In the context of wireless receiver system design, what is the optimum choice of amplifier placement? Should the received signal be amplified first then filtered or should it be the other way round?

And what are the pros and cons in practical applications?


2 Answers 2


GrapefruitIsAwesome's answer is right in mentioning that Friis noise formula says to put the most amplification possible as far up front as possible to minimize overall noise figure, but I'd like to add that you often have no choice:

If you have another signal out of the frequency band you're interested in, it will still be amplified. Now, if amplifiers were perfectly linear (and used no power for amplification), that's not a problem – either it gets filtered out before or after, same effect.

Sadly, amplifiers are not linear. Meaning that if you have a strong signal out of band, it can saturate your amplifier, meaning that its output is no longer simply a multiple of the input. Any nonlinearity like that is distortion, and even worse, leads to intermodulation, so that you get signals into your band that weren't in your band before.

So, in most practical wireless systems, while you want as much amplification as close to the antenna as possible, you need something to keep the out-of-band (OOB) interference out of your amplifier. Sometimes, having an antenna is selective enough. Rarely, you know your signal of interest is the strongest signal (with enough margin) that your antenna picks up, anyway. Then you don't need a prefilter. But honestly, more often than not, you want some kind of selectivity in front of your first amplifier stage, because what good is a noise figure that's better by the loss of the filter (usually, very low single-digit dB!) if in exchange you get the noise of a hundred bandwidths mixed into your signal of interest, with some unknown, fluctuating intensity.

There's also the power efficiency and price argument. If you can reduce the dynamic range your amplifier needs to be linear over, because you need to worry less about OOB interferers, then that means you can operate the amplifier closer to saturation, which makes it more power efficient. Or, you can choose a cheaper amplifier, and invest the money you've saved into a filter with lower passband loss (thus reducing the loss in overall noise figrue

  • $\begingroup$ Thank you for your detailed response. I get the idea that we would need some sort of pre-filtering to avoid the saturation issues. If that was handled and still the main goal was to get high sensitivity to recover a signal with very low or infact negative SNR at the input and recover it using processing gain, Amplification before filtering would be the way to go? $\endgroup$
    – malik12
    Sep 6, 2022 at 6:05
  • $\begingroup$ Well, that really depends on what kind of signal from what kind of interference environment you need to pick up! But if you open up an active GPS antenna, for example, you'll find relatively lax filtering and an LNA integrated behind the antenna. The "lax" is of course probably an effect of the "consumer-grade"; if the GPS doesn't work with interferers close to the GPS frequencies, that's because there's illegal interference, and cost considerations suggest you simply give up. Really, look at friis noise formulas and figure out how bad it is for your specific filter to be upfront! $\endgroup$ Sep 6, 2022 at 7:19

Placing the amplifier first will result in a lower noise figure for the receiver and hence more sensitivity but the amplifier will be exposed to all of the ingested signals, including undesired interference, potentially resulting in amplifier saturation / compression which will degrade the response of the desired signal.

Placing the filter first will make the system more robust interference outside of the bandwidth of interest at the expense of an increase of noise figure due to the loss of the filter.

See also: Filter before amplification or amplify before filtering

  • $\begingroup$ Thanks for your concise answer, with regard to the link you posted, the top answer says to always amplify before filtering as the filter will attenuate the signal. I am a little confused about this point if he is talking about insertion loss, that would be very small (relatively) as mentioned by @marcus below, but my SNR before amplification would still improve wouldnt it? $\endgroup$
    – malik12
    Sep 6, 2022 at 6:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.