Thanks to a file for running and testing the MC1496 by Ray Anderson WB6TPU, I was able to test out my dual mixer theory I posted here last year sometime.
I split the MC1496 Gilbert Cell into two independent demodulator / modulator sections. I have devised a way to test the circuit to prove the separation of each half. You can run and test the simulation with LTSpice. See if'in its a gonna work for ya'? http://groups.yahoo.com/group/soft_radio/files/SDR%20Spice%20Analysis%20Files/MC1496_Test.zip Note: In order to magnify the I/Q separation, the Transient time domain simulation is set for 10u ( .tran 10u). R2 is on pin 1 (S+) and R3 is on pin 4 (S-). If you adjust the value of either resistor, you can balance the I and Q level, or diminish one of them. If you remove R2 and C2 and their nodes. Or R3 and C3 and their nodes. you can completely remove either the I or the Q outputs. This way you can prove the separation of the two halves of the Gilbert Cell. What you do in one half of the circuit, will not effect the other half. R2 (Q) is set for 75 ohms to diminish the Q, and R3 is set for 1 ohm so that (I) can be full amplitude. And you will see the effect as soon as you run the Transient analysis. I worked on R7 and R8 to get the best separation waveform I could. I wanted to do this so that one chip could be used for a QSD mixer. Or for quad phase modulator or for both. And so, this will handle both the I and Q. My vision of this is that we can use perhaps a 74HC74 to split, shift and divide down a local oscillator, and run this split I /Q ~ LO IN to pins 1 and 4, and then the incoming signal to be sampled from a receiver or an antenna can be applied to pin 8. You might be able to make something like SR40 with this and even without a broadband input transformer. But I am thinking this would be a good idea in a conventional off the shelf radio as a SDR hybrid. Dannie Jackson ka9rza
