
This two stage buffer/amplifier
configuration appears in many designs, usually with
fixed gain, ie resistor 'Rf' and it's associated dc
blocking capacitor not fitted. The voltage gain when
driven by a high impedance source is determined by
the shunt resistance of the 620 ohm resistor in
parallel with the BFS17 base to emitter equivalent
signal impedance which together form the lower half
of a series pot down network, where the upper part
is that of the 2400 ohm resistor.
Driving the circuit from a 50 ohm source will lower
further the equivalent impedance of the lower part
of the pot down circuit, so the voltage gain will
increase as a result.
With current projects in mind, I thought it would be
interesting and useful to plot the frequency
response of this circuit with source and load
impedance's set at 50 ohms. The first graphs are
taken using a trusty Marconi TF2370 analyser and
tracking generator, so have an upper frequency limit
of 100 MHz whenever the lower limit is set to the
origin. Initially, resistor 'Rf' is not fitted.
Further graphs will be added to show the reduced
gain but increased bandwidth of fitting Rf .

Rf not fitted
Rf = 407 ohms
Rf = 105 ohms
Interestingly, the 3dB gain frequency for the 10 and
20dB gain amplifiers were similar, ie, 291 Hz for the 10dB
gain amplifier and 265 MHz for the 20dB gain amplifier. For
the projects currently in mind, the highest frequency used
will be 230.4 MHz and in it's case a 10 dB gain is required.
Generally though, the working frequency will be less than
100 MHz, so the TF2370 can be used for all the remaining
tests.
For buffer use, the amplifiers
reverse isolation figure is worth knowing. Since Zin
and Zout are both 50 ohms,
this is the same as S21.
Using this buffer in the current GB3CAM 10 GHz
baacon would (should) have been beneficial. At the
time of writing, a modified 14.4 MHz tcxo is used
and this directly feeds the multipliers. Although
the modified 'tcxo' was housed in an ovened
enclosure, non of the multipliers were, so they are
subject to varying ambient temperature. This
inevitably will show up as a slight loading
variation with temperature on the ovened oscillator
 a bit of a silly oversight really. So the next
beacon (and/or CAM's eventual replacement) will
include one of these buffers within the ovened
enclosure.


