IQ Modulator for 5 MHz to 20MHz Wireless Application

[flameboy666]

Hi,

Would like to know anyone has experience on IQ Modulators?

Need more explanation and I am stuck in the design of the IQ Modulator.

Thanks!

 

[karesz]

Hallo,
Analog Devices & Maxim-IC`s home sides?
K.

 

[Garbonzo]

What do you want to know? I've done work on instrumentation grade IQ modulators a couple jobs ago, though a few decades higher in frequency than you are looking for.

 

[flameboy666]

What would the basic standard design have? What are the main sections we need to build?

I would know it will be naive to expect a complete answer, but I am a bit clueless on the project. I am currently in the double balance mixer section.

Please do reply. Thank you.

 

[Garbonzo]

OK, broad subject, but here goes...

Let's start with a simple mixer. If you apply a low frequency input to the IF you can at best get AM modulation of the carrier. The output spectrum is symmetrical, and you'll also likely get some LO leakage (aka carrier feedthrough, aka carrier leakage, aka etc).

If for example you want to create an output at a frequency, say an LO of 1 GHz, and a desired output at 999 MHz, you can try putting a 1 MHz signal into your simple mixer, but you get 999 MHz, a little leakage at 1 GHz, and 1001 MHz on the output.

IQ:

IQ simply refers to In phase, and Quadrature. So lets for the moment use 2 mixers. MixerA will get the same 1 GHz LO, Mixer B will get a slightly delay signal (90 degrees). We will combine the outputs of these with some sort of summing device (resistive splitter, power combiner, etc). Importantly you need a 90 degree phase shifter at either the LO or RF, but not both (i.e. use one 90 degree splitter, and either a 0 degree combiner or 180 degree combiner).

First you should see there are 2 IF inputs, IF_A, and IF_B. Simply putting in our previous 1 MHz signal into just IF_A, or IF_B gives the same set of output signals as far as we can see with a spectrum analyzer. However it should be noted that due to the 90 degree phase shift on Mixer_B's LO the output spectrum is 90 degrees off as well (i.e. if we call the tones from MixerA 0 degrees, the tones from MixerB will be 90 degrees shifted at the output, but of the same amplitude).

Next, lets take out 1 MHz IF signal and put it through a 0 degree power splitter. We'll put one output into IF_A. We'll pass the second output through a phase shifter (such as a long piece of coax line) before putting it into IF_B. For these purposes lets adjust the delay line to be right at 90 degrees long at 1 MHz.

So what happens? Mixer_A still ouputs the same 3 tones at 999, 1000, and 1001 MHz. Mixer B puts out a very similar set of tones, though at a different phase. If phases are carefully added up you'll see that the 1001 MHz signals will be 180 degree out of phase (90+90=180), while the 999 MHz signals are in phase (90-90=0). The leakages at 1000 MHz are simply pot luck. The resulting spectrum is a strong signal at 999 MHz, a little leakage at 1000 MHz, and a mostly cancelled out signal at 1001 MHz.

With a little more work you can also sum in a little DC voltage at IF_A and IF_B to null out the carrier leakage at 1000 MHz (you are creating a second 1000 MHz signal which you adjust to null out all the leakage signals). Similarly slightly tweaking in the gain and phase to compensate for subtle differences in the mixers and errors in your splitting circuit you can very nearly null out the upper sideband at 1001 MHz.

So, from there, what's your next question on IQ modulators? The real fun happens when you have to make the LO a decade wide (phase shifters on the LO get wacky), and make the whole thing super temperature stable. The IQ modulator I worked on is in the Agilent E8267 20 GHz complex modulated source.

 

[flameboy666]

Basically, IQ Mods will help to allow us to get a more accurate reading that we desire?

IF_A will be fed with a RF signal while IF_B be fed with LO.

A phase shifter is needed at:
1. Before(90 degrees) LO of IF_B
2. Before(0 degree) LO of IF_A - if it is 0 degree, it is not necessary to put it?

A power splitter is needed at:
1. Before(90 degrees) of IF_B
2. Before(0 degree) of IF_A

For the summing device, what is recommended? And is it easily available or self made?

For the phase shifters, is it easy to DIY?

How do you manipulate with the DC voltage, gain and phase to compensate for subtle differences in the mixers and errors in my splitting circuit?