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

"Arny Krueger" wrote in news:UridncxI4vbh4VvZnZ2dnUVZ snipped-for-privacy@comcast.com:

command=text&

Tell me about it. The McCurdy consoles used an arrangement with several cells and an incandescent bulb inside a 9-pin tube envelope (like a 12AX7) painted black. Clockwise rotation of the potentiometer driving the things always gave an increase of level at the monitor. That's all they were supposed to control, though, so you could say that it worked.

Why are you replying to yourself ? It was you who introduced that topic !

dc. Try

higher

Eh ? Zipper noise *is* gain steps. It's different to the rotational noise from a pot though.

Graham

Of

those

direction.

More accurately 'normal' noise goes positive and negative randomly, but they are similar in some ways. A single step doesn't really count as noise though.

I suspect you're trying to obfuscate your way out of the corner you painted yourself into ! lol !

Graham

Perfectly true. I've used CdS cells in a crossfader though ( minimum effect at unity gain and hence very linear at that setting ) . The only nice, suitably priced VCAs were on 1 yr lead time !

Graham

Of

mention

those

direction.

ones

yourself

I have no emotional investment in this, other than to understand how a pot can generate noise, and why a cermet might be worse than a plastic pot. So far, not much illumination.

John

! Of

mention

those

direction.

ones

are

yourself

Neither have I. I just like to understand the mechanisms at work.

You're having trouble understanding that when a pot is moved, it moves in 'micro jumps' ? Those jump sizes vary with the materials used.

Graham

Sorry! Having realised that your expression was dimensionless, I really should have cottoned on you'd normalised with respect to the value at f->0.

Cheers,

rotated ! Of

mention

those

direction.

seems.

ones

are

yourself

'micro jumps' ?

Of course I understand that; I'm an engineer. And I understand that the gain of any signal present will be tweaked in a (probably) monotonic but not necessarily absolutely smooth manner. What I don't understand, or maybe appreciate, is why this should produce audible noise. The mechanisms are, probably, gross mechanical static friction and some sort of micro-granularity in the resistance track itself. Both should be in the 0.1% range, maybe a lot less. And the 0.1% or whatever noise is relative to the existing sound level, not absolute voltage on its own. Of course, a bunch of tube or opamp bias current dumped into a flakey wiper will generate real noise, a bad part compounded by bad design.

Really, there are so many myths around audio, it's rational to question anything that doesn't seem to have a sound physical foundation.

I think people prefer cp pots because they feel better.

John

rotated ! Of

mention

that those

direction.

seems.

cheaper ones

they are

painted yourself

'micro jumps' ?

I think you'll find it's quite non-monotonic actually.

Yes indeed.

That of course is yet another problem. If the wiper goes hi-Z and there's a DC path there'll be significant trouble ! That's why top end gear usually capacitively couples the wiper circuit.

I do this too.

That's yet another advantage. Their cost is something else though, although Vishay-Sfernice have some nice parts that aren't too shocking.

Graham

I've had a couple of Vactrols to evaluate in the past but the were pricier than the Silonex parts I used eventually. I've also used the cheap 'n cheerful 'Eagle' parts from China that are just an led and a little CdS cell in a black plastic box. Perfectly fine actually but I have had to get some graded for use. Actually the Silonex parts are graded too but you can buy them like that.

The very well known and popular UREI 'levelling amplifiers' used vactrols IIRC incluiding a neon one in an early toob based version.

Graham

--- Come on, John, get real.

You've got a huge emotional investment in this thread since you seem to want to be one of the gurus here, but you stated, wrongly, that a step change out of a voltage divider wouldn't generate noise and you've been proven wrong.

Hey, it happens to the best of us.

Acknowledge your error and get on with your life.

Or spend the rest of your life covering up

-- John Fields Professional Circuit Designer

No, really, it's mildly interesting, but since I don't design audio it doesn't matter much. I do use cermet trimpots and have never been bothered by their noise during adjustment. If it mattered, I'd do an experiment and cut through all this expert opinion.

Innocent again. I learn stuff here and sometimes have fun, and might even find an occasional friend, consultant, or employee, but am not interested in guru-dom. I find that sort of thing embarassing.

I do admit I sometimes like to mess with the heads of audio types, such as those heads are. I mean, they do have their own newsgroups where they can express subjective opinions unemcumbered by physics.

Gosh, has anything been proven? I said a pot wouldn't generate voltage noise on its own, and that it could slightly modulate the amplitude of any signal present, and questioned whether it was audible or whether cermet was worse than cp. As far as I'm concerned, the audibility remains an open question. And I did point out the the function of a volume control is to modulate amplitude; I sure hope that's not still being debated.

What, carry my shame all the way to the grave?

Listen: this is a newsgroup. It doesn't matter. Good things happened today that do matter.

John

I did a quick hack with available stuff: power supply; grossly excessive RCRC filter; 100K+100K resistive divider, using various resistor types; Tek AM502 differential amplifier module (gain 10K) fronting a Fluke 8842A trms meter.

With 1k-10k bandpass settings (9 KHz net), the Johnson noise should be around 2.8 uV RMS. The Tek+Fluke setup measures about 8.1 uV with the power supply off, pure Johnson, and about 7.2 dead shorted. So my instrument setup can't usefully resolve the Johnson noise.

But on the other hand, I still see 8.1 uv with the power supply at 10 volts, using either RN55 0.1% metal film resistors or cheap 5% carbon films, when full shot noise should be something like 20 uV.

So it looks like carbon films don't have anything like full shot noise.

More later, maybe.

John

They still do. The incandescent is slower than the LDR, but the LED and neons are faster. The LED has a pretty sharp knee, the neons have a sharper knee, and the incandescent hardly has any knee at all. All good tools for different jobs.

--scott

When you say 'knee' here Scott, which 2 parameters are you comparing ?

Graham

Hi John,

Since Graham has forced us to this dance, may I say you seem to have your priorities straight,

Are you smarter than Graham, or simply older? Both?

Please, if you are younger and dumber than G, keep it to yourself.

If you can do that, you are smarter than you know who ;)

Happy Ears! Al

John Lark> So it looks like carbon films don't have anything like full shot

I just have more test equipment.

John

But not up to measuring the noise from 50k ohms though !

Graham