Receiver Sensitivity Measurement

[the_falcon]

Hi all,
I am aware of the formula used to find the sensitivits of the receiver which is -174+NF(dB)+10log(BW)(Hz)+SNR(dB).
But my query is , which SNR should be used in this formula? The deifnition of the formula explains the SNR for a minimum detectable signal which I could relate to SNR at the input side but I am having a doubt on it. SO could you guide me in that.

so is it the SNR at the input side or the SNR at the output side..??


Falcon

 

[Manuel Maritescu]

Hi Falcon,

Sensitivity in a receiver is normally taken as the minimum input signal required to produce a specified output signal having a specified signal-to-noise (S/N) ratio ...so the answer is the SNR for a minimum detectable signal is the SNR at the output side.Why that?because the minimum detectable is meassured at the output of a "black box", not at the input.If we bring a signal at the input, and the black box has some noise, then we will see at the output a signal that is equivalent with the minimum signal that we can detect.

 

[the_falcon]

Thanks

But if you have a system A having a SNR of 50 dB and the system B has a SNR of say 30 dB and by considering the NF to be 5 dB, then according to the formula of the sensitivity, it seems like the sensitivity of system B is greater than system A. Why would that be.Could you guide me as where I am going wrong.

Falcon

 

[jiripolivka]

Thanks

But if you have a system A having a SNR of 50 dB and the system B has a SNR of say 30 dB and by considering the NF to be 5 dB, then according to the formula of the sensitivity, it seems like the sensitivity of system B is greater than system A. Why would that be.Could you guide me as where I am going wrong.

Falcon

There is a confusion in terms: you search for "sensitivity" but your formula defines noise power at a detector. I would advise to use S/N ratio as a measure of sensitivity, where N is the noise power calculated by your formula in W, and S is the signal power at the same point, again in W.

To communicate, various modulation types are used, and they require different S/N ratios to operate well. Then there are digital "modulation" types with which the bit-error ratio is defined which is a function of S/N, but more often Eb/No is preferred. Read a good radio communication textbook for more details.

 

[tony_lth]

What system is your target? WCDMA, GSM, Radar, etc?

 

[the_falcon]

Ok people. Thanks for the suggestions. I will refer some literatures
My system is mostly targeting GSM now

Falcon

 

[BigBoss]

Ok people. Thanks for the suggestions. I will refer some literatures
My system is mostly targeting GSM now

Falcon

If your application target is GSM, you should take BER take into account stand for SNR.Because SNR is an analog measure and it can not be used in Digital Modulation systems.
For instance, for GSM there should be a definition ( I don't know how much) for BER let's say 1E-3 and the minimum signal level at the input which gives you this BER level will be your sensitivity.
This is valid for other types of digital schemes such as DECT, etc.

 

[vfone]

The sensitivity formula you posted is referring to the SNR at the input of the detector/demodulator.

For GSM generally is used SNR=9dB

**broken link removed**

 

[jiripolivka]

The sensitivity formula you posted is referring to the SNR at the input of the detector/demodulator.

For GSM generally is used SNR=9dB

**broken link removed**

The referred paper by Lou Frenzel makes things more complicated. I real world, in addition to thermal noise your formula defined, there is interference from adjacent channels, from other interferers, etc. It is not easy to include and understand all effects. Refer to a good textbook, and learn the basics.

 

[vfone]

Unfortunately interferes from adjacent channel doesn't have anything with initial (and basic) sensitivity calculation. Most of the RF systems use the posted sensitivity formula.

After the initial receiver sensitivity is defined, and function of this, other parameters as: adjacent, alternate, co-channel, blocking, etc., can be defined.

 

[tony_lth]

GSM system is easy to find. You can find one and test the sensitivity with a signal generator. As jiripolivka said, there should be many interferences.
I guess there should be some software to simulate the whole GSM system performance, because GSM system are always optimised to get better performance. The optimised way is such as hop-freq etc.