# GPS - Noise Figure Analysis

I am trying to analyze all components in one home-made GPS receiver and I am stuck with figuring out how to determine Noise Figure at the mixer output.

See picture below:

If we are going from the input - ground noise level is at -174dBm and LNA amplifies it with 28dB of gain and adding 0.8dB more, so by now we have: -174+0,8+28.OK, going further - since SAW, COAX and MIXER are attenuating signals they are also attenuating the noise therefore their sign is negative(-174+0,8+28-1,5-3,9-6). What is confusing me is why after RF amplifier MGA-86563 we are not adding +20(GAIN)+2(NF)?

Why is this wrong: -174+0,8+28-1,5-3,9+20+2-6+10*log(BANDWITH)

What is wrong in my analysis? Did I misunderstood what NF is?

EDIT: I forgot to write +28 in the last expression.

• your "why is this wrong" doesn't seem to match the paragraph before it. You have as difference left out the +28 and added a +2 but your text says "why not add +20 and +2"... – PlasmaHH Feb 2 '18 at 12:43
• Where does your first analysis that results in -73 dBm come from? A link to the document please. – Andy aka Feb 2 '18 at 17:14
• Here it is aholme.co.uk/GPS/Main.htm – Bip Feb 2 '18 at 17:14
• You're not cascading the gain and noise figure properly. The Friis equation is the appropriate way to cascade noise figures. For example the noise figure looking into a two stage cascade is the NF_1+N2/(G1-1). The noise figure of the 2nd stage is reduced by the gain if the 1st stage. For this lineup, the noise figure is approximately the noise figure of the 1st LNA. The 28 dB of gain is sufficient to reduce the noise figure of the rest of the system to ~ 0. – rfdave Feb 3 '18 at 3:13