"Don Pearce" wrote
On Sat, 01 Jan 2005 16:08:29 GMT, "Fleetie"
wrote:

Well I presume that the closest approach to the ideal capacitor is
a pair of parallel plates in a vacuum! A bit bulky for anything more
than nanofarads though!

Hey, what limits the voltage on a vacuum-plate capacitor? I should totally
know this, but above a certain voltage for a given separation, is there
any kind of breakdown?


Martin

I have a graph...

http://www.donepearce.plus.com/odds/DielectricGraph.gif

Thanks, but this doesn't mean much to me. What is "dielectric strength"?
And presumably that is per (some unit of length), is it?

I don't understand the meaning of the legend in the middle of the graph that
seems to say "olts DC / MII". Uh? I assume it should say "Volts", but what is
"MII"?

Does the (hard) vacuum break down when the electric field exceeds
X volts/metre? When that happens, what, well, happens? I can't imagine there
is a visible spark, because in the ideal case, there is no gas to glow.

Excuse my ignorance, but at least I acknowledge it!


Martin
--
M.A.Poyser Tel.: 07967 110890
Manchester, U.K. http://www.fleetie.demon.co.uk

On Sat, 01 Jan 2005 16:45:57 GMT, "Fleetie"
wrote:

"Don Pearce" wrote
On Sat, 01 Jan 2005 16:08:29 GMT, "Fleetie"
wrote:

Well I presume that the closest approach to the ideal capacitor is
a pair of parallel plates in a vacuum! A bit bulky for anything more
than nanofarads though!

Hey, what limits the voltage on a vacuum-plate capacitor? I should totally
know this, but above a certain voltage for a given separation, is there
any kind of breakdown?


Martin

I have a graph...

http://www.donepearce.plus.com/odds/DielectricGraph.gif

Thanks, but this doesn't mean much to me. What is "dielectric strength"?
And presumably that is per (some unit of length), is it?

I don't understand the meaning of the legend in the middle of the graph that
seems to say "olts DC / MII". Uh? I assume it should say "Volts", but what is
"MII"?

Does the (hard) vacuum break down when the electric field exceeds
X volts/metre? When that happens, what, well, happens? I can't imagine there
is a visible spark, because in the ideal case, there is no gas to glow.

Excuse my ignorance, but at least I acknowledge it!


Martin

Sorry - not a very good graph. The vertical scale is volts, assuming a
separation of 1/1000" (1 mil). That legend in the middle should read
Volts/mil. At normal pressure, the strength of air is about 3 million
volts per metre.

When you are below the glow region, where the air ionises visibly
there will be sparks rather that a quiet discharge. Above the glow
region, the air will ionise invisibly, then flash over when the volts
are high enough.

In a vacuum, breakdown occurs when electrons are torn from the surface
of the electrode.

1 Torr is a pressure of 1mm of mercury, or 1/760 of an atmosphere.

d

Pearce Consulting
http://www.pearce.uk.com

"Don Pearce" wrote
Sorry - not a very good graph. The vertical scale is volts, assuming a
separation of 1/1000" (1 mil). That legend in the middle should read
Volts/mil. At normal pressure, the strength of air is about 3 million
volts per metre.

Yeah, I remember 3kV/mm.

In a vacuum, breakdown occurs when electrons are torn from the surface
of the electrode.

Ok so I was asking what the field has to be for that to happen. Or does it
rather depend on the metal constituting the electrode? Is this anything to
do with "work function" or something? (Sorry didn't do degree level
physics.) Is there a property of an ideal vacuum that specifies the maximum
field strength it can "withstand", and beyond which breakdown (of some kind)
occurs?

Your graph, as far as it goes, suggests maybe about 250V/mil, or about
9843 V/mm. But the graph is NOT anything like asymptotic to any value at the
lowest pressure shown (10^-4 torr), so I don't reckon that figure means much.

1 Torr is a pressure of 1mm of mercury

Yeah, I know. /LittleBritain


Thanks.


Martin
--
M.A.Poyser Tel.: 07967 110890
Manchester, U.K. http://www.fleetie.demon.co.uk

In article , Don Pearce
wrote:


In a vacuum, breakdown occurs when electrons are torn from the surface
of the electrode.

And even if you were able to prevent this, if the field is high enough,
pair creation from 'nothing' would occur a la Dirac. Vacuum is far from
empty... :-)

However I *hope* the above has no relevance to audio equipment. 8-]

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

b) a capacitor is a capacitor, and all caps of the same
value and voltage rating sound the same.

Correct.

Well, only if they are all *perfect* capacitors with no leakage,
inductance, internal resistance, etc etc of course ... ;-)

Al.

In article , Al wrote:
b) a capacitor is a capacitor, and all caps of the same
value and voltage rating sound the same.

Correct.

Well, only if they are all *perfect* capacitors with no leakage,
inductance, internal resistance, etc etc of course ... ;-)

Different capacitors have different parasitic properties as well as
their nominal capacitance.

All of these parasitics, if modelled in the ususal linear fashion, will
modify the frequency response of a circuit. A good designer should
specify circuits and capacitors in such a way that the parasitics have
minimal effects on frequency response.

Out of curiosity I did some simple SPICE models of dielectric absorption
in the capacitance of loudspeaker leads. DA is often quoted as the
underlying reasons for choosing PTFE (Teflon, Tefzel, ...) as a preferred
dielectric in some audio circles.

However, my linear simulations show only tiny, mathematical, variations
in frequency response even for extreme amounts of DA. Certainly nothing
that would normally be considered as audible. Perhaps I should go back
and check them - and also extend them to normal capacitors to see if
the variation is greater in some circumstances (e.g. in high impedance
circuits more typical of inter-stage impedances in valve amplifiers).

Non-linear behaviours in capacitors, to the extent they exist, are
probably more interesting and may be audible. However I have not had
a look at these.

I have only taken a quick look at the non-linearity of resistors and
copper wire. There may be a case for preferring metal film resistors over
carbon for their improved linearity. The non-linearity of copper wire
is detectable in very high power RF circuits. However it is difficult to
distinguish from equally small non-linearities from connector-to-connector
contacts. At levels circa -135 dB (IIRC) WRT some very high level RF
signals I suspect the audio equivalent would be insignificant.

--
John Phillips

In article , John Phillips
wrote:
In article , Al wrote:
b) a capacitor is a capacitor, and all caps of the same value and
voltage rating sound the same.

Correct.

Well, only if they are all *perfect* capacitors with no leakage,
inductance, internal resistance, etc etc of course ... ;-)

Different capacitors have different parasitic properties as well as
their nominal capacitance.

All of these parasitics, if modelled in the ususal linear fashion, will
modify the frequency response of a circuit. A good designer should
specify circuits and capacitors in such a way that the parasitics have
minimal effects on frequency response.

Out of curiosity I did some simple SPICE models of dielectric absorption
in the capacitance of loudspeaker leads. DA is often quoted as the
underlying reasons for choosing PTFE (Teflon, Tefzel, ...) as a
preferred dielectric in some audio circles.

The difficulty here is that some effects do not show up on 'topological'
models as they may involve factors like the amount of EM coupling between
components for a given physical layout, etc. You can SPICE model these, but
to do so you need to know the relevant coupling factors. These aren't
easily determined by making measurements on 'isolated' components. However
see my comments lower down in this response...

However, my linear simulations show only tiny, mathematical, variations
in frequency response even for extreme amounts of DA. Certainly nothing
that would normally be considered as audible.

In general this is also my experience, both with modelling, and with
measurements on built amplifiers where I experimented with various types of
component in the past.

However I have also encountered situations where changing one type of
component to another - of nominally the same value - had effects on things
like stability and interference/rail rejection, and hence sometimes
distortion. Due to the kinds of things I mentioned above. Again, in general
I'd expect this to be small, but it may crop up at a serious level in some
situations.


Perhaps I should go back and check them - and also extend them to
normal capacitors to see if the variation is greater in some
circumstances (e.g. in high impedance circuits more typical of
inter-stage impedances in valve amplifiers).

Non-linear behaviours in capacitors, to the extent they exist, are
probably more interesting and may be audible. However I have not had a
look at these.

Various people have looked at this. I recall series of articles in WW/EW,
for example. And some 'distortion measurements' in HFN by MC and others
published by Doug Self. However as with the above, in general I haven't (so
far) personally found these very convincing as sources of audible changes.
Ditto for microphonic or electro-acoustic effects.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

On Sat, 1 Jan 2005 12:48:37 +0200, "Iain M Churches"
wrote:

Happy New Year to All:-)

It would be interesting to know how many on this
group think/find that:

a) capacitor "upgrade" results in a clearly audible
improvement.

b) a capacitor is a capacitor, and all caps of the same
value and voltage rating sound the same.

By and large - b.

If anyone disagrees, they are invited to post the results of
double-blind listening tests.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

It would be interesting to know how many on this group think/find that:
a) capacitor "upgrade" results in a clearly audible improvement.
b) a capacitor is a capacitor, and all caps of the same value and voltage
rating sound the same.

My own listening experiences in A-B substitutions - alone and with several
others over many years and about 60 capacitor types - leaves me, personally, in
no doubt that the differences are audible in a high resolution transparent
system. My own testing led me to use polypropylene caps in the power supply of
valve amps together with chokes, polypropylene bypasses for cathode resistors,
and in particular Russian ex-mil teflon coupling caps. These last are
absolutely marvellous, and about as transparent as I've ever come across. The
largest values are .1uF, so larger values have to be made of multiples. I love
these caps so much I bought about 200 so as to be future proof. They are still
pretty cheap, but the stock will dry up. I get mine from Ukraine, and
personally slightly prefer the FT-3 type to the more common K72.
You will probably get quite differering responses to your question for various
reasons. The above is mine, and based on very many hours of painstaking A-B
comparisons. If you use small value coupling caps for valve equipment I'd urge
you to listen to teflons somehow just so can believe your own ears rather than
what I or others write.

=== Andy Evans ===
Visit our Website:- http://www.artsandmedia.com
Audio, music and health pages and interesting links.

On 02 Jan 2005 01:16:08 GMT, ohawker (Andy
Evans) wrote:

It would be interesting to know how many on this group think/find that:
a) capacitor "upgrade" results in a clearly audible improvement.
b) a capacitor is a capacitor, and all caps of the same value and voltage
rating sound the same.

My own listening experiences in A-B substitutions - alone and with several
others over many years and about 60 capacitor types - leaves me, personally, in
no doubt that the differences are audible in a high resolution transparent
system. My own testing led me to use polypropylene caps in the power supply of
valve amps together with chokes, polypropylene bypasses for cathode resistors,
and in particular Russian ex-mil teflon coupling caps. These last are
absolutely marvellous, and about as transparent as I've ever come across. The
largest values are .1uF, so larger values have to be made of multiples. I love
these caps so much I bought about 200 so as to be future proof. They are still
pretty cheap, but the stock will dry up. I get mine from Ukraine, and
personally slightly prefer the FT-3 type to the more common K72.
You will probably get quite differering responses to your question for various
reasons. The above is mine, and based on very many hours of painstaking A-B
comparisons. If you use small value coupling caps for valve equipment I'd urge
you to listen to teflons somehow just so can believe your own ears rather than
what I or others write.

=== Andy Evans ===
Visit our Website:- http://www.artsandmedia.com
Audio, music and health pages and interesting links.

But the system you have used for the assessment has been a valve one -
and whatever else that may be it most certainly is not acoustically
transparent. It has a very powerful personal sonic signature, which is
of course what is sought by most valve amp users. If, as you say, you
need an acoustically transparent amplifier to assess capacitors, you
absolutely must go for the best of solid state designs.

d

Pearce Consulting
http://www.pearce.uk.com