On Wed, 18 Feb 2009 12:12:53 -0800, Richard Clark wrote:
On Wed, 18 Feb 2009 12:06:09 -0600, Frnak McKenney
wrote:

What has puzzled me is that I have run across designs that use
(e.g.) a JFET isolation amplifier hooked to a whip or hunk-o-wire
with the statement (or implication) that this is done to ",atch the
antenna's impedance".

Matching provokes heated debates that in times past ran to 600+
postings - few knew what they were arguing (but enjoyed arguing
nonetheless) and little was offered.

What I think I'm looking for would be a point of reference that
would let me, if not exactly evaluate the facets of such an
argument, at least be a foundation for forming a testable opinion of
my own.

A JFET at these frequencies does satisfy the naive requirements of
"matching," but that giving you a reception solution doesn't always
follow.

In my current state of ignorance of the subject, this sentence has
the appearance of a Zen koan: something that sounds non-sensical at
first glance, but which, after sufficient time and effort studying,
will undoubtedly become so blindingly obvious as to appear trivial.
Thank you... I think. grin!

So I;ve been trying to figure out how to
calculate/estimate what it would be, without much success.

The Mohican schematic says quite planely Hi-Z input. This is
borne out by the antenna connection feeding a tank circuit in the
front end where the input stage is fed from a low tap into a 470 Ohm
resistor. ...

Ah! You have a Mohican? Or just access to the manual? Mine is missing,
burioed somewhere in my basement; I was fortunate enough to locate a
copy of a GC-1A PDF some kind soul posted online.

... This would be your JFET feed Z, but you could choose any
suitable close value. As for the input Z, the JFET input resistance
is perfectly capable of mismatching horribly high - although this is
not about optimal power transfer at these impedance levels. What is
at risk, is the JFET input capacitance which could present a low Z
at some frequency. Naturally, you select your JFET against this to
optimize. It will be in some ratio to the antenna capacitance (if
it is bare, short wire) and that will establish the proportion of
signal that gets in by divider action.

I imagine that the inductance of a 6' extension cord (not plugged
in, just dangling from a planter hook grin!) is down in the
uH-or-less range, which would mean that most of the "tuning"
inductance would have to be supplied to achieve 60kHz. I have this
image of a big (tens of mH) inductor in series with a moderate
capacitor and my (electrically) short wire;

That is one way, other ways work too and are electrically equivalent.
Loops help tune and match by a slightly more elaborate means, but
still fairly holds to simple requirements.

You don't need wire to build an inductor. At these frequencies you
can use a capacitor in a Gyrator design.

"Gyrator"? I thought that was the rooftop dance that follows an
antenna adjustment in mid-thunderstorm. grin!

Google led me to the AAVSO site (www.aavso.org) which led me to the
Yahoo VLF_Group. Aaaaaaaaaaaaaaauuuuuuuggggh! A circuit that can
replace capacitors or inductors? My first reaction is "technology
at a level indistunguisable from black magic".

I don't think I'm in Kansas any more.

(On the other hand, I have lots more toys to play with. grin!)

all of the surrounding
EM sets the electroncs in the wire to dancing, but the series RC
blocks those which are wiggling "off-key" (e.g. not dancing at the
"proper" rate of 60kHz).

And this responds to the filtering capacity (selection AND
rejection). This is called "Q" which also serves the yeoman's
task of matching as well

Um. I don't think I ever got past the simplistic "High Q = Good,
Low Q = Evil" stage. Looking back, I can now see cases where an
excessively high Q might be... counterproductive, but as always, it
depends on what one means by "high" or "low" in a given context.

Noted as something else I need to review and not depend on instinct
for.

... (observe the input tank design for the
conventional bands).

Of the five bands (A-E, SW3 positions 5-1), the only one which seems
different is "E", with an additional 130pF cap between the antenna
and the tank circuit.

Is that what you're referring to?

--snip--
... The module's 12V power and 120V
switching connections are done through a 9-pin tube socket with
mating connector/cable; remember those? grin!

Remembering isn't difficult. I broke into electronics through
TV/Radio repair during high school. If I could fix it, I got paid.

_You_ got _paid_!?? grin! Woody Maiden, WA4GMV, ran a radio/TV
shop near my home and didn't object to someone hanging around and
watching over his shoulder in the afternoons, but he was a bit
concerned about his liability insurance. (Don't know why. I didn't
get zapped by a "discharged" CRT until my college years. grin!)

Anyway, power connections from that era brought "ground" notoriously
close to lethal if you plugged the radio into the wall wrong. Some
used AC noise reduction circuit design that could almost guarantee
your chassis was floating at 70V if things went wrong.

"We don' need no steeken' transformers!" I think I saw a 50C5 in one
of my basement boxes a few months back. grin!

I had an ET striker (Navy parlance for a student electronics tech)
who connected a TV antenna input to ground, and the insulation
melted off of the wire in a heartbeat. This was in the day when we
called TV power line interlock replacements "suicide adapters."

I learned about them from my father, a power company EE who wasn't
above swapping the odd tube or building a color TV from a kit with
my little sister's help.

73's
Richard Clark, KB7QHC

* * *

If you're getting bored, please feel free to skip the following; on
the other hand, someone with access to a GC-1 manual might enjoy my
story...

A few months back I pulled my old Mohican out from under a pile of
magazines and blew the dust off it. It had been buried so long that
I had forgotten why I never built an outdoor antenna for it or
showed it to my nephews an nieces; I remembered the moment I powered
it on: there was a loud buzz coming from the speaker.

This wasn't _hum_ mind you, which even when loud has a sort of
even-ness to it; this was an intense, jagged, and metallic "I'm
shredding your speaker cone and I don't care!" kind of
low-frequency buzz. Worse, it went _away_ when I turned the volume
_up_, and reappeared when I turned it (the volume control) down.

I was able to isolate it to the audio section by unsoldering the
capacitor connecting the volume control wiper to the base of the
first audio stage. It went away (blessed relief!), and when I hooked
up a RadioShack "utility" amplifier to the volume control wiper the
sound was clear and crisp. I considered this A Clue. grin!

NTE listed their NTE102A as a replacement for the Mohican's
2N407s... at nearly $7 each (apparently Germanium is on the
Endangered Elements List). I replaced all three AF transistors, and
the audio level was much better than it has been. Unfortunately,
the BUZZZZZ was still present, and also louder. I did notice that
turning the ANL ON (diode and resistor connected between the base of
X7 and ground) eliminated it.

I tried all sorts of things. I jumpered the volume control wiper to
ground to make sure the resistance arc hadn't cracked, leaving the
wiper floating near its low end. I tried swapping the transistors
around. I tried paralleling the electrolytics I could reach, but no
luck.

As you've seen (or know personally) the Mohican's IF and AF stages
are all on one printed circuit board which is bolted down to the
chaissis with ?2-40 machine screws around its border. I couldn't
see much of the underside of the AF section of this board because my
vision -- and fingers -- were blocked by a large AF transformer, so
I unbolted it and let it hang free (but not shorting anything; thank
&deity for stiff transformer leads).

With that pushed out of the way I could see all three AF transistor
sockets and even reach them with my test leads. I jumpered a 100uF
electrolytic from the chassis to the emitter pin of X7, the AF stage
1 transistor, parallelling it with the existing 150uF emitter bypass
capacitor (C55(?)), and a miracle occurred: the buzz disappeared!
I can't tell you how good it felt to hear normal-sounding hiss,
static, and the odd "sqrgrl" coming out of the Mohican's speaker.

No 150uF in the parts bin, so I wired two caps in parallel for a
147uF, shrink-wrapped them, unsoldered the decades-old 150uF and
wired the glob-cap in its place. Didn't even burn my fingers in the
confined spacegrin!. And then I turned it on.

BZZZZZZZZZZZZZZ!!

I'm sure you recall the feeling: "What did I do wrong?" Bad solder
joint? No... Overheated the component? Not as far as I could
tell. Dropped a solder glob across two traces frying by $7
transistors? No... in fact, although the buzz existed over a
larger percentage of the volume control's span, the louder end of
the audio worked just as it had before. My "proven cure" had made
the problem slightly worse.

It was late, I was tired, so I decided that, since my previous
jumpering had created 250uF total, I needed to hit the problem with
a bigger hammer: I jumpered in a _470uF_ this time. And the buzz
went away. Again.

Okay. Now I was _sure_ I had fixed the problem, so I unsoldered my
obviously-too-clever glob-cap and soldered in the 470uF, this time
triple-checking the polarity markings. I inspected the traces for
possible solder bridges. I inspected the new solder joints under a
large magnifier and wiggled the new leads; the joints looked good,
so I hooked up the power supply again and turned it on.

BZZZZZZZZZZZZZZ!!

By now I was tired _and_ cranky, so I jumpered in the original
decades-old 150uF cap. No buzz.

At that point I had two options: go down in the basement and bring
up a sledge hammer, or quit to go eat supper. It was a close call,
but supper won out. There's something very weird about watching the
news, eating supper, and looking over a schematic and trying to work
what I was doing differently when it worked.

If the emitter of X7 is wired to R37 and C55, and if the other ends
of R37/C55 are wired to ground, how on Earth could a new capacitor
jumpered from X7's emitter to the chassis cause different behavior
from the _same_ capacitor soldered to a trace going to the same
emitter and the PC board's ground trace?

All I could think was "this makes no sense, but it does appear to be
happening". R37/C55 were clearly grounded. The trace was unbroken.
If they weren't grounded, I'd be getting _no_ audio. So how could
_soldering_ to ground be any different from _jumpering_ to ground?
Even stranger, how could a jumpered connection be _better_ than a
soldered connection?

* * *

You've probably figured it out by now. The answer is that when I
was jumpering, it was to the metal chassis; when I was soldering, it
was to the PC board ground trace. The "ground" trace on this PC
board includes a wide band around its outer edge; the strip makes
contact with the chassis when the board is bolted down, creating a
really solid round-the-board ground. Somehow, over the past
half-century, the contact between the chassis and trace had become
slightly less than perfectly conductive, and was "floating" in the
Never-Never Land between zero and infinite resistance.

Loosen all ten machine screws slightly, spray some contact cleaner
in the extremely narrow gap, tighten the screws, and guess what? NO
BUZZ!

Re-mount the transformer(*). Reconnect C54 to the volume control
wiper. Check for accidental shorts, dropped bits of solder and
component leads, and make sure the transistors haven't fallen out of
their sockets while I had the chassis upside down, Turn it on.
STILL NO BUZZ!

By then it was 2230, so I went to bed. Next step will be to improve
the PC board/chassis contact by running a thin bit of fine steel
wool around a bit to get rid of any residual crud; with luck, it'll
be another couple of decades before this problem reappears.

As for the 470uF capacitor, I think I'll leave it in place. The
original is likely good, but 470uF should work as well, and all I
need now is to have that portion of the trace lift off from the PC
board due to overheating.

* * *

What's interesting here is the contrast between the schematic and
reality. On paper, or on my CRT, those components were clearly
connected to ground, and, had I hand-drawn my jumper wires, their
wiring lines would have looked identical. It took me a long time to
find the cause of the buzz because I was stuck thinking about the
lines on the schematic, even though my hands were working on
physical components, wires, and traces; how do you know when to stop
trusting your compass and check it against the sun and stars?

Definitely a "Learning Experience". grin!


Anyway, thanks again for your comments. My head is about to explode
from all the new ideas, but please don't take that amiss. If I
don't expand it occasionally, it starts collapsing into a dull,
super-dense mass (think neutron star grin!).


(*) GlueStic(tm) is great for assembling hardware in confined
quarters. It's much better than spit for making lockwashers and
nuts stick to your fingertips so they don't drop off into the
chassis interior.


Frank
--
"If language is not correct, then what is said is not what is
meant; if what is said is not what is meant, then what ought to
be done remains undone." -- Confucius
--
Frank McKenney, McKenney Associates
Richmond, Virginia / (804) 320-4887
Munged E-mail: frank uscore mckenney ayut mined spring dawt cahm (y'all)