EMRFD Message Archive 3266

Message Date From Subject
3266 2009-06-30 10:01:34 timshoppa Measuring noise in front-end pre-amps and mixers
My homebrew RX seems to do just fine on 80, 40, 30, and 20M with no pre-amp, just the TUF-1 diode mixer. But it lacks some oomph on 15M and could use a pre-amp. Maybe come wintertime the atmospheric noise will be less on 20M and I'll want a pre-amp there. And who knows, maybe someday 10M will open up wide :-).

A J310 as a common gate pre-amp, like in fig 6.32 or 6.33 in EMRFD, seems ideal. A feedback bipolar amp around a 2N5109 or 2SC1252 seems common too.

Since better is the enemy of good enough, it would be nice to follow the spirit of EMRFD and measure any front ends to compare overall noise performance, and maybe quantify.

To actually measure noise figure... the right thing for me to build is a noise source, like in section 7.11 of EMRFD. But I've never made quantitative noise measurements like this before, and not really sure how much I need to exactly quantify noise (like in getting an actual noise figure) vs just comparing different preamp variations to find the one I like best. Certainly I do not own a calibrated noise source but I'm not sure what I'd learn from a calibrated one over just picking any random zener diode.

But is a noise bridge the right way for me to go? I have other measurements that I already can do on the bench, like MDS (just a signal generator and attenuators and a voltmeter). Where is the line between "get a calibrated noise bridge to measure noise figure" and "measure MDS to get to the best preamp you can build into your RX"?

Tim.
3274 2009-06-30 12:18:01 w4zcb Re: Measuring noise in front-end pre-amps and mixers
Take a look at G8KBB's home page for his noise meter. It doesn't get any better than that.

W4ZCB
----- Original Message -----
3276 2009-06-30 13:26:58 Chris Trask Re: Measuring noise in front-end pre-amps and mixers
>
> To actually measure noise figure... the right thing for me to build is a
noise
> source, like in section 7.11 of EMRFD. But I've never made quantitative
noise
> measurements like this before, and not really sure how much I need to
exactly
> quantify noise (like in getting an actual noise figure) vs just comparing
> different preamp variations to find the one I like best. Certainly I do
not
> own a calibrated noise source but I'm not sure what I'd learn from a
calibrated
> one over just picking any random zener diode.
>
> But is a noise bridge the right way for me to go? I have other
measurements
> that I already can do on the bench, like MDS (just a signal generator and
> attenuators and a voltmeter). Where is the line between "get a calibrated
noise
> bridge to measure noise figure" and "measure MDS to get to the best preamp
you
> can build into your RX"?
>

To properly measure NF, you can make a simple and fairly accurate
measurement using what's known as the "3dB Method". You first of all need a
calibrated noise source where the noise density is known in terms of
Watts/Hz. Yoau also need a 50-ohm load, an accurate variable attenuator, a
fixed 3dB attenuator, and some sort of power measuring instrument such as an
RMS voltmeter or a spectrum analyzer.

You begin by placing the 50-ohm load on the receiver input and then
measuring the noise power at the output. Next, you connect the noise source
and variable attenuator to the input and place the 3dB attenuator somewhere
later in the receiver, such as after the preamp or mixer (the later the
better). You then adjust the variable attenuator until you have the same
power reading as before. Now, the noise power at the receiver input (after
the generator and the variable attenuator) is the same as the equivalent
noise power of the receiver input. There are some caluculations to be made
after that to correct for the NF of the rest of the test setup, which are
discussed here:

http://www.home.earthlink.net/~christrask/Noise Figure 3dB Test.pdf

When measuring really low NF, you sometimes need an additional LNA after
the unit being tested. In some cases, I've had to have as much as 50dB of
additional gain when measuing the NF of LNAs using an HP141T spectrum
analyzer and an RMS voltmeter as the noise power indicator.

One of the best noise generators for HF work is the General Radio
GR-1383, though it only goes out to 20MHz.


Chris

,----------------------. High Performance Mixers and
/ What's all this \ Amplifiers for RF Communications
/ extinct stuff, anyhow? /
\ _______,--------------' Chris Trask / N7ZWY
_