EMRFD Message Archive 3727

Message Date From Subject
3727 2009-11-03 11:27:23 Pat Villani Digital mixers and impedance matching
Now, let me preface this with a disclaimer. I started out as an RF
engineer over 30 years ago but have not done any real RF design in a
long time. That's one of the reasons why I'm here. This book has
become a handy reference, and I respect the opinions of the folks in
this group.

I'm looking at various SDR designs, and one thing I'm finding is what
appears to be a total lack of concern for impedance matching the
mixers in these beasts. I may be totally wrong, and will go back into
my corner if told to do so.

In looking at some of the designs, predicting the impedance of the
mixer seems difficult, at best. In most cases, the RF input to the
mixer is a 1:4 center tapped transformer, followed by analog switches
that then conduct at different times during the period of the LO,
typically presenting a load to only one half of the secondary at a
time. It looks to me in some designs that the load on the switches is
there for only about 90 degrees of the LO cycle, so the impedance is
time varying. Even if there was a path to the IF output for the full
conduction of any pair of switches, it seems that none of the designs
I've looked at so far present a 200 ohm load to the secondary to be
transformed down to 50 ohms. If this is so, then wouldn't the mixer
wreak havoc on any filters that may precede it? As a good RF
designer, shouldn't I be concerned about this?

Sorry, but I really am confused. Thanks.

Pat
WB2GBF
3728 2009-11-03 11:49:13 Chris Trask Re: Digital mixers and impedance matching
>
> In looking at some of the designs, predicting the impedance of the
> mixer seems difficult, at best. In most cases, the RF input to the
> mixer is a 1:4 center tapped transformer, followed by analog switches
> that then conduct at different times during the period of the LO,
> typically presenting a load to only one half of the secondary at a
> time. It looks to me in some designs that the load on the switches is
> there for only about 90 degrees of the LO cycle, so the impedance is
> time varying. Even if there was a path to the IF output for the full
> conduction of any pair of switches, it seems that none of the designs
> I've looked at so far present a 200 ohm load to the secondary to be
> transformed down to 50 ohms. If this is so, then wouldn't the mixer
> wreak havoc on any filters that may precede it? As a good RF
> designer, shouldn't I be concerned about this?
>

If you look at the two 1:2CT transformers as being hybrid transformers,
you'll find that if the RF, LO, anf IF terminals are terminated in 50 ohms,
then everything works out fine. Each diode is on for half (180 degrees) of
the LO cycle.

One thing that might help diode ring mixers would be to terminate the
centre tap of the one transformer with 50 ohms instead of grounding it.
That provides isolation between the two sides. Without it, there is no
isolation and I sometimes wonder if there would be any IMD improvement by
adding this part. I've yet to see a paper where anyone has investigated
this aspect.

Chris

,----------------------. High Performance Mixers and
/ What's all this \ Amplifiers for RF Communications
/ extinct stuff, anyhow? /
\ _______,--------------' Chris Trask / N7ZWY
_
3732 2009-11-04 11:43:24 dixonglennb Re: Digital mixers and impedance matching
3733 2009-11-05 03:02:23 jr_dakota Re: Digital mixers and impedance matching
3734 2009-11-05 06:20:05 Robert Cerreto Re: Digital mixers and impedance matching
Correct me if I am wrong. I am still learning too.
 
What is meant by a resonant antenna? Are you inferring that a resonant antenna's feed point is 50 ohms? I was taught that a resonant antenna was one with no reactive component i.e. 300 -/+ j0 or 50 -/+ j0 etc. at the feed point. So, a resonant antenna may have many different impedances as long as there is no reactive component.
 
 Would it be more correct to say "a resonant antenna with a 50 ohm feed point"?
 
Bob WA!FXT
 

---
3736 2009-11-05 11:03:42 Lou Burke Re: Digital mixers and impedance matching
You are correct....resonant antenna simply means the absence of reactance, regardless of the actual impedance.....it can be any impedance. Your last statement is true if you are describing a resonant antenna with a 50 ohm feedpoint impedance.

Lou, W7JI

----- Original Message -----
3741 2009-11-05 13:26:03 Glen Re: Digital mixers and impedance matching
3744 2009-11-06 03:27:28 jr_dakota Re: Digital mixers and impedance matching
3746 2009-11-06 20:35:18 jr_dakota Re: Digital mixers and impedance matching
3747 2009-11-07 08:37:46 victorkoren Re: Digital mixers and impedance matching
In order to get good quadrature performance with a divide by two divider you have to feed it with perfectly 50% duty cycle double frequency signal. The divide by two uses the positive and negative transitions of the input signal to generate the quadrature signal.
The divide by 4 quadrature generator uses only the positive (or negative) transitions so they are perfectly spaced, giving very high accuracy quadrature signal.
You could get better performance when using a divide by 2 quadrature generator by filtering the double frequency input signal with a very steep response low pass filter which will make sure that the input signal to the divider has precise 50% duty cycle.
Victor - 4Z4ME