was much higher inductance than the one for the HR-10.

The answer is that the HR-10 DOES hook everything in parallel.   The fixed
(parallel) cap of 100 mmf is just right to give the proper tuning range with
their small coil.    The HBR designs, however, use a much smaller fixed cap,
a much larger coil, AND ON 40M, THEY TAP THE TUNING CAP DOWN on the coil to
get the proper range.  (The 80M coil puts the tap just short of the very top
of the coil -- that's why I didn't notice this on that band.)

The Heath method was simpler but a lower inductance coil is almost certain
to be lower Q.   A key part of the HBR design is very high Q coils in the
front end for both selectivity and sensitivity.

So, I have to rewind the 40M mixer coil, connecting the tuning cap to a tap
at a location to be determined.   Ugh ... but the good news is that the
tuning range is small and tracking doesn't have to be nearly as good as for
the oscillator coil.   A Q of 200 (it won't be that high) would give a half
power bandwidth of 35 kcs on 40, whereas the oscillator should follow the
dial within a kc or so.

Because the oscillator has to track so closely (in order to use the existing
dial) it's best by far to use the identical circuit and tuning arrangements:
Except for the substitution of an ECO for a straight triode oscillator I
have done that.   But there's much more flexibility for the antenna and RF
coils -- and once one of them is right, the other is just a copy.

In other news the set suddenly developed slightly raspy audio.   An hour of
troubleshooting homed in on a new warble coming from the LO and replacing
the 6BH6 tube eliminated the problem.   That's a Hartley ECO so
heater-cathode leakage will do this.

This stuff is a LOT easier to troubleshoot on this set than many others
because the two semi-balanced mixers (the product detector is also one)
serve as cathode follower buffers if you disable one input.   So to check LO
injection (waveform, presence of hum, frequency ...) just disconnect the
antenna and hook the appropriate instrument to the cathode of the mixer.
And similarly at the product detector for the BFO (pulling an IF stage if
necessary), as well as the ratio of BFO to maximum signal (BFO should be
several times greater to avoid distortion) .   I have above chassis test
jacks at both points.

I'm betting it'll take three tries to get that 40M mixer coil right.  But
the PVC form coils can be just pushed together (nothing glued or soldered)
for such testing so at worst it costs you a bit of wire and a couple of
extra holes in the coil form.

I do have to change the circuit for the 80M coils which means a change down
inside the coil form but that's no different than a builder using molded
forms would have to do for every step.

Walt
KJ4KV