Hi guys. I'm planning to do a video tutorial about CMOS Triode region mixers (also sometimes called Potentiometric mixers). I am not that familiar with what HAMs like to use, but I don't think I've come across this topology in any HAM stuff before. I have however seen it more than once in integrated circuits.
The circuit can be seen in the figure here. The concept is simple - the FETs are biased in their triode region. As such, they appear as resistors. The LO modulates their resistance, and as the gain of the amplifier with feedback depends on the ratio between the feedback resistor and the input resistors, the gain of the circuit depends on the LO. Since Vout = A*Vin, you are multiplying the LO with the RF input, and hence mixing.
The great thing about this circuit is that it can operate a good bit above the ft of the process it is built in, at least when we are building a downcoverter. In order to do so, we replace the X component in the circuit with a capacitor, which will "short" the two inputs at RF, and hence the amplifier doesn't need to provide the virtual ground for RF frequencies anymore. The amplifier onlyneeds the bandwidth required for the IF. (this has for example been done to build mixers with a 200+ GHz RF/LO in a 170 GHz ft CMOS process).
For upconversion, the amplifier needs to be able to have sufficient bandwidth to drive the feedback network. However, we can "help" it by tuning the capacitance at the input out with inductors (ie, replace X with the right value inductor)
For the tutorial video I'm building a downconverter based on this architecture (though just a single ended one, using a BS270 as triode). It works to about 35-40 MHz (I don't have enough RF sources to test it to any higher frequency).
Most of the times I see people discussing mixers in HAM use it tends to either be diode mixers or Gilbert cells. Has anyone else used this topology before in radio work? Anything interesting to add about this that is (or isn't) potentially worth mentioning? Any reason why you don't use it?