Mixers and Time Sidelobes
This forum is always filled w/ people who know wayyyy more than I do, so I am here once again to ask a general question. Its a good one.
If I put a filter infront of a mixer then another filter after the mixer, will the phase sidelobes of the impulse / step response add together? Will the phase sidelobes of the first filter pass through the mixer?
Good luck and thanks for the help.
Kyle
I am not sure if I understand your question. Phase relations concern two separate transmission lines; one can determine the phase between the signals. One line is usually taken as a reference.
If you use a mixer, its time delay is usually given by its physical length only. Phase and group delay is only defined by selective components like filters.
To measure e.g. the time, phase or group delay of a frequency converter (mixer with RF and IF filters), you need another "reference" mixer without those filters. The reference mixer and the tested converter mixer are fed the local-oscillator signals from one common source.
Sorry,
My question has to do with the impulse response of a mixer.
ie: the impulse response of a simple filter takes on the shape of sin(x)/x. If I have a filter at the LO port, filter at the IF port and filter at the RF port what will happen?
Will the IF, LO, and RF filter responses combine?
Will just the IF and RF filters combine?
Or will the output RF filter dominate?
Is it safe to assume that if the LO is not the signal channel, I can disregard the impulse response of the LO filter?
The local oscillator in principle switches the diodes alternately on and off and if there is a filter, it only serves to suppress unwanted signals from getting to the mixer. It never affects mixer parameters if the LO power is nominal.
The pulse response is always determined by its spectrum, and then by the narrowest filter in the RF to IF chain through the mixer. Mixer alone does not affect it but if the bandwidth of the narrowest filter (usually the IF bandpass) is comparable with the impulse spectrum, then any spectrum limitation by that filter results in smoothing the impulse response.
I am used to handle mm-wave mixers, which typically have the IF filter ~10 GHz wide. This means that such filter can affect impulse response if it is shorter than 0.1 nanosecond. You can scale these numbers for longer pulse rise time and/or narrower filter. A 10-GHz wide filter can easily pass data rates > 10 Gb/s without distortion.
If you operate the mixer in a linear-response condition ( no saturation), then mixer will not introduce any distortion to pulsed signals. In balanced (one diode pair) mixers, the saturation typically occurs when the input RF signal power is larger than the LO input power -7 dB. Mixers with a better linearity require a higher LO pump power .