Question about noise voltage notation

However, you have to make sure that you don't loose the information contained in your signal by narrowing too much :)

I'm afraid I don't follow you. Consider a square wave (between 0 and 1) of frequency f_1 amplitude modulating a carrier at f_0. We get frequency components at f_0 +/-n f_1. To keep the "information" of the signal, at least we have to keep a BW>2*f_1. If we don't take the next component (BW If you're willing to time gate the signal (ideally by multiplying it by

Right. This is how spread-spectrum techniques work! However, the required synchronization adds a significant overhead and is far less than trivial.

In essence, it is the same principle although without coherent detection.

Pere

In engineering there are always other constraints, but it's always worth paring it down to ideal cases to start out with, so that you know what the tradeoffs are.

You're assuming a lowpass filter, I'm assuming a bandpass. Apart from needing upper and lower sidebands, it's just like the lowpass case for unipolar pulses. The math involved in thresholding a tone burst gets difficult fast, but simple ad-hoc tricks work pretty well--for instance using a window comparator driving a retriggerable one-shot to make sure you oget exactly one trigger per pulse.

The simple way to do it is to use I and Q, lowpass, compute the modulus (which are both easy to do in analogue, if you don't mind a little bit of scallop loss), and gate that.

It's a little more subtle than that. Try working it out--the SNR _increases_ with increasing deviation, with constant power and a matched receive bandwidth. That only works when you're in the high SNR limit to start with.

Cheers

Phil Hobbs

It's not a 'reminder', it's a description of the behavior _of the resistor_. And your description is incomplete at any rate -- thermal noise depends on temperature, because it's, well, _thermal_. And there are other sources of noise, which may or may not occur in a given resistor.

It's up to you to figure out how the _part_ works in your _system_ -- then, the bandwidth of the _system_ outside of your _part_ become important. If your system bandwidth were different, or of your system behavior couldn't realistically be described by a bandwidth (i.e. if you were to integrate that resistor's voltage), then you have to do the math to figure out what the impact of that resistor's noise is on your system.

So no, I don't agree at all.

e quoted text -

What a bunch of strange answers to such a simple question. It appears as if the OP has never run across the fundamental equation for thermal noise: V^2 =3D 4kTBR K =3D boltzmann=92s constant T =3D temp, usually in kelvin units B =3D bandwidth in hertz R =3D resistance The noise is assumed white and uniform. It is not correlated. The equation makes the bandwidth part pretty self explanatory. You can also easily see what happens when you work into an optimum load. Which btw, is the assumed situation in 50 systems. NF and RF system measurements work with Nout =3D kTBG, but the assumption is that it is a matched system.

gl

I know the formula includes Temperature. It is not my description, the nv/sqrt(Hz) is a common description in data sheets. Obviously, it assumes a room temperature, or thereabout condtion. Anyhow, I appreciate all the comments and I got my confusion cleared up on this matter.

That's what i would like to have said, but you said it better.

?-)