[HBR] Another HBR Project -- Chapter 12

[hbr@mailman.qth.net]

"This whole thing has sure been a lesson to me."

Getting enough injection on 20 meters turned out to be simple.   On 
this band the 1st oscillator tunes 6.1175 to 6.4175 and the second 
harmonic is used to cover 13.9-14.5 Mcs, however looking at the 
injection waveform, the 2nd harmonic was fairly low.   The solution 
was to add a choke across the ECO output to resonate it at the 
second harmonic.   The choke's inside the 20 meter coil set -- uses 
a previously unused pin -- so it's only in the circuit on that band.    A 
fix tuned circuit could be used on higher bands if desired.

But the gain on 20 was still very poor -- maybe 1000 uV for S-9?

I *knew* that the L/C ratio sets the gain of a stage (particularly a 
pentode) but it still took me four tries jacking up the L/C ratio to get it 
to the practical limit -- just short of where stray capacitance (of which 
there's a lot, the way the tuned circuits are laid out in this set) 
makes it impossible to get an adequate delta-C to tune the desired 
range.   

But even at the limit it was still 'numb' on 20 ... maybe 300 uV for S-
9?  The sensitivity was fine -- about 0.3 uV for 10 db s+n/n, but very 
poor gain in the RF stage.

I woke up in the night with an inspiration:  I had used a cracked and 
repaired toroid for the RF plate coil -- maybe the Q was ruined?    
This morning, I wound a coil on a new toroid.   No change so it 
wasn't the coil ... but gee, the Q *did* seem low, indicated by a 
broad shallow dip on the GDO.  

The only other part in the coil assembly was the fixed shunt 
capacitor.  I swapped that for another make of the same value, and -- 
ta-da!   The Q shot up and when I replaced the cap on the antenna 
coil (same make) and re-tweaked both coils, the gain did the same.   
S-9 is now 50 uV which is perfectly reasonable.   

I had several more of the identical caps and I tried three of them in 
parallel with the same coil.   All bad.   Sometimes fleamarket cheap 
parts really are just trash -- Walt's third law of homebrewing?   Three 
days work that $1 worth of caps cost me.

I need to look at some other projects and see if any of those caps 
got into them.   They're probably *all* bad.

While working on the gain issue, I discovered yet another place 
where an 85 kcs stage had an 0.01 mfd bypass cap instead of an 
0.1.   Replacing it didn't cure any obvious problem but it made the 
receiver sound slightly different so there must have been some 
regeneration there.   I note that when I turn the volume up fully there's 
a trace of a howl which is tunable by the VFO; that has to be 
feedback in the 85 kcs IF section.   You don't notice it in operation 
because of AGC action.

While polishing my 20 meter coil winding skills, I found the answer to 
one annoying problem.  Many command receivers have ceramic coil 
forms, making a Hartley oscillator coil difficult because you can't drill 
the form for a center tap.   I found a diamond burr for the Dremel tool 
that will grind a hole about a 3/32" diameter in the form.   I don't think 
it will last for very many holes, but I'm very much hoping not to need 
many forms.

The STANDBY switch which takes the AGC line to cutoff via a 
steering diode scheme works -- except -- the diode that should 
isolate the AGC side of the circuit when on STANDBY is just leaky 
enough for the caps there to slowly charge to cutoff, so if you leave 
the receiver in STANDBY for a while, it comes to life very slowly.   
Back to the back of the envelope -- I mean, the drawing board.   I'll 
try changing the diode, but probably I just need to hard switch the 
line from the AGC detector to the cutoff bias.

(Another interesting trick that I've seen in some military sets is to 
switch to an adjustable bias which is set to provide just enough gain 
for sidetone.   Most hams don't use sidetone, but once you get used 
to it, it's pretty nice.)

(Another possible approach would be to switch in a large value 
resistance in the RF state cathode, leaving the AGC functioning.   
Depending on the resistor value that would also give sidetone but 
with AGC action.)

With the compensation cap at the oscillator tube soldered by the 
shortest possible lead to a short ground lug, the set drifts up 
(overcompensated) maybe 200 cps in 15 minutes or so and is then 
approximately stable.   I can lengthen the lead to the grid pin on the 
tube socket -- where most of the heat comes from during early warm 
up -- by maybe 1/4" and perhaps get it spot on.  When changing 
coils (after warmup) to another set from the drawer, it goes down 500 
cps or 1 kcs over half an hour and is then pretty stable.   Fixing that 
will require moving some compensation from the coil socket in the 
radio to inside the coil set.  I don't think I have the right caps for that 
... the command set ones (3 mmf/N750) are about the right value but 
too big for the space inside the coil can.   

Right now I have a 3.5 mmf/N1500 on the back of the coil socket.

Those are 80 and 40 results; the drift on 20 is much worse because 
of the high L/C ratio.   It's a trade off ... I haven't tried to do anything 
about it yet but I'll be surprised if I don't wind up with more drift than 
the low bands.

Another advantage of not having the 1st oscillator track the signal is 
that you can set the osc L/C ratio independent of the front end.   Of 
course in a more generous design (not a 'simplicity' approach, like 
this) you'd not need the RF stage to work at maximum gain so you 
could use a higher C arrangement.

Remaining work:  20M drift, fix the STANDBY circuit *again*, wind 
160M coils and possibly some for an SWBC band.   Calibrate the 
dial and figure out how to print it from the computer -- I had a way to 
do that a few years back, so probably I can do it again.  I don't think I 
have the headphone jack wired quite right.  (How-can-something-so-
simple-be-so-hard department!)   I'll probably do something to 
balance the gain across the bands -- maybe put the RF stage 
cathode resistor in the coil set.   Get rid of the 85 kcs feedback.

I'm convinced my S-meter is the same make Ted Crosby used on the 
HBR-16 shown on the web site.   The face is identical except mine 
has 'CalRad' where his meter has 'S Units.'   But mine has 'S Units' 
above the topmost scale.   It probably was imported by a different 
U.S. company but came from the same Japanese manufacturer.  It's 
dated 1958 in teeny-tiny hand lettering on the dial.   It looked new, 
but it wasn't in the box and the lugs had been soldered to; maybe it 
came from an old HBR?

I have two more matching meters, one 0-100 uA and the other 0-300 
mA.   So I guess that settles the question of whether a companion 
transmitter might eventually be required.   

The TC HBR-16 front photo from the web site makes great Windows 
wallpaper. 

I'm quite pleased with the 1MHBR.   It's a *good* receiver -- allowing 
for the plug-in coils and relatively coarse calibration (high end sets 
had 1 kcs markings), it's better than anything I know of in the 1960-
70 ham marketplace and due to the extensive use of command set 
parts it's certainly cheap enough to build now -- if you have some of 
the tubes the Radio Shack transformers are about the only costly 
parts.   I think it would be an even better performer with the extra IF 
stage and reduced RF gain (somewhat higher overall gain but less 
ahead of the 1st mixer) but it's just fine as it is.  I wish I'd had the 
sense to either leave out the 100 kcs calibrator or put it where it 
belongs, near the rear of the set so it wouldn't get tangled up in the 
BFO signal, but it works okay.

I have ordered a good handful of 6.5536 Mcs clock crystals -- several 
days of listening at various times to that frequency failed to turn up 
any important signal there, so it gets the nod for a modified G2DAF 
front end.   I have stared at the FT-101 chassis a while, but I'm not 
encouraged.   It converted to vacuum tubes just fine (in the HBR2K) 
but that was using the existing tuning components and conversion 
scheme.   There's *no* extra space in the front end area, and 
changing from the slug-tuned preselector to a five-gang cap;  from 
dual conversion to single with premixing; and from a single ended 
VFO to push-pull has many issues.   The biggest looks to be the 
need for space for coils around the bandswitch -- the FT-101 has *no* 
space there.

I may wind up with the FT-101 VFO assembly (minus all the 
oscillator parts -- that will become a vacuum tube circuit) on a blank 
chassis.  I'll also take a look at my Tempo One junker chassis.

Walt
KJ4KV