salar
New Member
Posts: 2
|
Post by salar on Aug 25, 2011 13:46:40 GMT -5
I am creating an FM modulation by using 13 MHz carrier signal and 100 Hz signal. Than I am suppose to increase the frequency from 100 Hz to maybe 100 KHz in order to measure the loop bandwidth in a PLL. I have some question that maybe you can help me. 1. when I create this FM modulated signal, I guess I will have some component in my spectrum in …., 12.8, 12.9, 13 MHz, 13.1 M, 13.2 ….. . The low pass filter that we have is supposed to remove the high frequencies. Am I right? If we assume that the bandwidth of the filter is 50k then what would be remain?how would be the shape of spectrum? 2. what should I set for FM deviation in this case? Do you have any suggetion for my measurements?
|
|
|
Post by manzana on Aug 25, 2011 21:12:17 GMT -5
I do not believe she can answer this without knowing what kind of lp filter you are using. If it isnt a deep secret, how many poles does it have? Is it a simple RC?
|
|
|
Post by kalyanasv on Aug 28, 2011 12:19:39 GMT -5
1.) you might consider using a bandpass filter. than just an lpf to remove side band harmonics. 2.) consider your system seems to a wideband FM. If you take beta=0.3 then you get a del f1=0.3x100 = 30hz. Then del f2 =0.3 x100K = 30000. This means you have a del-f variation of 30000/3000 i.e. 10000 . This implies you need a multi-stage FM atleast 3-4 stages for stabel operation. I would suggest to reduce the range to maybe 500Hz - 20KHz and deal with a two -three stage modulator which is more typical.
|
|
|
Post by manzana on Aug 28, 2011 21:02:46 GMT -5
(Isnt this fun!) I believe his loop filter has already been chosen and he is just trying to measure the loop bandwidth. (?). Why not just give the loop a good step; try changing from the lowest frequency to the highest and deliberately try to lose lock. You will learn a lot about the loop dynamics by looking at the error signal (output of the lpf) and see how long it takes to recover and what shape the response is. PLL's have lots of gremlins in them so repeat the experiment many times and then with a multitude of step sizes. PLL's have sent more engineers to rubber rooms than any other contraption... ;D
|
|
salar
New Member
Posts: 2
|
Post by salar on Aug 29, 2011 4:25:23 GMT -5
Yes I have my low pass filter and I am just trying to measure the bandwidth of loop filter. My first problem is how to choose the FM deviation for my reference signal? It is not that much clear for me what is the effect of FM deviation. When I increase the frequency from 100 Hz to above, the amplitude of the signal in the output of LPF will change (Due to LPF tranfer fuction). But I do not know in which FM deviation should I do my measurements
|
|
|
Post by manzana on Aug 29, 2011 20:48:08 GMT -5
Salar, How pregnant your questions are... As Charan describes in her FM tutorial, as you increase the FM deviation you are increasing the frequency spread of the signal you are putting into the loop. This increases stress on the loop . But it seems like you have a 2 dimensional problem. As you increase the frequency of the reference signal this stresses the loop also. Just to be clear; I assume we are talking about a PLL and not a Frequency Locked Loop since the 2 behave differently. If you are using a PLL with a digital phase detector (an exclusive OR) I dont see why the amplitude of the LPF would increase with frequency. Seems like it would increase in *frequency* and thus decrease in amplitude. Would welcome other insights.
|
|