Choosing the best FM demodulation technique

[Archimedes]

So I have an FM signal: 200 mV_pp, 11.5 MHz carrier, max deviation - 500 kHz, modulating signal frequency range is 40-10000 Hz. And I need to demodulate it with lowest noise possible. The circuit has to be low voltage (3.3V or lower), low power and low profile (<4mm, so no fancy inductors and transformers). What are the best ways to do this?
I've found **broken link removed** circuit, it has super low power consumption, can be made in low profile and doesn't require much tuning, which is also great, but I'm not sure about it's noise performance. I also think about multiplying signal frequency 4-5 times with PT7C4511 clock multiplier and then demodulating it with TLC2933 PLL.

 

[BigBoss]

The specifications sound to me an analog FM system. Right ??
I used SA612 for a kind of purpose ( analog satellite reception with analog FM demodulation ) so long time ago . IF frequency was 10.7MHz and the own VCO of SA612 had a temperature drifting problem but we solved the problem by adding an extra PLL to fix the VCO freuency.Crystal VCO can also be used.

 

[Archimedes]

Yes, it is an analog FM system.
From what I understand, you suggest quadrature FM detection. I can't use SA612 since it requires 4.5V min supply, but that's not a problem, I have low voltage mixers. As far as I know, noise performance in quadrature detectors depends on Q of the resonant tank and since I can't use large inductors and only chip SMD inductors, the noise performance may be not so good. Or it still better compared to others?

 

[FvM]

Or it still better compared to others?

I guess it is. If you worry about quadrature detector noise you don't need to consider simple pulse counting detectors.

The general term noise should be better replaced by detail analysis. Which noise source do take into consideration? How about parasitic AM sensitivity?

 

[filip.amator]

I also think about multiplying signal frequency 4-5 times with PT7C4511 clock multiplier and then demodulating it with TLC2933 PLL.

The PLL bandwidth of PT7C4511 is 10 kHz and it is close to your upper limit of modulation signal frequency. I would guess that such a clock multiplier will attenuate your signal.
Have you checked the parameters of the all-digital receivers ?

 

[Archimedes]

Which noise source do take into consideration?

It's more about high SNR, not low absolute noise. My bad. So the demodulator has to have high F-V ratio in the first place and fairly low noise to keep SNR high. Now the question is, what types of noise can I encounter while demodulating FM besides white noise? Nonlinear distortions below 3% are fine (this includes parasitic AM, I guess).

The PLL bandwidth of PT7C4511 is 10 kHz and it is close to your upper limit of modulation signal frequency

I'm fine with that.

Have you checked the parameters of the all-digital receivers ?

I haven't seen any digital receivers for my carrier frequency. The only useful ICs I've found are quadrature IF detectors.

- - - Updated - - -

I've also found an interesting patent which claims low noise demodulation. What do you think?

 

[biff44]

why not just an LC/schottky diode frequency discriminator? Not gonna be much noise in that, where as digital systems have a ton of noise.