PDA

View Full Version : Re: Capacitors and Resistors


Scott Dorsey
July 1st 03, 03:03 PM
Rifa Roederstein > wrote:
>controls. I'm missing like two values and then I'm set. Roederstein
>MK3's
>and MK4's throughout the circuit, polypropylene and polystyrene caps
>entirely,
>etc,etc. These preamps I'm building are really something and I'd just
>like to go top of the line all the way. An obsession?, definitely,
>necessary?, probably not. I'm just in way too deep to give up now
>man!:)

The thing is that resistors are all very good today. They are all pretty
nonreactive and they are all pretty close to the theoretical noise floor,
which is something you couldn't have said 25 years back. And they are
available in 1% precision for reasonable price.

I can see the occasional need for exotic resistors here and there in things
like differential amps (where you might need more than 1% matching) or low
noise head amps (where you might want low noise plate resistors), but I
can't see just dropping them willy-nilly into circuits without first doing
the analysis to see what they are going to buy you in each location.

Try the Yageos. You'll be impressed, and they are cheap. Twenty-five years
back they would have been high-priced exotica, but today they are cheap.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

MS
July 1st 03, 06:50 PM
Even though the original post doesn't mention capacitors, the title
does, so I'll chime in with a suggestion on caps...

Concerning sonic degredation when placing electrolytic caps in the
signal path, try selecting your electrolytic by ESR instead of by
cutoff frequency in low level blocking/coupling applications. You
might find yourself very pleased with the results in your ensuing
listening tests. And you'll probably find that your selection by ESR
instead of cutoff also helps address the relatively short lifespan of
electrolytics as well as minimize some of the phase issues in
coupling/blocking apps in general...

Probably the biggest reason electrolytics have such a bad rep for low
level signals is because so many designers opt to size them by cutoff
frequency instead of by ESR, resulting in some crappy sounding gear.
This happens because sizing low level coupling/blocking by ESR is
always a bit more expensive... But it's a heck of alot cheaper than
going poly and it holds it's own sonically with the expensive
alternatives.

ThePaulThomas
July 2nd 03, 12:31 AM
>...try selecting your electrolytic by ESR instead of by
> cutoff frequency in low level blocking/coupling applications...

OK. But how is that done? -Paul

MS
July 2nd 03, 03:07 PM
(ThePaulThomas) wrote in message >...
> >...try selecting your electrolytic by ESR instead of by
> > cutoff frequency in low level blocking/coupling applications...
>
> OK. But how is that done? -Paul


Scott has good suggestions and following up on those ideas, you can
get a general trend by looking at some of Illinois Capacitors product
lines. Their site is good in that it lists the nominal (120Hz) ESR
for all values of a given model in a table, here's an example:


http://www.illcap.com/RadialExtLife.asp?Alum7_Action=Find('SERIESID','10 ')&Alum7_Position=FIL%3AORD%3AABS%3A1KEY%3A10PAR%3A

You can see real quick that small electrolytics pretty much suck, they
have a pile of ESR. And though you don't see this in the tables, that
ESR isn't linear, it reduces as frequency increases but on a curve
particular to that device. The larger caps do the same thing, but if
you have .95 ohm nominal ESR and scale down to point-0-something ohm
over frequency, it's alot different than starting at 1K or so and
scaling down... I ESR sized the blocking caps of an INA103 based pre
and it sounded great.

....Since I'm sharing some sound and tone preservation things I've come
across for low level stuff, another is hookup wire. More specifically
the insulation thereof. I actually found this on the "audiophool"
boards. Those guys are always singing the virtues of silver wire and
one person said they found it wasn't the silver, but it was the teflon
insulation doing the deed. They contended that you could use copper
with teflon insulation and actually get warmer yet "silver clear"
sound than with silver/teflon.

So, I took an Oktava MK319 with some Scott Dorsey mods (and a few of
my own), that sounded alot better than stock but still not great, and
rewired from the capsule to the board (all hand wiring inside). I
used standard Belden 24ga tin clad copper/teflon and 4% silver solder
for the entire signal path to the JFET. All I can say is "you got to
be freaking kidding me that wire insulation can make that kind of
difference" after hearing the results (it's actually a great mic
now)... Of course if it ever will, it's going to be in this
application, there's just not very many electrons carrying the signal
from the capsule to the JFET...

Justin Ulysses Morse
July 2nd 03, 03:31 PM
MS > wrote:

> You can see real quick that small electrolytics pretty much suck, they
> have a pile of ESR. And though you don't see this in the tables, that
> ESR isn't linear, it reduces as frequency increases but on a curve
> particular to that device. The larger caps do the same thing, but if
> you have .95 ohm nominal ESR and scale down to point-0-something ohm
> over frequency, it's alot different than starting at 1K or so and
> scaling down... I ESR sized the blocking caps of an INA103 based pre
> and it sounded great.

I'm not calling this untrue, but I think it's an odd way to think about
it. Of COURSE the ESR at 120Hz is higher in a smaller-value capacitor.
That's what capacitance does! "equivalent series resistance" at a
given frequencyis simply the net impedance (resistance plus capacitive
and inductive reactances) at that frequency. We already know that
capacitors block low frequencies, and that's the same as calling it an
increased (frequency-dependent) equivalent series resistance. So what
you're really saying is "use bigger capacitors" than what the arbitrary
LF cutoff might ought to be (which a lot of people ass-ume should be
20Hz or so). It's no big secret that picking a LF cut-off point a
decade or two below that (by using a cap 10 or 100 times as big) will
improve a bunch of things like LF phase response.


> I used standard Belden 24ga tin clad copper/teflon and 4% silver solder
> for the entire signal path to the JFET. All I can say is "you got to
> be freaking kidding me that wire insulation can make that kind of
> difference" after hearing the results (it's actually a great mic
> now)... Of course if it ever will, it's going to be in this
> application, there's just not very many electrons carrying the signal
> from the capsule to the JFET...

This makes a lot of sense. Teflon is used as a dielectric in very
high-precision capacitors because it is so stable and makes a capacitor
that performs very well, but it also has a very low dielectric constant
which means teflon caps don't have very much capacitance for a given
physical size. If you consider the insulation on your wire to be a
capacitor of fixed size, then using teflon insulation will result in a
lower-value (and higher-quality) capacitor than other common insulator
materials. In ordinary circuits this capacitance would be negligible
in any case, but in a 10,000Mohm circuit this tiny capacitor becomes
meaningful. So, was it the decreased parasitic capacitance that
improved your mike or was it the improved qualities of said
capacitance? In what way did the sound improve? Simply increased
high-end or some kind of improved clarity or lower distortion? Hmm?

ulysses

Scott Dorsey
July 2nd 03, 04:59 PM
MS > wrote:
(ThePaulThomas) wrote in message >...
>> >...try selecting your electrolytic by ESR instead of by
>> > cutoff frequency in low level blocking/coupling applications...
>>
>> OK. But how is that done? -Paul
>
>Scott has good suggestions and following up on those ideas, you can
>get a general trend by looking at some of Illinois Capacitors product
>lines. Their site is good in that it lists the nominal (120Hz) ESR
>for all values of a given model in a table, here's an example:
>
>
>http://www.illcap.com/RadialExtLife.asp?Alum7_Action=Find('SERIESID','10 ')&Alum7_Position=FIL%3AORD%3AABS%3A1KEY%3A10PAR%3A

I tend to pick caps with
1. low ESR
2. good high frequency performance
3. 105'C temperature rating.

>You can see real quick that small electrolytics pretty much suck, they
>have a pile of ESR. And though you don't see this in the tables, that
>ESR isn't linear, it reduces as frequency increases but on a curve
>particular to that device. The larger caps do the same thing, but if
>you have .95 ohm nominal ESR and scale down to point-0-something ohm
>over frequency, it's alot different than starting at 1K or so and
>scaling down... I ESR sized the blocking caps of an INA103 based pre
>and it sounded great.

Yup. So you can sometimes gain a lot by going to larger value caps
(or sometimes higher voltage caps... just make sure peak voltage is
at least half the voltage rating of the cap so it doesn't deform).

But you can also decide to select a tantalum or a film cap for smaller
values in some of these cases.

>...Since I'm sharing some sound and tone preservation things I've come
>across for low level stuff, another is hookup wire. More specifically
>the insulation thereof. I actually found this on the "audiophool"
>boards. Those guys are always singing the virtues of silver wire and
>one person said they found it wasn't the silver, but it was the teflon
>insulation doing the deed. They contended that you could use copper
>with teflon insulation and actually get warmer yet "silver clear"
>sound than with silver/teflon.

I like teflon wire. On very high-Z circuits, I could believe you could
hear a difference because of the improved dielectric quality. But that's
much less of an issue than the fact that teflon wire is very easy to work
with; you can't melt the insulation with an iron so you can get very small
joints built with it. On top of that, it lasts forever and never becomes
brittle or sticky decades later.

>So, I took an Oktava MK319 with some Scott Dorsey mods (and a few of
>my own), that sounded alot better than stock but still not great, and
>rewired from the capsule to the board (all hand wiring inside). I
>used standard Belden 24ga tin clad copper/teflon and 4% silver solder
>for the entire signal path to the JFET. All I can say is "you got to
>be freaking kidding me that wire insulation can make that kind of
>difference" after hearing the results (it's actually a great mic
>now)... Of course if it ever will, it's going to be in this
>application, there's just not very many electrons carrying the signal
>from the capsule to the JFET...

Hand wired a 319? THAT sounds silly to me. Next thing you're going to
be ordering teflon PC boards. (The Teflon stuff is a lot nicer than FP4,
but you don't even want to think what the board houses charge for it).
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

MS
July 2nd 03, 06:53 PM
Justin Ulysses Morse > wrote in message >...

> I'm not calling this untrue, but I think it's an odd way to think about
> it.

I'd say that the nominal ESR of a 1uF electrolytic specs at about 500
ohms or so, and a 1uF poly specs at about .1 ohms. And they sound
quite different. I'm thinking the ESR specs are correlating closely
to the results, since I found that an electrolytic with an ESR close
to .1 (big uF) sounds remarkably similar to a poly 1uF with an ESR of
..1...

It's definitely a different angle at looking at it, and that's why I
mentioned it. It delivered the goods for me on the several occaisions
I've done this, thought it was worth sharing...

> So, was it the decreased parasitic capacitance that
> improved your mike or was it the improved qualities of said
> capacitance? In what way did the sound improve? Simply increased
> high-end or some kind of improved clarity or lower distortion? Hmm?
>
> ulysses

I'm not sure. The properties of Teflon make it the best choice to
"give the signal it's best chance" to remain unaltered in this
impedance environment. My main point is that it has a valid,
understandable basis (the dielectric properties of teflon, if not the
mechanism for improvement over the previous, unknown wire), and more
importantly, it delivered in the listening test. To quantify how the
sound differed before and after is pointless on a board, you have to
hear it for yourself, that's why I didn't do the wine tasting thing
trying to describe it...

Don't get me wrong, my intent is simply to say "these ideas have a
quantifiable basis and they actually delivered the sonics for me. You
might want to try them yourself sometime, if you haven't already".

MS
July 2nd 03, 07:37 PM
Justin Ulysses Morse > wrote in message >...

> I'm not calling this untrue, but I think it's an odd way to think about
> it.

I forgot to add, this idea of sizing by ESR isn't mine, I actually saw
it in one of the cap mfgs white papers on applications. They (maybe
Ill-Cap?) stated ESR as a primary selection criteria for electrolytics
used in coupling/blocking applications...

MS
July 3rd 03, 02:22 AM
> Hand wired a 319? THAT sounds silly to me. Next thing you're going to
> be ordering teflon PC boards. (The Teflon stuff is a lot nicer than FP4,
> but you don't even want to think what the board houses charge for it).
> --scott

Scott - of ALL people, YOU should know that a 319 is handwired between
the capsule and the pc board for the roll-off and pad switches, and
then from that pc board to the main pcb... sheeeeesh!!! It sounds
silly to me that you seem to have forgotten that fact!!!

As for the next thing I'm going to do, it's probably stop wasting my
time posting stuff that actually works and can help people improve
their craft (which I thought was the whole point of this group).
Looks like the egos are a bit inflated for my tastes...

no regards,
skitron

Justin Ulysses Morse
July 3rd 03, 02:25 AM
Okay, I have a refined appreciation for the concept now. I was
thinking in terms of choosing the _capacitance_ of, say, a Panasonic FC
capacitor based on the ESR. Which would, of course, just mean using a
bigger capacitor. But you're also talking about choosing the _type_ of
capacitor by ESR specs, which is of course the right way to do things.
I have a handful of capacitor types I rely on, and I made my choices
based on ESR among other parameters. Now if I know I need a big
electrolytic, I look at size/capacitance needs and choose the lower-ESR
type (which is usually Panasonic FC) if it will in fact give me lower
ESR than the more compact, cheaper capacitance of, say, Panasonic NHG.
Which I think is what you suggested. So cool.

ulysses

In article >, MS
> wrote:

> Justin Ulysses Morse > wrote in message
> >...
>
> > I'm not calling this untrue, but I think it's an odd way to think about
> > it.
>
> I'd say that the nominal ESR of a 1uF electrolytic specs at about 500
> ohms or so, and a 1uF poly specs at about .1 ohms. And they sound
> quite different. I'm thinking the ESR specs are correlating closely
> to the results, since I found that an electrolytic with an ESR close
> to .1 (big uF) sounds remarkably similar to a poly 1uF with an ESR of
> .1...
>
> It's definitely a different angle at looking at it, and that's why I
> mentioned it. It delivered the goods for me on the several occaisions
> I've done this, thought it was worth sharing...
>
> > So, was it the decreased parasitic capacitance that
> > improved your mike or was it the improved qualities of said
> > capacitance? In what way did the sound improve? Simply increased
> > high-end or some kind of improved clarity or lower distortion? Hmm?
> >
> > ulysses
>
> I'm not sure. The properties of Teflon make it the best choice to
> "give the signal it's best chance" to remain unaltered in this
> impedance environment. My main point is that it has a valid,
> understandable basis (the dielectric properties of teflon, if not the
> mechanism for improvement over the previous, unknown wire), and more
> importantly, it delivered in the listening test. To quantify how the
> sound differed before and after is pointless on a board, you have to
> hear it for yourself, that's why I didn't do the wine tasting thing
> trying to describe it...
>
> Don't get me wrong, my intent is simply to say "these ideas have a
> quantifiable basis and they actually delivered the sonics for me. You
> might want to try them yourself sometime, if you haven't already".

Justin Ulysses Morse
July 3rd 03, 03:09 AM
MS > wrote:

> As for the next thing I'm going to do, it's probably stop wasting my
> time posting stuff that actually works and can help people improve
> their craft (which I thought was the whole point of this group).
> Looks like the egos are a bit inflated for my tastes...

Woah, bessy. You need to grow some scabs over those open sores.
Things get pretty salty around here, but we all love a lively
discussion and it's all in good fun and for the betterment of the audio
world. Take a minute to assume (correctly) that the replies you've
gotten were intended much less bitterly than you seem to have thought.
In other words, lighten up. I'll let you know when I'm being rude. Do
I have to go back to using emoticons? That's so late 20th century.

ulysses