Microwave Office 2010 - FM discriminator question

Microwave Office 2010 - FM discriminator (demodulator) question

Hi folks! I'm just observing FM demodulation using Microwave Office (2009/2010 version), and I'm just trying to observe how well a very typical FM system really is against added noise.

I've got this simple setup (microwave office .emp file) attached, and I'm actually getting some results. But I've also got a couple of simple questions for any Microwave Office users with similar area of interest.

In my system, I use the FM_discriminator (FM_DSCRM from the MWO library) as the demodulator device. And I use a noise source to add noise to my FM signal (in order to study what will happen when the noisy FM signal eventually goes through this demodulator).

When I observe the bare output of the FM discriminator block (ie output of the FM_DSCRM block), I see a very noisy demodulated time-domain signal. This is when I'm using 0 dBm noise power for my noise source. So I decided to add a low-pass filter to the output of the FM_discriminator (ie the demodulator block). And I find that the output of the filter is quite a clean waveform (the recovered signal is 10 kHz, which matches the 10 kHz baseband/message signal).

I would really love to know whether or not the filter that I added is actually needed, or is the very noisy signal directly at the ouput of the FM_discriminator due to something that I've done wrong with my setup? I was thinking that if my filter is not supposed to be added, then the noisy output of the FM_DSCRM block doesn't really allow me to see any noise suppression quality of an FM system.

For this simple FM demodulator project - I'm only attempting to observe the ability of an FM system to deal with (suppress) noise. Can anybody tell me if I'm doing the 'wrong' thing by adding that extraneous filter after the FM_discriminator module? I see that when this filter is included, then I get a clean demodulated signal. But without that filter, I get a really noisy signal. Any nice recommendations?

Thanks very much in advance for any help!!

Your setup is fine and the noise that appears at the demodulator output (due to added WGN noise at the input) is normal.
The level of the noise at the demodulator output is so high because the WGN noise at the input is very high (-3dBm), which is unusual for a normal FM communication. If you reduce the level of the noise you will see that is looking better.
Adding a LPF after the demodulator always is a good choice to clean the signal.

Vfone - thanks very much for your time and effort in taking a look at my MWO setup for the FM demodulator! Very greatly appreciated here. Your comments help a lot, and will definitely help others. Thanks once again for everything.

FM demodulator push input RF noise power spectrum to higher frequency on demodulated baseband output. The output S/N is dependent on the modulation index. The baseband noise spectrum density goes up as a differentiator and high freq demod output spread is based on RF bandwidth. Modulation gain is therefore based on having an output LPF that attenuates at the modulation maximum frequency and the modulation frequency is much less then the input RF bandwidth (high modulation index).

Original Squelch circuits worked by detecting high frequency noise component at output of demodulator, above the voice audio frequency. The high frequency noise dropped off rapidly as input C/N increases giving the squelch detector good sensitivity.

SCA broadcast (elevator music, MUZAK) used 67 kHz subcarriers on baseband modulation. Stereo used DSB suppressed carrier subcarrier at 38 kHz. Because these components have lower modulation indexes relative to the baseband audio they require a higher C/N RF input.

With constant power envelope of FM, an amplitude limiter in front of demodulator improves demodulation S/N output. There is definite 'knee' to an FM demodulator at a C/N ratio around 10 db (with some variance to modulation index). This is due to a pronounced increase in sudden 180 deg phase reversals below this threshold, where the input noise power begins to statistical have peaks that exceed the input carrier power, which causes 'clicks' in demodulated baseband output.

Thanks RCinFLA for your additional helpful info. Very much appreciated! The things you mentioned is definitely going to be useful.