EMRFD Message Archive 4702

Message Date From Subject
4702 2010-05-18 13:08:48 w1kilofoxtrot Mixer terminating diplexer tuning
I'm playing with a simple superhet receiver, to reinforce my understanding of how they really work. (I'm using W8DIZ kits, and retuning for my purposes.) Yes, I could go ugly, but it's a challenge for me to make adjustments to designs. The tinker-toy approach appeals to me.

Anyway, I'm re-designing the diplexer on the output of the first mixer. I know how to calculate the values of the components, based upon the Q (and bandwidth) I'm trying to get. I know how to test it, using my VNA. But I'm a little stumped on the practical method for tuning the series and parallel resonant circuits.

When I simulate the circuit (using RFSim 99), I see small changes in component values affect the response dramatically. I plan on using my LCR meter to preset the inductors to design values, but there is likely to be some variation when they are in the board. I could tweak endlessly, and never quite get a good result.

Also, is there a similar procedure for the bandpass filters on the front end of an RX? That does seem a little more straightforward...

Any helpful hints? Or, am I just overthinking this?

73,
Steve
W1KF
4703 2010-05-19 10:54:00 ha5rxz Re: Mixer terminating diplexer tuning
"When I simulate the circuit (using RFSim 99), I see small changes in component values affect the response dramatically. I plan on using my LCR meter to preset the inductors to design values, but there is likely to be some variation when they are in the board. I could tweak endlessly, and never quite get a good result.

Exactly. Using some software I designed a 30 MHz nine element low-pass filter that also provided a bandstop function at 45 MHz. Some of the component values were none-standard and exact, when I departed from these values the filter response was poor. So.....

1) Using my AADE LC meter I adjusted all the inductors to the required values and marked them. I then assembled sets of capacitors to the required values and built the circuit.

2) It was there, but not quite as there was a dip in the response between 20 MHz and 28 MHz. Leaving the capacitors alone I then adjusted the coils by tiny amounts, working from the input to the output.

3) If a minor adjustment produced a positive result then I left it and moved on to the next coil. If there was a negative result such as an extra peak then I adjusted the coil back to the original position and moved on.

I ended up removing all the coils and resetting them back to the original values but I only had to do this once as my second attempt produced the filter I needed. As your diplexer will only have one or two inductors you should have an easier task.

HA5RXZ
4708 2010-05-23 14:31:13 ajparent1 Re: Mixer terminating diplexer tuning
Oddly enough I undertook the same exercise and found it
4710 2010-05-24 11:12:10 w1kilofoxtrot Re: Mixer terminating diplexer tuning
Allison,

I'm OK with SWAG :-)That's why I was asking for a process, not a solution!

I'm using what appears to be a fairly common topology, where the output of the mixer goes to a series connected LC, often with higher L; in parallel with a T of two 51 ohm resistors and a parallel LC (with higher C to match) to ground at the junction of the two resistors. It would be fig 3.88 in EMRFD if it was labeled... L is fixed (toroids) C is variable.

I built it up in part yesterday. I left the resistors out, so I could sweep the LC separately. The series was easy to set at exactly 9 MHz (my target IF). I'm embarrassed to note that miscalculated the parallel LC, and don't have the 680 pF cap I need. I tried a 750, but was out of range to tune to 9 MHz. Final results to follow...

So, my plan was to try to tune them individually, then add in the resistors, and tune them as intended, with tiny tweaks until I could see how they interact. Not exactly rigorous, but I think it will work.

The band pass filters were simple, and I did much the same. I tuned the end elements, then connected the coupling cap (double tuned BP filter) and swept the pair. Not so hard to tune, but then my simulations weren't so sensitive to component variation either.

I am achieving my goal of learning practically, at least!

73,
Steve
W1KF

4713 2010-05-25 08:30:58 ajparent1 Re: Mixer terminating diplexer tuning
4715 2010-05-25 14:14:05 ajparent1 Re: Mixer terminating diplexer tuning