Vishay talk bollocks - was re: Teflon Film Tin Foil capacitors

What's the origin of the noise? If the wiper drives a high impedance, unstable wiper contact resistance will generate no noise.

John

Classical equipartition says that the mean energy per degree of freedom is kT/2. There are simple ways to show that this is flat with frequency, classically speaking. This breaks down where the Planck rolloff sets in, at a frequency of order kT/h. The proportionality constant k/h is about 20 GHz/K, so at room temperature it rolls off somewhere around 6 THz. [There's actually another factor of about 1.26 since the Planck function is (hf/kT)/(exp(hf/kT)-1) instead of the usual one-pole rolloff we're used to in circuits, so the 3 dB frequency is actually closer to 8 THz at room temperature.)

The optical emission isn't flat with frequency, but has a peak frequency proportional to temperature, and a FWHM of about two octaves. The reason for this is that there are a lot more degrees of freedom as you go to higher frequency--each 1-Hz slice corresponds to an annular sphere in k-space, whose volume goes as f**2. Optical spectra are commonly quoted in per-wavelength units, and converting from differential frequency to differential wavelength gets you another factor of f**2, so in per-wavelength terms the number of degrees of freedom goes as lambda**-4.

Cheers,

Phil Hobbs

It just so happens that I am about to get involved in a project that includes the design of low noise sensor electronics in a rather small space. Having never worked with surface mount components before, what would you gents suggest for prototyping equipment? Would the use of flex circuit boards make any difference in your choice?

Many thanks in advance,

Stuart Welwood

The Johnson noise voltage is there with no bias, and the shot noise depends on the current through the part.

John

"Stuart Welwood" wrote in news:bqudnT0XytnNmVjZnZ2dnUVZ snipped-for-privacy@comcast.com:

We've seen good results using boards with rigid sections coupled with flex sections. You might also want to investigate Sealed Chip On Board techniques. I can't say more, but here is one university paper that deals with the subject:.

Perfectly correct, but that would be a current dependence profile, which doesn't seem to be quite what was being described.

Bob didn't specify that it was a frequency profile, you did. All he said was "profile." I think his correct point is that one can use simple techniques to separate them. Additional argument is just being a pain in the ass.

John

You will note that carbon comps are

I think carbon comps are better in handling pulsed power, like in radar or flash tube circuits etc.

Mark

On Mon, 24 Jul 2006 10:09:16 GMT, spud Gave us:

INCONCEIVABLE! ;-]

On 24 Jul 2006 10:41:54 -0400, snipped-for-privacy@panix.com (Scott Dorsey) Gave us:

ELECTRONIC variable resistor.

On Mon, 24 Jul 2006 08:24:37 -0700, John Larkin Gave us:

The surface quality of a first surface mirror is very flat, and very fine grained in the macroscopic view.

The "drag surface" of a variable resistor has to be a couple hundred times rougher than that. That is going to produce "noise" when turned, and even when stopped as the contact surface upon the wiper can and likely will always be different tiny points on the wiper surface. It is rougher even than basic grain oriented polished stainless at a certain finish. (maybe a number 4). Pretty rough is the macro surface quality/texture realm.

On Mon, 24 Jul 2006 11:33:03 -0600, "Stuart Welwood" Gave us:

Flex is expensive. Especially in short run numbers.

But how does that create electrical noise?

John

People have told me a million times not to exaggerate but sometimes I still do. At some point, like the THz range as you suggest, the energy in a photon starts to become too much to be ignored.

On Mon, 24 Jul 2006 20:07:24 -0700, John Larkin Gave us:

While the wiper is still or moving?

While moving, it should be obvious. The locations where conduction is taking place are moving all over the wiper surface, and there are even microscopic arcs taking place If the two mediums were bound together covalently, I'd say no noise, but since oxides and oxygen and dissimilar mediums are concerned (wiper metal over some form of carbon slurry or deposition) I'd say there is room to argue that some noise might occur through those pathways.

Wiper still... Interfaces between mediums and oxides, less noise.

Tastes great, less filling.

Not to me. Noise requires potentials, and resistors don't do that.

Arcs? At audio signal levels?

John

difference

But as connections are varying, Vout / Vin also varies. This creates the noise. Think of the condition with a DC input and it's more immediately obvious.

Graham

Something doesn't stack up there, Phil. It's common to replace the exponential term with the series expansion 1 + hf/kT, which is a good approximation down to a few Kelvin (I assume this is the method you talked about which gives the classical result.)

The series expansion gives you (with your formula) (hf/kT)/(exp(hf/kT)-1) ~= (hf/kT)/(1 + hf/kT -1) = (hf/kT)/(hf/KT) = 1 which is an interesting (dimensionless) result!

Your bracketed numerator should just be (hf) rather than (hf/KT) ... (hf)/(exp(hf/kT)-1) ~= (hf)/(1 + hf/kT -1) = (hf)/(hf/kT) = 1/(1/kT) = kT which is the classical result.

Incidentally, the solution of the full (non-approximated) Planck function for the half-power frequency is in terms of the Lambert (Omega) function, which we've encountered before on this group, though in a different context: f(-3dB) = -1/2*k*T*(2*LambertW(-1,-1/2*exp(-1/2))+1)/h = 7.8 THz at 300 K which stacks up with your figure of 8 THz.

Cheers,

On Mon, 24 Jul 2006 22:51:37 -0700, John Larkin Gave us:

difference

You are not thinking in the microscopic level.

Instead of volts per mil, it would be like millivolts per micron. If oxides keep a "wiper" from making "intimate" contact, there will be tiny little breach events that take place as electron bore through the oxides and the wiper scrapes through. Generally though it would simply be shot noise magnified by passage through a smaller surface area of contact. And end cap has more total surface area of contact than does a pair or a few wiper noses. Sniff, sniff... Hey... that's it! Sniffing Noise!