single-ended BJT MIXER design help!

[e2]

please help me with my design on metal detector.i need to mix my two
sine inputs(few hundreds of KHz)and extract their difference.i opted
for the bjt.my questions are:

1.biasing
-- using the buil-in transistor of lm389,how do i bias it such that
it operate in "mixing" mode?

2.diode-based
-- can u share any other simple circuit to implement a mixer in the
low frequency range?


thanks!!!!

[W3JDR]

There are many, many ways to implement a frequency 'mixer', and the choice
of which one to select is based on even more considerations and tradeoffs,
so it's a little diccicult to recommend a topology without knowing exactly
what the requirements are.

However, from your short description, it sounds like you're trying to make a
metal detector that works on the principle of an oscillator whose frequency
changes when the search head comes into close proximity of a metal object.
The way these usually work is that the search head has a rather large
inductive coil that is part of an oscillator tuned circuit (usually at a few
hundred KHz). The search-headd oscillator frequency is 'mixed' with a fixed
oscillator a few KHz different in frequency, producing an audible beat note.
When the search head comes near a metal object, the change in the oscillator
frequency produces an audible change in the audio beat note.

For this type of application you can use a very simple digital 'mixer' since
there is no useful information in the amplitude of either mixer input - the
only thing that matters is the frequency of the beat note. Assuming that
your two oscillator signals can be formed into clean square-wave logic
levels, I would suggest an HCMOS exclusive-or gate for this application,
followed by a simple RC lowpass filter prior to the audio ampliifier.

A little trick you might try is to connect an unused exclusive-or gate as an
inverter by connecting one input to the supply. Then connect a large value
(100K is fine) from the remaining input to that gate to its output. This
will bias the gate into a linear mode and the output DC voltage 'idle' at
approx 1/2 Vcc). If you capacitively drive this biased gate with your
filtered mixer output, it will have quite a bit of audio gain. You can use
the remaining gates as amplifiers in a similar fashion.

I'm sure that this thread will produce many other very good suggestions, But
this one is simple and pretty idiot-proof.

Joe
W3JDR


"e2" > wrote in message
oups.com...
> please help me with my design on metal detector.i need to mix my two
> sine inputs(few hundreds of KHz)and extract their difference.i opted
> for the bjt.my questions are:
>
> 1.biasing
> -- using the buil-in transistor of lm389,how do i bias it such that
> it operate in "mixing" mode?
>
> 2.diode-based
> -- can u share any other simple circuit to implement a mixer in the
> low frequency range?
>
>
> thanks!!!!
>

[e2]

thanks..you hit my application!
actually im targeting either diode or bjt based mixer becoz they're
available for me now.
i tried the single-ended bjt mixer..
i set the oscillators around 460khz...with a few hundred of millivolts
peak to peak..
i use the transistor from lm389n audio chip..what i did was to bias it
in its "active mode"
end then single-fed the two outputs from oscillator to the base of
bjt...
the problem is i cant discern the signal from the spectrum..but in the
oscilloscope,its
close to a cliped periodic wave with fundamentals close to the one of
my oscillator frequency...can you help me with another design
configuration with the bjt?

next i tried the diode based but i could hardly get a differnce
frequency..only harmonics though..my arrangement is each input is fed
to a 1k resistor which is then single fed to the negative side of the
diode...any help?or new configuration?

thanks!!