System noise figure with IQ demodulator
I'm trying to measure a receiver' NF I designed. This is an heterodyne receiver and I'm measuring the I channel output and therefore I'm not clear at all with what is the loss (gain and NF) introduced by not considering the Q channel. Does someone have a technique for considering both channels in the measurement?
Thanks in advance
Ghost
what do you mean ? explain better your doubts.
I mean not all the signal power goes into the I channel and therefore when measuring the NF between the input of the receiver and the IQ demodulator I output, the performance is rpbably degraded? Isn't it?
Hi. When you performed your receiver cascaded chain analysis (with IQ demodultor), you probably specify a conversion gain/loss for the IQ demodulator. This conversion gain/loss of the IQ demodulator includes the power splitter's splitting&insertion loss.
Hence, when you are doing Receiver NF measurement, don't worry that NF measurement not accurate because of the Q channel. What you should do is to terminate the Q output when you perform measurement on the I channel. Be sure to supply all the LO powers to the receiver.
Cheers!
Which type of NF do you want to measure? SSB? DSB?
If you don't consider Q-channel, your demodulator is not a single sideband mixer, so that the measured NF will be 3dB higher than DSB NF.
BR.
I'm not clear with what I'm measuring, it's my first IQ demodulator. I don't think that I can measure the I and Q channels at that point since some further processing is required in the digital domain to cancel the image. Therefore I am only measuring one of the channels and I would like what degradation I can expect or if another measurement is more appropriate (put the Q channel on the I channel for the measurement)?
As you say, to measure DSB NF, further processing using a VSA(vector signal analyzer) in the complex domain is required.
I recommend the following procedure to make a job easier.
1. measure the NF of IQ demodulator at either I or Q channel.
2. Only compensate the measured NF by 3dB for image noise.
Of course, when the conversion gains of your mixer at the frequency of the real RF and image are exactly same, the compensated 3dB will be correct.
BR.