[HBR] The long, SLOW HBR project

Thu Sep 1 09:33:47 EDT 2011

Walt.  I want to thank you for your posts concerning your receiver. Your
progress has been an education, and I look forward to hearing more!  I am in
a similar situation with my receiver, it seems that unexpected stuff, always
rises up to add to the learning curve.. In my case, my HBR is based on an
R392 receiver...all is going very well but I had a 'motorboating\,
oscillation in the IF that I could not get rid of--I am running 26 volts on
both plate /fils.  so power supply decoulping was my first order of
concern--plus the fact that there are 6 IF stages...I expected some
problems, but bypassing the daylights out of everything would not make it go
away....long story, short, the problem was in the power supply....RFI hash
from the bridge rectifier was gating the agc-and it was getting into the
radio--not on the power supply rail, as you would expect-but coming in
through the antenna-just like any other signal would.  Drove me nuts for a
bit, as I would pull the antenna off and the motorboating problem would go
away!
Anyway, 4 .01's around the bridge took care of 95% of the problem.  At least
I know what I'm fighting!!
--Keep up the good work--looking forward to more posts!
de Don Haworth wb0rai Nashville

On Wed, Aug 31, 2011 at 3:06 PM, Walt Hutchens <waltah at earthlink.net> wrote:

> > I'll probably do a 20M oscillator coil next, just to see how that goes.
>
> Done.   But what a bunch of work, with more to come.
>
> Since I'm not building an established design (new oscillator circuit and
> tracking tuning to an existing dial)  AND I'm making the coils from 1" Sch.
> 20 PVC pipe the coil work is t-e-d-i-o-u-s.   I'd guess 3 hours plus per
> coil, down from about 10 hours on the first (80M) oscillator coil.
>
> And -- as usual -- this was more educational than intended.
>
> The W6TC (and many other multiband superhet) designs operate the LO on half
> the needed frequency.   A trick they use that I've never seen discussed is
> that feeding the LO to the same mixer grid as the signal lets the mixer
> grid
> tuned circuit help select the harmonic signal over the (unwanted)
> fundamental.
>
> This tuned circuit won't be right on, but considering the fact that
> harmonic
> injection is used only on 20M and up, it's not that far off, either.   The
> impedance of the mixer tuned circuit will certainly be much higher on the
> harmonic than on the fundamental.
>
> Use of the second harmonic has the advantage that wiring and tube
> capacitances and inductances, transit time, and other unwanted effects in
> the LO are only half the percentage of the total circuit values, meaning
> they have a smaller (percentage) effect on frequency.   A few hours
> experimenting with a 12.3-12.78 Mcs local oscillator (existing dial is
> calibrated 14-14.48 Mcs, IF 1.7 Mcs) convinced me that I wanted that
> advantage.
>
> Use of the 19J6 dual triode mixer with signal on one grid and LO on the
> other made that impossible.  It's a great circuit but everything is a
> compromise: I replaced it with a 19JN8, using the pentode section as a
> mixer
> and connecting both signals to the control grid in a manner similar to the
> HBR-series.
>
> These tubes are always run very close to cutoff: Cathode resistors are 2400
> to 12,000 ohms, depending on which design you look at.  I picked 2200 ohms,
> pretty much out of thin air however a brief tour of the various handbooks
> would be a good idea.
>
> You'd expect some reduction in ability to handle very strong signals but I
> haven't noticed anything yet.
>
> One problem that resulted is that the former smooth slightly-humped-at-the-
> edges passband is now two pronounced peaks with a valley in between and
> audio quality has suffered accordingly.   This is most likely due to the
> much higher plate resistance of the pentode mixer driving the half-lattice
> crystal filter.
>
> With the mixer circuit change I was able to redo the 20M oscillator coil to
> cover 6.15-6.39 Mcs.   Other than more than the usual number of assorted
> wiring and arithmetic  mistakes (but practice with a slide rule is a good
> thing!), that went smoothly and the oscillator is certainly more stable.
>
> At least the antenna and mixer coils should be straightforward, right?
> Err
> ... 80M being the broadest band (percentagewise) I was able to do those
> coils by connecting the tuning cap across the whole coil.   And I neglected
> to provide for bandspread on the other (narrower) bands.
>
> One can either use a capacitor in series with the tuning cap or tap the
> coil
> and hook the tuning cap across just a part of it.  Each way has advantages
> but  the tapped coil seemed best in this case:  Fewer hard to find parts
> (silver mica caps!)  are needed and it allows me to start by copying the
> antenna and mixer coils of an existing HBR design.
>
> Using a tapped coil, the tap has to be brought out for connection to the
> tuning cap.   A few hours were devoted to rearranging the set wiring and
> making new 80M antenna and mixer coils to conform.  Other minor things got
> tidied up at the same time.
>
> The next few hours went mostly to making up the 20M antenna and mixer
> coils.
> Using the coil info for the 'Deluxe HBR' in the Bill Orr handbook these
> coils came out right the first time.   They peak up at the right point and
> tracking is in the ballpark, certainly tweakable to perfection.
>
> The new problem is that gain on 20M is completely inadequate:   The 3.5 Mcs
> band edge marker nearly pegs the S-meter on 80; the 4th harmonic on 14 Mcs
> is only just audible and doesn't move the meter; only the strongest 20M
> sigs
> can be heard at all.
>
> The overall gain of this receiver is less than for the real HBRs.   That's
> both a principle -- a whole bunch of gain you won't use is a whole bunch of
> potential trouble you don't need -- and a consequence of designing with a
> minimum number of stages and including a crystal filter having some loss.
>
> It's not a big deal to boost the overall gain and it is time to do so.   I
> think the double-hump passband of the filter can be fixed by better
> matching
> to the mixer (more turns on the primary of the 1st IFT) and that will also
> increase the gain.   That's next.
>
> There are a couple of other things:  The IF and RF stages can be run at
> higher gain.   (The 19JN8 pentodes are comparable to the 6BZ6 in gain -- in
> other words, plenty hot enough for a receiver of this type and number of
> stages.)   They're designed to work in FM receiver front ends -- even 10M
> is
> no stretch, frequency wise.
>
> The set is rock stable now -- no trace at all of regeneration -- so gain
> increases shouldn't be a problem.   (Yeah ... that's what he said ...)
>
> Raising the gain will force improvement of the AGC and/or BFO level
> (whatever is the problem causing roughness on strong signals) but that has
> to be done anyhow.
>
> When the gain is ample on 80M (rather than 'enough' as at present) it'll be
> time to deal with the much lower gain on 20.   That almost has to be weak
> LO
> injection.   There's lots to look at in that department.
>
> The place to deal with band-to-band gain variation (that can't be fixed in
> the LO/mixer circuitry) is in the coils.   I'm not eager to wind ANOTHER
> set, but if that's what it takes ...
>
> About the only good news from all the practice coil winding is that it is
> easy to salvage the octal tube plug and APC cap from those that are
> replaced
> -- I just cut off the top and bottom of the coil form and peel off the PVC
> from the critical part.   Both the PVC cement on the APC and the JB Weld on
> the plug work perfectly -- rock solid in use but come away clean for
> recycling purposes.
>
> Even the wire from discarded coils can be recycled into coils for higher
> bands!
>
> This was definitely the time to tackle a higher band as new issues were
> exposed and possible solutions to a couple of known ones were narrowed.
>
> Walt
> KJ4KV
>
>
> ______________________________________________________________
> HBR mailing list
> Home: http://mailman.qth.net/mailman/listinfo/hbr
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:HBR at mailman.qth.net
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>