The simulation of the Gilbert mixer gave results closer to the integral than the product of the signals.
A low-pass filter using a resistor + capacitor is used at the Gilbert cell load stage.
I wondered if the cap at the load stage performed integration, so I removed it and simulated it, but there was no big difference.
I would like to know what could be the cause of the mixer acting like an integrator.
The test circuit I built differs from the standard Gilbert Cell Mixer, we've added resistor biasing and capacitors to one side. it works but I wonder the purpose of these serve based on how we're inputting the RF and LO signal
-What is your LO and RF frequency ??
-Do you drive the commutation part with rail-to-rail square wave ?
-Have you ever checked the Bias voltages and currents ?
-What are for 50 Ohm resistor in parallel to generators ?
Show the waveform at active nodes from transisent analysis.