FM band filter and amplifier

[XRISTOS]

Has anyone an idea where i can find a bandpass filter in fm band range(88-108 Mhz) and an IC amplifier working in the shame frequency range?

 

[audioguru]

FM radios already have the tuned circuit and RF amplifier.
Why do you need more?

 

[XRISTOS]

Iam working my thesis....iam supposed to build a digital fm demodulator with FPGA so i need a bandpass filter and an amplifier as the rf front in of ADC and fpga!!!Is there anything in the market or i have to build them myself?

 

[RadioRon]

You will probably have to build the filter yourself from inductors and capacitors. This is how your typical FM receiver does it. As for the amplifier, there are many ready-made ones in IC form that you can choose from. It isn't really too difficult to make your own from a transistor and most older FM receivers typically integrated the bandpass filters into the input and output sections of the first amplifier, making it a tuned RF amplifier.

For an easy-to-use amplifier, you could choose one of these:

https://www.electro-tech-online.com/custompdfs/2009/03/upc3237tk.pdf

https://www.electro-tech-online.com/custompdfs/2009/03/ZX60-33LN.pdf

https://www.rfmd.com/pdfs/3827DS.pdf

 

[Roff]

Iam working my thesis....iam supposed to build a digital fm demodulator with FPGA so i need a bandpass filter and an amplifier as the rf front in of ADC and fpga!!!Is there anything in the market or i have to build them myself?

Do you also have to make a digital tuner?

 

[audioguru]

An FM radio has an IF amplifier with a lot of gain and sharply tuned circuits. The sharply tuned circuits are crystal filters today and provide selectability for the FM radio. The very high gain allows the IF amplifier to limit which eliminates AM (static type of interference) pickup. The very high gain also provides good sensitivity.

I can't remember what in an FM radio circuit provides a good capture ratio.

 

[XRISTOS]

I''ll tune my signal inside the fpga with the help of a numerical control oscillator and a multiplier.....but i want my incoming signal to be band limited within the range of 88-108 Mhz.

 

[flat5]

Capture is automatic. The stronger signal wins.

 

[audioguru]

Capture is automatic. The stronger signal wins.

Yes but a cheap circuit has a poor capture ratio and a good circuit has an excellent capture ratio. The capture ratio helps to avoid distortion caused by multipath.

 

[RadioRon]

If you assume a clean, interference-free, input signal, such as might be the case if this were an exercise for schoolwork, then perhaps you don't need to work too hard on the input filter and amplifier. The filter will exclude out of band energy. If you are only feeding a signal generator into this receiver such as one might do to demonstrate principals in a lab environment, then you really have no other signals to alias into the desired one, so a bandpass filter isn't really too important. The amplifier's job is mainly to lower the overall noise figure of the receiver and so increase the dynamic range by increasing sensitivity. Again, not really critical for a lab demonstration.

On the other hand, if this is supposed to be a practical design, then I hope you are calculating the noise floor of your A/D and determining from that how much gain you will need in the amplifier. In addition, you probably should consider including a bandpass filter both before and after the amplifier. The one before the amp serves to keep strong signals out of the system and avoid overload. The one after the amp helps with preventing out of band signals getting to the A/D and also knocks down the higher level broadband noise floor that is coming out of the amp, noise which could alias into your signal.

In a practical receiver there are many considerations. Dynamic range is always an important one. A practical dynamic range for an FM receiver might be 70 dB. So you either need to have an A/D that can handle that range, or you need to implement an AGC prior to the A/D.

What sampling rate are you considering?

 

[XRISTOS]

This is my thesis!!I want to implement a software defined radio fm broadcast receiver!!! I consider to sample at 80 Msps!!I already find by the way the bandpass filter....https://www.electro-tech-online.com/custompdfs/2009/03/2722-1.pdf i consider to use an amplifier with 20dB gain at range of 20-500 Mhz from analog devices....so the rf front end circuit would be antenna-b.p filter-amp-b.p filter!!!

 

[ecerfoglio]

This is my thesis!!I want to implement a software defined radio fm broadcast receiver!!! I consider to sample at 80 Msps!!

If you want to receive a 88-108 MHz signal, don't you need a sample frecuency of al least 216 Msps? (Nyquist's theorem?)

 

[stevez]

The ARRL Radio Amateur's Handbook and similar publications have a lot of information on filters - high pass, low pass and bandpass. Much of the information is in a form that is not specific to amateur bands.

 

[RadioRon]

This is my thesis!!I want to implement a software defined radio fm broadcast receiver!!! I consider to sample at 80 Msps!!I already find by the way the bandpass filter....https://www.electro-tech-online.com/custompdfs/2009/03/2722-1-1.pdf i consider to use an amplifier with 20dB gain at range of 20-500 Mhz from analog devices....so the rf front end circuit would be antenna-b.p filter-amp-b.p filter!!!

This filter looks very good and should do the job well, although it is likely more expensive than doing it yourself.

80 Msps may be acceptable. This is Undersampling and is commonly done for bandpass signals. A quick review of Wikipedia on this even gives an example of sampling an FM band signal:
**broken link removed**)
The article points out there are several choices for a good sampling frequency and their example mentions 44 and 56 MHz. Usually one seeks to use the lowest practical sampling frequency in order to keep the hardware as cheap and simple as possible. I wonder why you chose a freqeuncy as high as 80 MHz?

 

[XRISTOS]

This filter looks very good and should do the job well, although it is likely more expensive than doing it yourself.

80 Msps may be acceptable. This is Undersampling and is commonly done for bandpass signals. A quick review of Wikipedia on this even gives an example of sampling an FM band signal:
**broken link removed**)
The article points out there are several choices for a good sampling frequency and their example mentions 44 and 56 MHz. Usually one seeks to use the lowest practical sampling frequency in order to keep the hardware as cheap and simple as possible. I wonder why you chose a freqeuncy as high as 80 MHz?


Yes i use a sub sampling procedure due to the fact that my signal is band limited!! You are right i can use even smaller sample rate!!Another question is there a circuit that i can use to limit the amplitude of my singal to avoid noise amplitude variations? As we know frequency modulation is invariant of amplitude....so variations of amplitude will cause errors in demodulation.

 

[RadioRon]

In an analog system this is commonly done using a "limiter". A limiter is usually an amplifier with so much gain that its output clips the waveform, hence it limits the amplitude. You can also create a simple limiter with two back to back diodes to ground shunting your signal path.

In your architecture, it will not be possible to include such a simple limiter because you will be amplifying the entire band. If you amplify the entire band with so much gain that the amplifiers are limiting your signals, then you have created a mixer and all of the different FM broadcast signals in your band will mix with each other and create a complete mess in the spectrum. Limiting really only works if you limit the incoming spectrum to a single signal, which is about 150KHz wide (approximately). But to do such limiting you would need a narrow filter, and (.....here we go down the entire justification for a superheterodyne architecture...) such a filter is not practical at 88 to 108 Mhz and so requires an IF frequency.

Anyway, I think you have no choice but to implement a limiter in numerical form.

 

[audioguru]

The input amplifier does not have automatic-gain-control so it will be overloaded by all the local stations anyway. Then it limits and mixes all the stations together into a big mess (like my cheap Sony Walkman FM radio).

 

[XRISTOS]

If i consider to implement the rf front end the same way that super hetero dyne receiver works with IF at 10.7 Mhz, where i can find staff like an rf amplifier an if amplifier filters and a local oscillator-mixer? Is there anybody know the implementation method-procedure I am supposed to use for mixing the rf signal to if?

 

[audioguru]

Parts to make super-heterodyne FM radios are hard to find today since manufactured radios are cheap and perform very well.
Buy a good radio, take it apart and use its parts in your circuit.

Radio tutorials explain how the mixer works.

 

[RadioRon]

If i consider to implement the rf front end the same way that super hetero dyne receiver works with IF at 10.7 Mhz, where i can find staff like an rf amplifier an if amplifier filters and a local oscillator-mixer? Is there anybody know the implementation method-procedure I am supposed to use for mixing the rf signal to if?

I think that you should stick with the direct sampling at carrier. It may be more challenging, but it is also more up-to-date as this is the direction that SDR designs are moving to. Sure, you are going to have to include an AGC, but that is easy, so put that in. Its much easier doing that than building a superhet front end.

If it is still early in your project, you might want to consider using part or all of the "DSP-10" radio design project that is a popular SDR for 144MHz reception that amateurs have been using for a few years. There are many sites on the web with info about DSP-10 and some kits available too.