Yes it certainly is possible. Your problem will be matching components for making the carrier feed through initially low and stay low over temperature and time.
There will be 8 transistor for full differential Gilbert cell and " Current Transducer " to control the currents of the each pairs.
But some pitfalls may ocur . You may not provide the symmetricity and if
there become 15-20 degree phase error , at the output this will return back with poor sideband rejection and LO rejection.
Instead of this , using ready-to use IQ modulator is more efficient.
Philips SA 900 is very useful Transmit IQ Modulator.
Datasheet and App. report are attached here.
A new question:how to measure a IQ mixer's amplitude and phase error accurately?
Now I should design a 9~10GHz IQ mixer,the IF is DC to 100MHz.
As I know,The HP 8753C or HP 8753D's freq offset mode can be used to measure the phase track of IQ mixer,but this network analyses's can not be used up to 6 GHz.whoever know any other method to measure it accurately?
I want to test baseband circuit.
Modulator has baseband IQ output and Demodulator receive 10.8 MHz input.
Are there any IC match or how to test it anyway.
The only idea I get is to build IQ modulator to output at 10.8MHz
Do you have any idea?