Frequency Deviaition

[Yaa]

Hi i got a question regarding to Frequency deviation.

Carrier frequency = 150Mhz , channel spacing = 12.5kHz. When audio frequency varies from 300 - 3kHz, the modulation analyzer will show deviation readings around 2~2.5kHz.

What is the deviation reading at the modulation analyzer, if the audio frequency changes to 5kHz?


----------------------------------------
per my understanding is the maximum frequency deviation is -+2.5kHz.
1) is the any possibilities that audio frequency affect the deviation?
2) can you guys explain more on this?

[fourfathom]

Basic FM deviation is not determined by the modulation frequency, just the modulation amplitude.

But in practice, in a communications or FM broadcast transmitter there is audio pre-emphasis (a high-pass filter of sorts) between the audio input and the modulator.  This will cause the modulation amplitude (and thus the deviation) to increase as the frequency increases.

PM (Phase Modulation) is a different story though.  There are different types of phase modulators, but the result is essentially FM with a particular pre-emphasis filter.

[MrAl]

Hi i got a question regarding to Frequency deviation.

Carrier frequency = 150Mhz , channel spacing = 12.5kHz. When audio frequency varies from 300 - 3kHz, the modulation analyzer will show deviation readings around 2~2.5kHz.

What is the deviation reading at the modulation analyzer, if the audio frequency changes to 5kHz?


----------------------------------------
per my understanding is the maximum frequency deviation is -+2.5kHz.
1) is the any possibilities that audio frequency affect the deviation?
2) can you guys explain more on this?

Hi,

It might help if you look up some methods for these kinds of circuits.
For example, in AM we can look at:
y=sin(t*w1)*sin(t*w2)*A*B

where w1 is the carrrier frequency and w2 is the modulation frequency and A and B are the amplitudes.  This can be resolved into:
y=(cos(t*w2-t*w1)*A*B)/2-(cos(t*w2+t*w1)*A*B)/2

and that is the same as above yet it involves just a subtraction, and the sum and difference of the two frequencies.

[CaptDon]

It is a difficult thing to picture mentally. As the modulating index changes so will the number and amplitude of the sidebands. So strangely enough you can have a 1KHz modulating signal with only a measured deviation of 100Hz at low modulation levels. But on an analyzer you will still see the sidebands at 1KHz! Carsons rule states that total bandwidth of an FM signal which will include 98% of the total signal power can be represented by (Tone + Dev) X 2. An example would be (1KHz Tone + 5KHz Dev) = 6KHz X 2 = 12KHZ occupied bandwidth. As you increase the modulation index (ratio of deviation to tone) sidebands and even the carrier will come and go. With a good quality spectrum analyzer a handy way to set 5KHz deviation is to use a 2.079KHz tone and slowly increase the deviation until the first instance of the carrier disappearing. This magic carrier disappearance occurs at a modulation index of 2.405, 5.52, 8.654 and so forth. So by Carson's rule occupied bandwidth will be bigger than the 'deviation'. Deviation alone does not describe bandwidth, the modulating tone must also be considered. It is amazing to look at an FM commercial broadcast station who has the 67KHz and the 91KHz subcarriers turned on. Even stranger to consider a 91KHz subcarrier on an FM station using +/- 75KHz deviation? These are my observations only. I am sure there is some Einstein math that describes my observations.

[newbrain]

Had started answering when I was you post. Some effort spared, thanks!

I am sure there is some Einstein Bessel math that describes my observations.

[fourfathom]

Hi,

It might help if you look up some methods for these kinds of circuits.
For example, in AM we can look at:
y=sin(t*w1)*sin(t*w2)*A*B

where w1 is the carrrier frequency and w2 is the modulation frequency and A and B are the amplitudes.  This can be resolved into:
y=(cos(t*w2-t*w1)*A*B)/2-(cos(t*w2+t*w1)*A*B)/2

and that is the same as above yet it involves just a subtraction, and the sum and difference of the two frequencies.

But this isn't normal AM modulation, your equation describes linear mixing.  In AM modulation the carrier phase never inverts (as it does in the equation).  Add a constant offset to the sin(t*w2) term so at the negative peak of the modulating signal the term does not go negative.  That's AM modulation.  An image should be easier to visualize -- let me search for one:

(wikipedia)

[TimFox]

In that equation of Mr Al for AM, the result is actually "DSB", meaning double-sideband, suppressed carrier.
The difference is when the "baseband" modulating signal goes through zero, the output of that modulator also goes through zero volts.
In AM (double sideband but with carrier), when the modulating signal goes through zero the output is an RF sinusoid whose amplitude ("carrier") is 1/2 the maximum value .
As shown in the animation above, the output level goes from zero (at max negative baseband signal) through "carrier", to maximum output.
True AM is less efficient than the suppressed-carrier versions, but the required demodulation is far simpler (merely following the peak values and filtering out the carrier frequency to obtain the baseband), allowing an elementary "crystal radio" circuit to work.

[ejeffrey]

Even if you move beyond crystal radios to more sophisticated devices like a superhet receiver, AM is easier to build a receiver for.  It's easier to lock a PLL to the signal when it doesn't go to zero, and it also provides an amplitude/volume reference.  With DSB-SC you can tell the difference between lower signal and RF attenuation.  You either have to manually adjust that or use compression.  It's OK for voice but not great for e.g., music.