RadioBanter - View Single Post - how to wind 10mH inductor for AF diplexer?

March 10th 06, 01:43 AM posted to rec.radio.amateur.homebrew

wrote:
wrote:

I understood that it is a good idea (or even mandatory?) to follow DBM
with a diplexer.

I know that in terms of the audio frequencies, op amps with R and C
passives have long been a viable option to big inductors - but can they
now satisfy the need to terminate the RF components?

I've been playing with this a bit in simulation, and came up with
something seems to work. Most likely it is either common knowledge, or
fatally flawed - would be interested to hear which.

What I started with is an op amp with a grounded non-inverting input, a
50 ohm resistor from the mixer output to the inverting input, and a 50
ohm feedback resistor from the op amp output to the inverting input.
With an ideal op-amp, the inverting input would match the non-inverting
input's ground with a virtual ground, and the mixer would be
resistively terminted into this at 50 ohms for all frequencies.

Two obvious failure modes come to mind. First, at some frequency the
op-amp feedback will no longer be able to maintain a satisfactory
virtual ground. I tried to address this with a capacitor from the
inverting input to ground. 220 nF seems to clamp it there below the RF
frequencies where a 1 Mhz gain-bandwidth op amp fails, but not roll off
the audio response too badly.

Second, in the presence of a strong intefering signal on the order of
volts, the op-amp probably can't drive enough current through the 50
ohm feedback resistor to reproduce this, so non-linear effects (and
intermodulation) may set in. I attempt to address this by shunting the
feedback resistor with a 470nF capacitor, to make the gain roll off
towards zero above audio.

There is some slight reflection at the transition frequencies, but it
looks like it can be kept very low. One major weakness of this circuit
is that the rolloffs are largely first order and thus gradual. Given
that the goal of the exercise was to avoid the use of large inductors,
one improvement might be to precede the mixer with some higher order
selectivity at RF or IF, where inductors, or even ceramic resonators
are easy to come by. Another obvious improvement would be to pad the
output of the mixer down by quite a bit, though that probably means
picking up some broadband noise from the op-amp when it is reamplified?

So far I've only played with this in simulation. Trying to think how I
would test it if I were to actually build it. I have some access to a
network analyzer and would guess I could use that to measure the
reflection from the input that the mixer would see over a wide
frequency range. Measuring the ability to handle a weak signal in the
presence of a strong interfering one might be harder.