I just returned from a little radio-related hiatus. GOOD News: A solid
weekend long DX contest is always a blast at a good station. BAD New: That's
if conditions are any good - bad conditions, like last weekend in the
Miodwest, means lots of mindless hours of listening for weak signals in much
stronger noise. I recall getting a few points from some calls that I
recognize from this forum - thanks.
I have some unfinished business here, regarding g-g amplfier input circuits.
When I attempt to read the postings I am somewhat frustrated by a certain
vagueness to the language. I do not believe in or participate in arguments
which are sematic only - if there is a REAL issue, the language problems can
always be eliminated. So in the present case I will simply post my own
comments, some of which are facts and some of which are opinions. I will be
very careful to distinguish bwteen the two, and to make the intended meaning
of my statements as precise as possible.
First, there are no resonances in the grid structure of a 3-500Z at
anywhere near 90MHZ. This is a FACT which I obtained by measuring two
different tubes, using three different measurment techniques. I would like to
dwell for a moment on this technique(s), since it is of prime importance here.
When you look at the base of the tube, it is clear that the shortest
possible lead from grid-to-cathode (i.e. the "grounded grid" lead itself)
will be about 1" long. This immediately implies a lower limit on the amount
of external circuit inductance to the range of 25nH. or so. In other words,
any conceivable circuit will have at least that amount of L. in series.That
situation also applies to the measuring fixtures which I used. I was able to
fabricate interconnect leads to my instruments which were in the 1" range
(total length.
The three different instruments I used were: 1. An HP4815A Vector
Impedance Meter, 2. An HP 8407A/8601A Network analyzer, and 3. An HP 8920A
Communocations analyzer.
On the direct Z measurment (4815) the Grid-Cathode circuit, as a
terminal pair with a driving point impedance, read -j120 (negligible R) at
100MHZ. This translates to about 10pF opf capacitance. I was able to sweep F
from 40-125MHZ (by means of some of my own proprietary enhancements to the
4815). As I did that sweep manually, there were virtually NO signs of any
resonant behavior. The thing just acted like a perfect capacitor!. The slope
of the dX/dF curve, which gives some insight into the inductance of the
circuit, also was consistent with an extremely low value of Linternal.
To cross check with the NA's, I simply fed the terminal-pair thru a 50
ohm source resistor and looked at the output amplitude/phase response. Since
the impedance happens to be not very from 50 ohms (although it is fully
reactive) this is an acceptable way of doing quick and approximate searches
for resonances. For example, connecting a short (5-20') piece of coax in
place of the tube shows the very market resonances at 1//4-lambda lengths.
Well, once again, no hint of resonance in the tube. With the 8920A, in the
Analyzer/Tracking Generator mode, I was able to sweep up to 1GHZ.
Conclusions: 1. The tube does not have any resonances visible up to
over 200 MHZ. I don't really know what happens at that frequency and beyond,
except that the conditions for a lumped-circuit analysis are probably not
valid any more. Trying to establish a meaningful terminal-pair to which an
impedance can be assigned becomes progressively meaningless.
2. The tube does act every bit like a 10pF capacitor. When you use it
in a GG amplifier, your "grounding" lead inductance had better be less than
100nH or so. This is not too hard to achieve. Recall that the entire L
required for a 6Meter PInet tank circuit may be 300nH .
3. If you are not careful to do this you almost certainly will have an
unstable amplifier.
Eric von Valtier K8LV
P.S. I also tried 811's, SV572's and an 833A. The latter has a minimum G-K
terminal path of almost a foot, and would resonate exactly as per simple LC
theory.
|