Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

I&Q Modulator questions

  1. May 22, 2013 #1
    I&Q Modulator questions....

    Hi. I cannot understand the I&Q Modulator...

    I am reading this one of mini circuits..

    It has I and Q inputs...and it write (DC to 2MHz) freq for the input..

    1)is the I and Q analog signals or bit stream ?If is a bit stream how can it has 2MHz frequency?

    2)The I&Q Modulators as i understand make QPSK/8QAM/16QAM etc modulation..so the message I and Q must be bit streams (symbols basic) and carrier analog signal..right?

    3)The output of a I&Q Modulator is exactly like the output of a AM-Modulator (?)...USB, LSB , carrier...what is going on here?

  2. jcsd
  3. May 22, 2013 #2


    User Avatar
    Science Advisor
    Gold Member

    1) I and Q are the (I)nphase and (Q)uadrature phase of an an analog signal.
    The 2 MHz is the modulation frequency. Remember that an IQ modulator is basically a fancy mixer which allows you to play with both amplitude AND phase. The 2MHz would be the IF on an ordinary mixer,

    2) They can be, but an IQ modulator is an analog component, I and Q can be whatever you want them to be. IQ networks can e.g. be used as single-sideband modulators (which is what I mostly use them for)

    3). Again, a IQ modulator is like a very versatile mixer. By varying the amplitude and relative phase of I an Q you can create all sorts of interesting waveforms.
  4. May 22, 2013 #3
    I like your reply!Thanks... i have some additional questions..

    At the output of my DSP i have a 16-QAM analog signal fc=10KHz and BW=20KHz like this one: http://i1284.photobucket.com/albums/...pse9ee0bf9.png [Broken]

    I have one output port at DSP.

    So i can use a 90 degree splitter to create the same 16QAM like above but with 90 degree shift? So i will have I and Q signals

    And then use a I/Q modulator to upconvert this one with carrier 100MHz?

    Last edited by a moderator: May 6, 2017
  5. May 23, 2013 #4


    User Avatar
    Science Advisor
    Gold Member

    I can't see the picture (I am at work and our firewall is blocking it)....

    I can't really answer your question fully, because I have no experience with digital data (I use I Q modulators for analog applications).

    However, two things comes to mind. Firstly, you can't use a passive 90 degree splitter for 10 kHz, you'd be better off doing this digitally or -if you have to- using an op-amp.

    Secondly, I don't quite understand what you are trying to do. I don't know much about 16quam, but shouldn't the DSP give you both the I and the Q?
    The phase shift should not be a constant 90 degree but should vary, that is -AFAIK- the whole point of QUAM. If you just use a constant 90 degree shift you will just be AM modulating, effectivly using the IQ modulator as a Single-sideband modulator (useful, but it is not 16quam).
    Last edited: May 23, 2013
  6. May 23, 2013 #5
    f95toli many thanks for your reply.

    1)I agree with you i cannot find a 90 splitter at 10khz. No way..maybe an operational yes.
    The DSP at this time has one port output. The output is an analog signal like this:

    16QAM_zpse9ee0bf9.png "] 16QAM_zpse9ee0bf9.png [/URL]

    My DSP output is already a 16QAM analog modulated signal. I want just to upconvert it at 20 MHz with one I and Q modulator SSB-SC.
  7. May 23, 2013 #6


    User Avatar
    Science Advisor
    Gold Member

    OK, so all you really want is to configure the IQ modulator as a single sideband modulator :)
    In that case, it should work.
    However, note that in order to get could surpression of the unwanted sideband you might need to adjust the phase a bit; 90 degrees is for an ideal mixer and in real life the phase that gives the best surpression might be a bit different. That said, you would need a spectrum analyzer to check this and if you don't have one 90 degress will be better than nothing.
  8. May 23, 2013 #7
    Yes exactly IQ Modulator as a SSB Modulator. Maybe so, i need tunable degrees shifter. (i don't know if exists).

    So, if i had a I and Q outputs from my DSP baseband with fc=0 and BQ=20KHz ... what i must do if i want to upconvert this in 20MHz ? Again IQ Modulator?
  9. May 23, 2013 #8


    User Avatar
    Science Advisor
    Gold Member

    You can just use the IQ modulator as an SSB mixer. The signal goes into I (or Q) and the phase shifted signal (+90 degrees) into Q (or I). All you need then is a 20 MHz LO of the right power (50mW according to the datasheet).

    You should be able to use an all-pass filter as your phase shifter. There are passive implementations, but they tend to use very large inductors so I would personally avoid that.
  10. May 23, 2013 #9
    Nice about I/Q MODulators...many thanks my friend!

    About attenuators can you propose me something? I have a sine source at 10khz power -20dbm that going to a mixer..L.O=30.010KHZ and RFport=30.000khz .conv.gain=-6db
    The l.o port is 7dbm and i have problems with power leakage..i watch to a oscillospoce the input IF port of mixer and it is not a sine...is has a lot of phase noice and amplitude variations maybe from power leak from l.o port..what can i do for this?>
  11. May 24, 2013 #10


    User Avatar
    Science Advisor
    Gold Member

    Which mixer are you using?

    Assuming you are using the right LO power etc, I don't think there is a lot you can do if the problem is really leakage (except get a better mixer, that is)
  12. May 24, 2013 #11


    User Avatar
    Science Advisor
    Gold Member

    You are making a distinction here that does not exist. The only time 'digits' are just digits is when you are talking about the information they carry: i.e. when the data is being treated as data. A real signal which is carrying digital information (binary or whatever) never has 'square edges' - it will always be a (filtered) analogue signal that modulates any carrier you use.

    QAM is, as you say, the result of modulating two quadrature carriers (any frequency you like) with the amplitude of two digital signals - which are filtered to limit the bandwidth of the resulting QAM signal. I'm not sure why your DSP has been chosen to produce a QAM signal with such a low carrier frequency. It makes the up-conversion process much more fiddly.
    Why not produce an original QAM signal with a practical IF carrier frequency (say 200kHz or higher) and then just use a straightforward up-converter, with a local oscillator, mixer and simple band-pass filter to get your QAM signal at any frequency you want. I thought that it is standard practice to produce an IF signal which is modulated / channel-filtered etc. and then up-converted to a convenient frequency for transmission.
  13. May 24, 2013 #12
    @f95 toli ..i am using this one of mini-circuits
    i feed it with 7dBm as it requires..i don't understand what is the L.O to RF isolation 45dB typical...isolation is the opposite of leak power..so what 45 dB means for my signal's power?

    @sophiecentaur.. yes..i think better definition is:

    In the begin, we have a digital stream of bits I and Q , then they get filtered to a Raised Cosine Filter so they form a BasebaNd Analog Signals I and Q with BandWidth..then this Baseband signals modulates a carrier...

    The DSP's clock are very low..it cannot upconvert up tp 10-15Khz...so i must make the upconversion on my own maybe with a analog I/Q Modulator
  14. May 24, 2013 #13


    User Avatar
    Science Advisor
    Gold Member

    Ok. So you need quadrature feeds of a local oscillator. Probably better to get them from the oscillator itself than using a phase shift circuit. Two divide by two circuits off a double speed clock will deliver two quadrature LO feeds. You can do the band pass filtering after the modulator to eliminate all the out of band harmonics, rather than having to use two well matched tight filters for two 'good' sinewave drives.
  15. May 24, 2013 #14

    why i must need quadrature feed of a l.o...?

    this modulator that i want to buy has a single port for l.o
    So, i feed this port with the carrier for example 20MHz right?

    I don't understand this"
    You can do the band pass filtering after the modulator to eliminate all the out of band harmonics, rather than having to use two well matched tight filters for two 'good' sinewave drives."

    What do you mean ABOUT" well matched tight filters"? Are you reffering about the raised cosine filter that DSP uses to make a bit stream more smooth like sine to eliminate the bandwidth? i think i need this type o filtering...

    At the output of I/Qmodulator what will be the result? An SSB-SC right? with a lot of harmonics?
  16. May 24, 2013 #15


    User Avatar
    Science Advisor
    Gold Member

    There are many ways of skinning a cat. You may as well ignore most of what I wrote because I was thinking you'd need to be producing your own quadrature LO feeds.
    I hadn't realised that mini circuit has its own phase shifter inside it, which provides it with the appropriate two quadrature carriers for free. So you just feed it with a sensible LO frequency of your choice and give it the two I and Q filtered baseband data streams. (I assume that is esy to do). The (slow) DSP machine is not needed, for producing the QAM. In fact you are just making life harder for you by using it because you then have to follow it with an SSB modulator with all the problems of balancing out the adjacent unwanted sideband and carrier. This IQ modulator will give you the QAM signal at any convenient carrier frequency and you can mix it up to your 20MHz. It's a piece of cake to make a mixer which will beat up a signal of 10MHz (or whatever you choose) to 20MHz and you can easily filter off the local oscillator and image sideband.
  17. May 24, 2013 #16
    Question: so i put the I and Q baseband to iqmodulator with carrier 20mhz.... thi i and q basebands have fc=0khz and bandwidth 8,5-0=8.5 khz.

    At the output of the modulator i wii have an AM modulation with SSB-SC?ANY harmonics? i must put filter after?
  18. May 24, 2013 #17


    User Avatar
    Science Advisor
    Gold Member

    The two data streams are just baseband data streams which will be used to AM each of the two quadrature carriers, surely (I can call you Shirley, can't I?). Why would you say they have a carrier frequency of zero?
    Each data stream will suppressed carrier amplitude modulate each of the RF carriers in a balanced modulator. I don't see why it would be SSB. The QAM spectrum is symmetrical about a carrier, is it not?
    That mini circuit spec sheet says it operates around 10MHz, so it would not work well at 20MHz. the phase angle would no longer be a good 90 degrees.
  19. May 24, 2013 #18
    What Shirley means? My name is George :P

    Not carrier, frequency center the basebands have zero. The carrier frequency is 20MHz.

    The output of the modulator will not be a QAM signal. The I cosx component baseband will AM as you said the LO carrier cosy and the Q baseband sinx will modulate AM the siny apo LO carrier. Then these summed...

    check here please:

    1) http://www.cnam.umd.edu/anlage/Microwave Measurements for Personal Web Site/5989-7057EN.pdf

  20. May 25, 2013 #19


    User Avatar
    Science Advisor
    Gold Member

    Sorry - "Shirley" sounds like "surely" - joke. Reference the film 'Airplane' with Leslie Nielson. Also, see "You have clearance, Clarence" and "Roger, Roger" etc....

    There may be some confusion here. Afaik, an IQ modulator can be used to produce SSBSC modulation when that is what you want (as in HF Communications or broadcasting etc.). To do this, you feed the Iand Q inputs with the same baseband signal but with one of them phase shifted. The result of adding the two RF outputs is to cancel one sideband and the carrier.
    It can also be used to produce QAM. With QAM, there are two independent baseband input signals and the resulting output consists of a carrier with a range ('constellation') of resultant amplitudes and phases (the QAM signal). The QAM spectrum is symmetrical about the (suppressed) carrier frequency. The reference (Agilent) you quoted says precisely this.
  21. May 29, 2013 #20

    thanks for your reply my friend . very helpful
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Similar Discussions: I&Q Modulator questions
  1. Q-factor question (Replies: 10)