|
From: | Marcus Müller |
Subject: | Re: [Discuss-gnuradio] Fwd: Creating a FFT plot like the one in this youtube variable |
Date: | Wed, 22 Jul 2015 22:55:48 +0200 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.1.0 |
Hi Ashraf, I don't know what frequency f_squarewave your square wave has, but rectangular signals have sinc-shaped envelope, with peaks every f_squarewave. Having a sinc envelope especially means it exists over the whole nyquist band -- so that explains why you see your complete spectrum rise! I think what you should do is have a look at the spectrum of your transmit signal. This will make it easier to understand what you see at the receiver. Try this: http://i.imgur.com/EARpJLx.png so, a squarewave with frequency of f_sample/32 (remember, there's no real frequencies in DSP -- this really just tells GNU Radio that a period is 32 samples long). Important: the QT sink is set to have an FFT length of 1024 points. You should see this: http://i.imgur.com/33xrCl6.png Now, if you have a look at this spectrum, you'll notice deep "wells" between the peaks, which aren't there in the receive PSD, right? There's a simple reason for that: In the simulation / the TX spectrum, your f_squarewave is probably an integer factor of f_sample. This means that the period is a whole number of samples, and that whole number of samples also happens to be an integer factor of your FFT length! Therefore, your signal is perfectly periodic (as far as the observer can tell), and hence, has line spectrum characteristics. Now, we'll play around with the frequency of the square wave: http://imgur.com/Kduv5fL set it to f_sample/32/2**0.5 ; the root of 2 is not a rational number, so no FFT window in this world could transform this without leakage. So you get this: http://imgur.com/1Fkl8f8 Looking familiar? Now, these are extremes. But the problem here is that frequency synchronization between your N210 and B200 will not be perfect -- hence, on the receiver side, the signal period might not be as exactly the integer factor that you have on the TX side, and you see the inter-peak leakage. I hope that has explained the most of this phenomenom. Best regards, Marcus On 22.07.2015 20:40, Ashraf Younis
wrote:
|
[Prev in Thread] | Current Thread | [Next in Thread] |