Message	Date	From	Subject
1470	2008-02-25 20:18:07	kanewderfish	Zout of a 2N4401 tuned RF amp
Hi gang, I've been doing some experimenting with measuring some impedance/bandwidth parameters of parallel tuned tank circuits. I have been using the methods similar to those outlined in "Solid State Design for the Radio Amateur" and "Experimental Methods in RF design. Specifically, connecting a 1pF cap from a 50 Ohm generator to the high side of the tank (other side of the tank to ground), and another equal value cap from the same point to a 50 Ohm resistor connected to ground. A 10X scope probe and 100MHz scope is then used to measure the voltage across the 50 Ohm. Work is being done around 10MHz. I appear to be getting reasonable results. What I really wanted to investigate however, was the effect that a common emitter 2N4401 BJT amp would have on a given tank circuit. So, I did the following: I built up dead bug style a voltage divider bias with emitter stabilization amplifier circuit with the following values: Resistor from decoupled Vcc (9V) to base--150K Resistor from base to ground-- 16K Resistor in emitter lead--100 Ohms (with a 0.01uF bypass cap) Collector connects to one side of parallel resonant tank, other side of tank has 100 Ohm to Vcc and a 0.01uF cap to ground (decoupling network). Emitter current is ~1mA There is no feedback network from collector to base.(Other than internal effects and stray). I tried connecting the fixture described above (1pf coupling caps, and 50 Ohm load) to the collector/tank junction to compare the bandwidth of a given tank circuit alone or connected into this powered amp circuit. This was problematic as I saw multiple resonant points that did not seem to correlate with the values obtained with the tank alone. I presume this is due to the transistor circuit with it's associated internal feedback paths upsetting the readings somehow, but I'm not certain. I then connected a 50 Ohm signal generator to the base through a 2.7pF cap, and coupled the collector to a 50 Ohm resistor through another 2.7pF and powered up the circuit. This seemed to produce much more stable results. Many of my results were as expected.....the tank circuits are heavily loaded by what I presume to be the output impedance of the transistor. Here is a sample result: A coil wound on a T50-2 to produce 2.717uH resonated with 99pF when tested alone in the fixture described above, produced these results: Fr......9620kHz 3dB bandwidth....56kHz Q.....171.8 X....167.5 Ohms As Z = Q x X Z = 28.76k Ohms When I placed the same tank in the amplifier circuit as described above, I got these numbers: Fr.....9694kHz (A bit higher freq.......??) 3dB bandwidth.....1640kHz Q.....5.91 X.....167.5 Ohms As Z = Q x X Z = 994 Ohms Since the raw tank Z was ~29k Ohms, and the value for Z was ~1k Ohms for the transistor circuit, I assume the impedance at the collector must be ~1k Ohms to cause that much of a drop. Any thoughts? Does ~1000 Ohms seem plausible as the total Z seen at the collector of this stage? Is my technique reasonable? "Why did the 2 caps to the collector fixture" not work well? BTW, I have more data with other tank values that seem to confirm, (at least >my< understanding) of things like L/C ratios, loaded Q, and bandwidth. If there is any interest I'll put it together and post it. Tnx Bob WB0POQ

Hi gang,
I've been doing some experimenting with measuring some
impedance/bandwidth parameters of parallel tuned tank circuits. I
have been using the methods similar to those outlined in "Solid
State Design for the Radio Amateur" and "Experimental Methods in RF
design. Specifically, connecting a 1pF cap from a 50 Ohm generator
to the high side of the tank (other side of the tank to ground), and
another equal value cap from the same point to a 50 Ohm resistor
connected to ground. A 10X scope probe and 100MHz scope is then used
to measure the voltage across the 50 Ohm. Work is being done around
10MHz.
I appear to be getting reasonable results. What I really wanted to
investigate however, was the effect that a common emitter 2N4401 BJT
amp would have on a given tank circuit. So, I did the following:

I built up dead bug style a voltage divider bias with emitter
stabilization amplifier circuit with the following values:
Resistor from decoupled Vcc (9V) to base--150K
Resistor from base to ground-- 16K
Resistor in emitter lead--100 Ohms (with a 0.01uF bypass cap)
Collector connects to one side of parallel resonant tank, other side
of tank has 100 Ohm to Vcc and a 0.01uF cap to ground (decoupling
network).
Emitter current is ~1mA
There is no feedback network from collector to base.(Other than
internal effects and stray).

I tried connecting the fixture described above (1pf coupling caps,
and 50 Ohm load) to the collector/tank junction to compare the
bandwidth of a given tank circuit alone or connected into this
powered amp circuit. This was problematic as I saw multiple resonant
points that did not seem to correlate with the values obtained with
the tank alone. I presume this is due to the transistor circuit with
it's associated internal feedback paths upsetting the readings
somehow, but I'm not certain.

I then connected a 50 Ohm signal generator to the base through a
2.7pF cap, and coupled the collector to a 50 Ohm resistor through
another 2.7pF and powered up the circuit. This seemed to produce
much more stable results. Many of my results were as
expected.....the tank circuits are heavily loaded by what I presume
to be the output impedance of the transistor.

Here is a sample result:
A coil wound on a T50-2 to produce 2.717uH resonated with 99pF when
tested alone in the fixture described above, produced these results:
Fr......9620kHz
3dB bandwidth....56kHz
Q.....171.8
X....167.5 Ohms
As Z = Q x X Z = 28.76k Ohms

When I placed the same tank in the amplifier circuit as described
above, I got these numbers:
Fr.....9694kHz (A bit higher freq.......??)
3dB bandwidth.....1640kHz
Q.....5.91
X.....167.5 Ohms
As Z = Q x X Z = 994 Ohms

Since the raw tank Z was ~29k Ohms, and the value for Z was ~1k Ohms
for the transistor circuit, I assume the impedance at the collector
must be ~1k Ohms to cause that much of a drop.

Any thoughts? Does ~1000 Ohms seem plausible as the total Z seen at
the collector of this stage? Is my technique reasonable? "Why did
the 2 caps to the collector fixture" not work well?

BTW, I have more data with other tank values that seem to confirm,
(at least >my< understanding) of things like L/C ratios, loaded Q,
and bandwidth. If there is any interest I'll put it together and
post it.

Tnx

Bob WB0POQ