[sdiy] Resistor War !!!

[Neil Johnson]

Grant,

> > Metal Film = very low noise
> > Carbon Film (most common today)
> > Not so good noise

> So how can two different resistors of the same value produce different
> noise figures? Are we talking about mechanical noise sources such as
> microphonics or vibration sensitivity?

You're both right and, ultimately, wrong.  Sure, a perfect resistor will
only exhibit Johnson noise (yes, all right, you can stop laughing now),
also known as thermal noise, which can be calculated from:

	Vn = SQRT(4kTBR)

where k = Boltzmann's constant
      T = absolute temperature in Kelvin
      R = resistance in Ohms
      B = bandwidth in Hertz

E.g 10k resistor at room temp over 10kHz bandwidth generates thermal noise
of 1.3uV open-circuit.

And yes, this figure is independent of what sort of resistor it is.  
Great.

But there are other forms of noise that are generated by real resistors,
from such things as fluctuations in resistance proportional to the DC
current flowing through the resistor.  This "1/f" noise is not white
noise, but has an approximate 1/f shape, and is due to such things as the
resistive material, end-cap construction, etc.

>From the Bible (Art of Electronics 2ed, pp.431-433) for a decade frequency
band:

	Carbon comp	0.1-3.0uV
	Carbon film	0.05-0.3uV
	Metal film	0.02-0.2uV
	Wire-wound	0.01-0.2uV

So, there you have it -- ideal resistors all make the same amount of
noise, real resistors don't.

Of course, AC performance is another matter -- you wouldn't want to use
wire-wounds at audio frequencies as the inductance would cause too much
grief.

Neil

--
Neil Johnson :: Computer Laboratory :: University of Cambridge ::
  http://www.cl.cam.ac.uk/~nej22           +44 (0) 1223 334 477