MDS vs Noise Figure isn't MDS better and more reliable ?

[BarNash]

Hi all
Noise figure of an RF amp is rather theoretical , computational, isn't it better to use the MDS = Minimum Detected Signal method to
get the real situation of an RF or If amp ?

Elico

 

[volker@muehlhaus]

Noise figure is more flexible because you can easily calculate the cascaded noise figure of a system with multiple stages, where each stage has some contribution to noise.

 

[BarNash]

Yes but this is only calculated not necessary the real status of those cascaded stages.
what one wants to know is the whole status off the cascaded system.

Am I right ?

Elico

 

[volker@muehlhaus]

Am I right ?

No, I don't think so. Using NF is a proven method that works well in theory and reality.

 

[BarNash]

Thanks
Elico

 

[zorro]

MDS is a tricky concept. Ir could be useful for some situations, but not for characterizing an amplifier.

Yet the name is wrong. There are people that believe that a signal of a level stronger tan MDS can be detected, and that a signal weaker that MDS can not. This is wrong.

Instead, noise figure or noise temperature say how much the amplifier impairs a signal to noise ratio.
Beyond the name, MDS depends of bandwidth. The same aplifier with different filters would have different MDS values. Noise figure or noise temperature don't.

Regards

Z

 

[BarNash]

But the bottom line at any RF/If amp is the minimum input signal that will be usable in working conditions, isn't that true ?

Elico

 

[BigBoss]

Noise Figure gives you an aspect about Signal to Noise ratio..
Does MDS ?? This is the fundamental difference.

 

[BarNash]

Right, the point is that MDS is what matters in the "real" physical world.
The designer uses transitors and chips for its design , and the real test of his design is the MDS of the real physical prototype.
This how I see it .
I admit that I have to spend some time to learn again the basics of the NF issue , and I intend to do that soon.

Thanks any way for your insights.


Elico

 

[volker@muehlhaus]

Right, the point is that MDS is what matters in the "real" physical world.

You don't need to repeat this again and again. It's not wrong, but there are these issue that "zorro" mentioned in post #6.

You might be right from a system measurement view. But you are wrong from an amplifier design viewpoint.
The MDS value for the exact same amplifier changes if you change bandwith/modulation scheme. That's a nightmare for design because it will create all sorts of possible misunderstanding. The noise figure is a much better measure of the amplifier's noise performance because it's clearly defined and does not depend on other parameters outside the amplifier.

 

[zorro]

I admit that I have to spend some time to learn again the basics of the NF issue , and I intend to do that soon.

OK. I advise to avoid MDS; it can cause a lot of confusion.

The concept of MDS had its origin in the old times of radar as an answer to this question:
Which is the level of the weakest pulse than can be detected (with some reliability) with an envelope detector in presence of noise?

For signals other than unmodulated pulses, the power level of signal than allows reliable detection can be very different.
For example, coded pulses like those used in modern (and not so modern) radars can be detected even when they have a level well below the noise floor.

 

[vfone]

Yes, the origin of using MDS or even the name of Sensitivity to characterize the noise figure (SNRinp/SNRout) of a device is old, back to 40'ties.
At that time was very common to mention: "sensitivity of a mixer" or "sensitivity of an amplifier".
Later when RF systems become more complex, noise figure concept was used to characterize a device, and MDS and Sensitivity to characterize a system.