Hi. I cannot understand the I&Q Modulator…
I am reading this one of mini circuits…http://217.34.103.131/pdfs/ZFMIQ-10M.pdf
1)It has I and Q inputs ANALOG right?
2)The I&Q Modulators as i understand make some kind of modulation… Analog or Digital or both modulation?
The questions is:
----if i want to make QPSK … i must have for input as i know bit stream I and Q (modulating signal) and i need a carrier (modulated signal) i compine them and i make QPSK siganl. With the I&Q Modulator i need an extra DAC to do this thing?
----If i want to make AM DSB modulation what i am doing? just use the I port or Q port and put my analog signal it it ?
for AM DSB-SC i use the I and Q port?
Many thanks
George
This may be of help: http://www.home.agilent.com/upload/cmc_ … =US&lc=eng
http://www.home.agilent.com/upload/cmc_ … lation.pdf
To do QPSK, I and Q must already be encoded. And in the case of the Minicircuit modulators, they need a bias (iirc). For AM DSB-SC you pick either I or Q. For standard AM; you modulate both I & Q and make sure to add offset on top of the bias (the offset adds the carrier - e.g. if the signal has 600mVpp, then it must be offset by 300mV - this is on top of the bias required by the modulator).
So…first of all .this Modulator of Mini Circuits has I and Q ports. From my device i put into ANALOG SIGNAL right? I=COS…
Q=SIN…
I don’t understand this with the QPSK how can i do?
You need to read-up on encoding of QPSK signals (http://www.maximintegrated.com/app-note … mvp/id/686). For QPSK the carrier can represent four distinct symbols. Each symbol represents 2 bits. For example one symbol can represent “00” (simultaneous 0’s applied to both I and Q), another symblol can represent “01”(simultaneous 0 to I channel and 1to Q channel). This encoding allows transmission of 2 bits instead of 1.
Fovakis:
So…first of all .this Modulator of Mini Circuits has I and Q ports. From my device i put into ANALOG SIGNAL right? I=COS…
Q=SIN…
I don’t understand this with the QPSK how can i do?
I don't know that you can do what I'd consider "normal" QPSK with that device.
Think about it this way. If you applied a single sinewave to just the I input, the RF output would be a DSB-AM signal with little carrier suppression. Now turn that I signal off and immediately apply exactly the same signal to the Q input. You get a similar output but the phase is now shifted by 90 deg to the prior signal. You’ve got a form of BPSK.
Now imagine phase shifting the I signal by 180 deg. Apply just that signal to the I input and you’d have another flavor of BPSK but the phase shift would be 180 deg btw signals. Do the same to the Q signal and now time that signal to be on/off when the I signal is off/on and you’d have a form of QPSK. Is that what you’re looking for ?
WRG to efficiency you’d be better off using delay lines to generate a carrier, and 3 other phases shifted by 90, 180 and 270 deg, and using a 4 throw RF switch to output the QPSK waveform.
And even better would be …
I agree with you languer and Mee_n_Mac but there is something i must consetrate to understand…WHAT IS THE I AND Q SIGNAL THAT ARE INPUT TO MODULATOR? HOW THEY HAVE CREATED?
The I/Q Modulator has four ports. I and Q (for information) , LO (for carrier) and OUTPUT.
So, for PSK,as i imagine ,the I and Q ports of modulator, will have a baseband signal of a bit stream that has become more smooth from a raised cosine filter. The process of raised cosine filter is being done in the DSP. So the output of DSP is a sinewave of I with some bandwith and frequency the max frequency of the bandwidth and the same signal with 90 degrees shift is Q. These signals will modulate the carrier of the L.O…am i right?
You are correct. The I & Q are (for most purposes) baseband signals. These modulate the LO. I can tell you that even though I could read and mathematically understand these concepts quite well, they really didn’t clikc completely until I tested them in Mathcad or Matlab (and later on in a real application). But I am going to try to elaborate a bit more.
Say you have two identical sinewaves for I and Q. Say they run at 1kHz and are sampled at 2kHz. If you sample them correctly you could effectively samples them at their peaks and zero crossings all the time. Now say you can delay one of the two (pick Q) such that every time it is sampled you can control whether it is a peak or a zero crossing. This is analogous to the encoding. In real life you create the waveforms, and use a look-up table to determine their phase. These I & Q drives the modulator. So one implementation of M-PSK could use a lookup table which encodes the symbol on the I & Q channels by adjusting the phase of one or both the sinewaves to match the sample rate.
Now what you do on the receive end. Well you demodulate the RF signal in the same way you modulated it. You feed an LO to the demodulator, you apply some filtering and get a two nice waveforms for I and Q. Now the trick here is that you have no clock to synchronize against. So you have two sinewaves that should be synchronous to each other, but unless you can sample them at the same rate, but more importantly at the same position as they were when modulated then you don’t get the same info. As an example, lets say you sample them a 2kHz, but your starting point has you sampling them not at the peaks and zero crossings, but half-way on their way up and half-way on their way down (sort of hard to see without pictures I know). The information you gather from them (i.e. the 00, or 01, or 10, or 11 - for QPSK) may not be the same. So the receiver must have a means to “recover” the sample clock (or a clock which is synchronous to it) so that you can have everything synchronized (i.e. make sure all you sample points on the receive, match the sample points at the transmit). To do this you implement some mechanism of “clock recovery” (or in some other systems where the clock is not recovered, it is “re-created”). That way when the transmit sends a 00, the receive can interpret it as 00.
I really hope I did not blunder this too badly.
i want to test it also in matlab. i know some basics and now i will see how to write code. any help about the matlab?