Problem of impedance matching of Gilbert cell mixer

Hi,

For a typical double balanced Gilbert cell, how to match the RF input to 100 ohm?

The zero frequency input impedance should be infinity. Even when frequency increases, the input impedance doesn't change significantly at sub-6GHz range.

I understand that inductive degeneration can help with input-matching, but sometimes we need very large value to make the real part 50 or 100 ohm, which might not be realizable.

Since the input impedance is very large, can I directly add a parallel 100 ohm resistor at the input? In my simulation, noise figure doesn't seem to sacrifice a lot. So can I do it?

Firstly, is it integrated or not? Integrated mixer's input is connected to a low noise amplifier (LNA) usually, matching between them is not necessary until the frequency is not too high and distance is low between them. Even with discrete devices this statement can be valid. Share more details, specifications what you want.

I think you had better show your input impedance to examine whether or not your process is possible of doing input matching.
Because any BJT has a rpi in parallel with Ccb hence we are possibly able to do matching with L-C fundamental structure. If inductors or capacitors are enormous challengingly to implement on-chip, you r supposed to employ Transmission line or off chip components instead.
In case of putting a 100 Ohm at base of RF input device, 100 Ohm here gets the NF significantly soar to the bad figure so i think it is not a wise choice for matching stuff at least in my design experience.
Have fun !