IQ demodulator required in RF front-end?

iq demodulation theory

Hi everyone,

I'm very confused about this point. To me IQ demodulation is mostly for image cancellation (yes, I'm more into RF design than wireless communications theory). Let's say I have a non coherent receiver. Do I need an IQ or not, let's say for binary phase keying modulation? I would normally say yes but I now see software radios with only an LNA and a mixer followed by an ADC! Does someone have a good articel or answer to this?

Please help!

iq demodulation

Cos & Sin (I & Q) are intrinsically orthogonal.
If you transmit BPSK without using I-Q structure , of course no problem
with it but you will decrease your data throughput.

I-Q can be regarded as "dual channel"
You transmit data by using con(wt).
Why not tranmit data by using sin(wt) simultaneously also.

They won't intefer each other and increase data throughput.

i q demodulation

by using iq demodulator, you can transform the RF signal to zero IF without frequency spectrum(positive frequency and negative frequency) overlap.

If you use mixer instead of iq demodulator, the RF signal can only been translated to low IF, not zero IF.
The low IF bandwidth is twice of zero IF, so the ADC must work with a higher sampling clock.

For example, the bpsk data rate is 1Mbps. you use iq demodulator, the ADC working clock is 2MHz. but if you use mixer, the ADC working clock must be 4MHz or higher.

Generally, mixer + ADC is easier to realize than iq demodulator + ADC. If the ADC and following digital signal processing part(FPGA, DSP or ASIC etc.) can work in higher clock and the power supply is not limited, the mixer + ADC are recommended.

regards

+iq demodulation

Thank you for your answers. I agree but if the receiver is non coherent, in both cases, you need to sync with the carrier in order to have all the energy on I (or Q why not?) for detection, don't you? but it doesn't require to receive the I and Q signals, am I still right? What I was worried about is that if the mixers in the analog front-end are not in sync with the signal, the signal becomes real and imaginary and I was afraid to loose that imaginary info in the ADC, filters, ... As a consequence I conclude that the IQ demodulator in the analog front-end is purely for image rejection...
Let me know if I'm wrong
Tanks for your help

iq demodulation with non-zero if

You can use a Squaring loop or Costas loop (it is same actualy) to recover your carrier phase. The RF mixer does not have to be coherent. You can still do coherent dectection at IF using Squaring loop to generate coherent LO. BUt RF mixer LO frequency need to be very stable. Most of the system, only TCXO for referent frequency is ok. You can also do IF sample and do Squarrinh loop digitally.

iq demodulator

Ok, things become clearer now. For all stages, a complex signal is not a problem. It's only prior to the detector that we want all the signal energy on either I or Q.

Thank you for your help
Ghoster