--- In qrsstxrx@yahoogroups.com, "Hans Summers" <hans.g0upl@...> wrote:

> Nice work, and very interesting how you have arranged the transformers
> in the H-mode mixer.
>
> Interesting and worrying! I don't completely understand that
> transformer... it worries me because you are using the same core for
> your input transformer and your output transformer, is that Ok? Also
> is it Ok to be using that transformer at AF? I know that previously
> you were considering just summing the two complementary baseband
> outputs of the H-mode mixer using a differential amplifier. I wonder
> why you changed now to using the transformer? What would be the
> relative advantages and disadvantages of each?
>

Hi Hans - It is difficult to reply to this without going into a
complete explanation of how the h-mode mixer works. Basically, in
this system there is no output transformer: I use the op-amp with its
100 Ohm input resistors instead.

If you compare my circuit with the classic design, you will see that I
have reversed the rf and if ports, and then replaced the output
transformer primaries (previously used for rf input) by 2 100ohm
resistors. Finally I have combined the 2 transformers previously used
as if outputs into 1, and used it as an rf input. The transformer is
an rf design, using a binocular core, and has no af function. I
placed an explanation of this configuration (but using 2 input
transformers) on the files section about a month ago. It is worth
refering to that:- fstmixerorg.pdf

Really there is no alternative to some serious thinking about how this
mixer works. It would resolve your doubts about the placing of 20.28
MHz filter capacitors as well (to remove the f*2 mixer output, and to
complete the 100ohm termination path to ground). Notice I have placed
them after the 100 ohm filter-terminating resistors, as previously
discussed.

It might be useful for you to try to work out an af-output mixer
circuit for yourself; you would then see the logic of what I am doing.
You might come up with an interesting alternative.

> It would be very interesting to measure. Do you have the right skills
> and equipment to measure IP3? I know I don't. That's another one for
> the G0UPL ToDo list.

I am still evaluating this design, and will give some thought to
trying to measure the IP3. Paolo published a spectrum lab picture of
a mixer output spectrum, and that might be a way forward, but it is a
complex measurement even given the right equipment. Basically you
have to measure how the 3rd order mixer output varies with input
amplitude, plot a graph, extrapolate the line, and see where the it
intersects the first order output line. One should be direct
proportion, the other a cubic. The data sheet for the SA602 gives an
example of the kind of plot that is needed.