Is it possible to build a ~unity gain RF preamp with extremely low noise figure?

[cvanc]

Well, it's not really a preamp, as such, but more of a buffer.  Main goal is NF reduction.

Design ideas:
75 ohm in/out
88-108MHz; filtered beyond that
Very low NF (1dB max)
Unity gain is fine, a small amount of gain is fine, but not too much (8-10dB would be too much)
High headroom to pass hot signals without trouble

Purpose:
Provide a back-of-the-tuner, easily installed improvement in reception, by reducing the NF of the existing tuner (which can be 5 or even 7 dB).

Possible?  Thanks.

[Andy Watson]

75 ohm in/out
Unity gain is fine,

If you keep the same input and output impedance, unity gain is not going to improve your noise figure. The receiver input will still generate the same noise, for the same signal level at the same impedance - plus any additional noise generated in your new buffer.
Assuming you want to keep the same input/output impedance you will need to add some gain to increase the signal level. Also, if possible, reduce the bandwidth.

[T3sl4co1l]

Yes, a piece of wire will do nicely.

...

Well, you can't get a useful noise factor without gain at all, you can only add noise in that case.  So it's certainly not going to be a "buffer".  10dB would be a good figure for a LNA at the beginning of a carefully designed signal path; 15-20dB might be more desirable for a poor one (which I guess is the case here?).

Tim

[Z80]

Hi,  you can't improve the NF of your receiver without performing surgery on it and giving it a better front end, but you can provide it with a signal that has a better S/N ratio.  The tried and trusted method to do this is to use a good outdoor antenna with some gain and directivity pointed at the weak station and away from the strong.  Adding a preamp at the top of the mast right next to the antenna improves the performance of the coax downlead.  If you have a strong local station causing problems, you may also need to add a tunable notch filter to stop it overloading the system.

[cvanc]

(Sighs, looks down at mug full of decaf)
Looks like I picked a bad week to give up caffeine 

Yeah it's totally obvious my idea won't work, now that you've explained it.  All right, with the goal still in mind of improving my system performance, let me ask another question:

Existing tuner has NF of 7dB.  Goal is to reduce this dramatically, by 5 or even 6dB.  What is the minimum gain a preamp would need to accomplish this?  (Assume preamp NF of roughly 1dB.)

[KJDS]

You need to convert from dB to linear gains and put them into the Friis equation

http://www.microwaves101.com/encyclopedias/noise-figure

The maths is fairly straightforwards, it's worth putting it into a spreadsheet and working the numbers yourself to get a better feel for it.

If you're feeling lazy then there are an assortment of noise figure cascade calculators available online.

[rjeberhardt]

There is a reason why FM tuners typically have noise figures of 5 to 7 dB.  In a typical urban environment the background noise level can easily swamp the thermal noise in the receiver so there is nothing to be gained with a better noise figure.  Dynamic range is more important.

If your requirement is for a single low level station your best bet is a high gain, narrow beamwidth, antenna.

Russell.

[Kelbit]

(Sighs, looks down at mug full of decaf)
Looks like I picked a bad week to give up caffeine 

Yeah it's totally obvious my idea won't work, now that you've explained it.  All right, with the goal still in mind of improving my system performance, let me ask another question:

Existing tuner has NF of 7dB.  Goal is to reduce this dramatically, by 5 or even 6dB.  What is the minimum gain a preamp would need to accomplish this?  (Assume preamp NF of roughly 1dB.)

Apply Friis' Equation:

F_total = F1 + (F2-1)/G1 + (F3-1)/(G1*G2) + ...

In your case, you only have a two-stage system, so this simplifies to:

F_total = F1 + (F2-1)/G1

Converting to linear ratio, 1 dB NF is equal to a noise factor of 1.26 and 7 dB is equal to a noise factor of 5.01. Your target NF is 7-5 = 2 dB which works out to a noise factor of 1.58.

We can then substitute:

1.58 = 1.26 + (5.01-1)/G1

Solving for G1 yields a linear gain of 12.5, or approximately 11 dB for an LNA with a noise figure of 1 dB.

Designing an LNA with this kind of performance should be achievable. However, your P1dB will be pretty bad unless you bias your transistor with a *lot* of current. For a good introductory reference on RF design, I recommend "RF Circuit Design" by Chris Bowick.

[retrolefty]

There is a reason why FM tuners typically have noise figures of 5 to 7 dB.  In a typical urban environment the background noise level can easily swamp the thermal noise in the receiver so there is nothing to be gained with a better noise figure.  Dynamic range is more important.

If your requirement is for a single low level station your best bet is a high gain, narrow beamwidth, antenna.

Russell.

 He will get there eventually if he is serious about getting a specific station. Otherwise he will have to research
antenna rotation and all the various directional beam antenna designs. When it comes down to improving the S/N ratio there is no substitute for antenna improvements. Antennas don't get the respect they once did, but they are fundamental.

[Mosaic]

S/N improvement by filtering/lowering the bandwidth (reduced Gaussian noise) and also directly attenuating high signals you don't want (also considered focused non gaussian noise in this case)
http://www.rtl-sdr.com/rtl-sdr-frontend-bandpass-filter/