# IQ demodulation

1. Sep 12, 2009

### likephysics

This suddenly came up in a MRI class. I kinda understand IQ modulation (both amplitude and phase are modulated). But how does IQ demodulating a signal help, specially MRI signal.
In case you don't know a MRI signal, its just a exponentially decaying sine.

2. Sep 12, 2009

### waht

In demodulating IQ, you have to separate the two channels that are out of phase and usually feed them to a microprocessor.

3. Sep 12, 2009

### f95toli

IQ demodulation is very useful if you need both the magnitude and phase of the signal (I suspect this is the reason is is used in this case). There are other methods for achiving this but the nice thing with an IQ demodulator is that you can easily (and quickly) record all the information by just measuring two voltages (as opposed to measuring one voltage and one phase).
Once you have both the I and Q channel data you can either calculate magnitude and phase or if you prefer stay in the "IQ domain" and do your data analysis there; in some cases the latter is actually more convenient.

Note also that there is nothing stopping you from just recording one channel if you are only interested in the magnitude; you can always use a phase shifter to "zero" one channel and then just record the other.

4. Sep 12, 2009

### likephysics

ok thanks. So to get IQ from any signal you just feed the signal into 2 buffers and multiply one by cos and the other by sin. The cos should give the magnitude and the sine part should give the phase?
What happens if I try to IQ demodulate an FM signal or the signal coming out of an op amp oscillator?

5. Sep 12, 2009

### f95toli

No, the magnitude is just given by the norm (sqrt(I^2+Q^2)) and the phase by the phase angle (arctan I/Q) .
I and Q are just representations of the signal in the complex plane ("x+iy") so the usual rules apply.

I don't know what happens if you IQ demodualte a FM signal (I've never tried); I suspect it might get messy and I can't think of a reason why one would do that.
You can obviously use IQ demodulation it for an AM modulated signal (this is typically how it is used when looking for small signals because you can then feed the output of the demodulator directly to a lock-in amplifier).

6. Sep 12, 2009

### waht

It's not multiplied by cosines or sine. The two channels carry data in variations of the amplitude. They are fed to two independent ADC (analog to digital converter) where they are turned into 0s and 1s. The microprocessor does some signal processing magic and then it spits out the data.

I don't think it's possible to incorporated FM modulation with IQ. it will change the constant 90 degree phase shift.

here is a constellation diagram representing all possible states of QAM 16.

http://72.232.229.42/thumb/1/1e/16QAM_Gray_Coded.svg/200px-16QAM_Gray_Coded.svg.png

Last edited by a moderator: Apr 24, 2017
7. Sep 12, 2009

### f95toli

Waht: I think we are talking about different things here. What you are refering to is the "typical" use of IQ demodulation in telecom for e.g. phase-shift keying.
But I am pretty sure that is not what the OP is asking about since the application was in MRI. In this context IQ demodulators are often used much like you would use a mixer in ordinary homodyning; but -as I wrote above- the added advantage that you get both magnitude and phase of the signal directly.

I use IQ demodulation for nearly the same reason. I e.g. use it to measure the decay shape/time of high-Q resonators; I send a pulse in at the resonance frequency (several GHz) and can then watch the decay of both the I and Q channel on an ordinary oscilloscope (the IF BW of my system is around 10 MHz) meaning I can see how both the magntude and phase changes. I would assume this is why IQ demodulation is also used in MRI, where similar techniques are used.

8. Sep 12, 2009

### waht

You are right, I was confusing the issue with the actual demodulation of a telecom signal.

I was able to pull up an abstract for a patent of an MRI transceiver. It appears that the IQ scheme is used as an image reject mixer.

http://www.patentstorm.us/patents/6259253/description.html [Broken]

This is really cool. I'm a rf/microwave junky.

Last edited by a moderator: May 4, 2017
9. Sep 13, 2009

### likephysics

When I first saw the MRI decay, it reminded me of ringing and the way it decays(some times).
Anyway, by measuring the decay of the pulse I can understand you see the amplitude decay. But phase also decays/changes?
For example in a exponentially decaying sine, the amplitude decay is obvious. But where is the phase decay? The frequency of the sine does not change.

waht,

Another term I don't understand is image reject mixer. Are they really rejecting an image using a mixer?

10. Sep 13, 2009

### f95toli

No, but real systems are never 100% ideal meaning while the decay of most real resonant systems (incuding MRI) looks very much like the decaying sine you would expect from theory there are always other factors that will affect it (various types of non-linearities); it is those factors that I am interested in (I am using the resonators to probe the properties of various materials).

No, it is a mixer that reject the (mirror) image of the signal, i.e. it rejects one sideband from the IF.

A good place to read about microwave/RF techniques and equipment is www.microwaves101.com

11. Sep 13, 2009

### skeptic2

What is it that you're trying to measure with the decaying sine wave? Is it important to know how fast the sine wave is decaying?

What are the two frequencies that are being mixed in the image reject mixer? How will the product be used? For instance is the frequency or amplitude of the mixer product more important?

12. Sep 13, 2009

### f95toli

Well, since it is MRI we are talking about I would assume they are interested in parameters like T1 and T2.
This means that in general a homodyne measurement in used and what you measure is the amplitude of the signal as a function of time; meaning the signal of interest is the decay envelope of the decaying sine wavefom.

13. Sep 19, 2009

### kirubanithi

Could you please explain the different between IQ and IF signals

14. Sep 19, 2009

### f95toli

Both I and Q are IF signals. An IQ demodulator will givbe you both the in-phase(I) and quadrature(Q) IF signals whereas an ordinary mixer will only give you one signal out.

15. Sep 19, 2009

### waht

In IQ, there are two independent signals at the same frequency, but are offset by 90 degrees in phase.

IF (intermediate frequency) is a result of multiplication of two other signals. It is composed of the sum and difference of frequencies being multiplied. The electronic circuit that does the multiplication is called a mixer. The mixer is often used when you want to down or up convert a signal in frequency.

16. Sep 19, 2009

### f95toli

It is perhaps worth pointing out that an IQ modulator/demodulator is just a couple of mixers and hybrids in a single package. Hence, the I and Q outputs of a demodulator are really just the IF outputs of its internal mixers.

17. Sep 20, 2009

### kirubanithi

what is the main advantage of splitting the signal into I and Q signals.

18. Sep 20, 2009

### f95toli

Did you read my first post?

19. Sep 20, 2009

### kirubanithi

In set top box, two types of tuners are available. One is IF tuner and another one is IQ tuner.Could you please explain about the two tuners.

20. Sep 20, 2009

### waht

What is the set top box?

IF could be input or output of a converter such as IQ