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

Cool. I do NMR gradient-coil drivers. What do you do?

John

You're joking, right? Lasers and photodetectors are the nastiest, noisiest things around; I've been fighting that particular issue all week. A photodiode has full shot noise, too.

Well, maybe a picosecond-speed optical link with KHz FM modulation could get close, at a system bandwidth penalty around 1e9.

John

Well...... I don't see them stocked by the usual suspects.

Graham

Which is what I said !

As for brainless, perhaps you'd care to read a post accurately first before making daft comments. Indeed, why would I have mentioned Boltzmann's constant in this context if I didn't know damn well the issues ?

Graham

mention

Leaving out of the widest used !

Graham

On Sat, 22 Jul 2006 11:48:12 +0100, Eeyore Gave us:

The entire selling point of the sheet was that they are very low noise compared to other resistor manufacturing mediums.

Noise in a resistor occurs because resistance mediums are actually heavily diluted semiconductors. The electrons bang around on their way through. It's a loose lattice.

On Sat, 22 Jul 2006 20:28:55 +0100, Eeyore Gave us:

In the audio industry?

In a current flow in metals, electron-lattice interactions tend to space electrons evenly, so their exit from the resistor is more regular than a random (Poisson) distribution, so current-modulation noise is far below the shot noise level you'd get for uncorrelated conduction. The physically longer the resistor, the better the smoothing effect. This works less well in non-metal conductors, with a few exceptions.

That doesn't affect Johnson noise of course; that only depends on the resistance and the temperature, as required by conservation of energy.

John

Perfectly true but essentially irrelevant in most audio usage. In fact I can think of no problems with using metal film anywhere in audio and CF is fine for the most part too.

Graham

Unquestionably, at least before widespread use of SMT. Actually, that not a subject I've yet looked into in depth. Any ideas what commodity SMT resistors behave like ? I've got reason to suspect some may be inferior noise wise to leaded CF.

Graham

In article , John Larkin wrote: [...]

It is not quite just a "gain modulation" there is also a phase modulation component to it. This phase modulation is less than what you would get from an equivelent amount of signal independant noise.

Eeyore wrote in news: snipped-for-privacy@REMOVETHIS.hotmail.com:

martin

Oh, man, this old chestnut again. Of course you're perfectly right, John, but the effort is doomed because half your audience appears to think that classical thermodynamics is a matter of opinion. Why don't we all vote on it and find out how resistors *really* behave?

Cheers,

Phil Hobbs

That surely can't matter until microwave frequencies, and who needs a ppm-accurate resistance at microwave?

John

Phil,

Do metal-oxide thickfilm resistors have shot noise? Any idea how much, as, say, % of full shot?

John

I don't know. The physics is quite a bit more complicated than in metal, so I can't say from first principles, and I haven't measured it. Most of my circuits are built from a huge supply of 1% metal film resistors I bought in about 1989!

It might be quite interesting to measure.

Cheers,

Phil Hobbs

I'm curious myself. My early low noise designs used metal oxide since that was then the readily available low-noise type.

I can't quote any data as such but ISTR that metal film when it took over that market was understod to be fractionally noisier.

Graham

I'd be very surprised if the excess noise in carbon film and carbon composition resistors had much to do with shot noise - carbon has a positive temperature coefficient of resistance, and any path that starts off carrying more current than its neighbours is going to carry more current still - you don't get active channelling at normal current and voltage levels, but the current distribution across the restive paths is going to change with changing current level.

Thick film resitors, built up from a mixture of glass and metal oxide, can do practically anything, so I suppose shot-noise has to be one of the possibilities. "Excess noise" is probably a better phrase to use, since it doesn't delude the user into thinking that "shot noise"is the only source of excess noie in these devices.