The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> antenna and LNA BW
https://designers-guide.org/forum/YaBB.pl?num=1213623890

Message started by aaron_do on Jun 16^th, 2008, 6:44am

Title: antenna and LNA BW
Post by aaron_do on Jun 16^th, 2008, 6:44am

Hi all,

a typical band-select filter is only matched over the desired band. However, it is normally designed by considering S-parameters. Furthermore, the LNA matching bandwidth depends on the Q of the input matching network. Likewise, and antenna is only 50 ohm over a certain BW. Therefore the filter transmission, T21, should be different from S21 and can be calculated using the S11 of the antenna and LNA.

I was wondering, for a commercial LNA is it customary to have a larger than necessary matching bandwidth in order to "idealize" the band-select filter's T21 over a wider band? In other words, if I have a band-select filter with a BW from 2.4-2.5 GHz and an LNA with a BW from 2.4-2.4 GHz, then the filter's T21 will not equal to its S21 for frequencies outside of 2.4-2.5 GHz. But if i make the LNA match from 2-3 GHz, then the filter's T21 will equal to its S21 from 2-3 GHz.

As a last point, a narrowband LNA has a matching network gain at the input which is frequency dependent. Is it normally taken into consideration when concerning the band-select filter's attenuation profile?

thanks,
Aaron

Title: Re: antenna and LNA BW
Post by loose-electron on Jun 16^th, 2008, 12:21pm

Aaron:

It is toughto generalize on some of the questions. As we all know al lot of designs are case specific.

As for the spectral profile of each of the elements being considered - well thats sort of a system question vs. performance of a particular block.

First order system design is often simply having sufficient BW arouind the frequency center and dealing with additive noise, and approximate spectral shapes (i.e. the -3 db point is far enough out) However, more fussy system designs can take into account a lot more variance and detail.

In my opinion, at the end of the day, the simpler system design with approximate spectral BW, additive noise, linearity, etc, generally gets it done. Process variance or model inaccuracy can make the more detailed analysis a waste of time.

my 2 cents,
Jerry