Our receiver chain is as follows:
IQ Mixer: HMC1113LP5E
The problem is as follows: The I and Q signals out of the HMC1113 are not in quadrature. In fact placing a shield over the receiver changed the phase difference.
The HMC903 devices are supposed to operate at 3.5V, we decreased the voltages to 1.5V (for both the devices) and the I,Q signals were in quadrature.
Question 1) : How is the bias voltage of the amplifiers affecting the IQ balance?
Setting the bias voltages to 1.5V seemed to solve the problem at least when the board/pcb was tested outside. However when the PCB was put inside an enclosure the phase imbalance again changes infact to the extent that one of the I and Q signal completely disappears (or gets attenuated by 30-40dB and is no longer visible on the oscilloscope but can be seen on a spectrum analyzer).
Q 2) We imagined a problem of cavity resonance and placed absorbers on the cover of the box... this changes the situation (phase imbalance and amplitude level) but does not solve it.
Is this problem somehow connected to the first fix of making the amplifiers work at 1.5V instead of the recommended 3.5V?
Making the devices operate at 1.5V doesn't seem to affect the gain or the increase the noise floor.
Thank you for your time in answering these questions,