Dielectric Absorption in Capacitors! Could it be the answer?

[mikebse2a3]

I thought I would post this information for those who might be interested since I haven't really seen it talked about much when questions about what is the best capacitor to use are asked.

I believe one of the important answers might be in the quality of the dielectric as it relates to the dielectric absorption qualities of different materials used in capacitors. This seems to correlate with what I perceive about different types of capacitors. If a capacitor's dielectric is bad about storing and releasing energy in time well after the orginal musical signal has passed then that residual energy will interfere or mask later musical signals passing through the capacitor. To me this is audible as a lack of clarity especially at higher frequencies.

I also seem to notice the most differences in interconnects and speakerwires at the higher audible frequencies in the form of better or worse clarity which is due I believe to many factors including dielectric absorption. Again I believe this lack of clarity could be due to the musical signal being smeared and distorted in time due to the construction and materials used in the wire.

I would also like to mention one of the largest problems we face when it comes to Clarity(AT ALL AUDIBLE FREQUENCIES) is the Distortation of the Orginal Musical Signal due again to the interferance by the rooms Time Delayed Reflections causing a masking and distortion of later musical signal events. Also as I mentioned in another thread when clarity is improved I also seem to notice an improvement in dynamics.

I know some will disagree especially about wire but simple measurements of resistance, capacitance and inductance definitly doesn't tell us all we need to know.

mike[]

[mikebse2a3]

More Capacitor Info

mike

[Deang]

I've brought up the differences in charge and discharge rates often. I think it's a factor, along with the quality of the film used and what happens to it as it's wound under high tension. Metallized types are presented in a positive light because of their "self-healing" attribute, but we should wonder what kind of damage is it healing itself from and how is it doing it. I've posted on this so many times I'm tired of typing it out.

[djk]

03-27-2004, 4:09 AM

djk




Joined on 09-01-2000
Posts 1,562

Re: "Conditioning", and diminishing returns with capacitors

"Al believes the biggest jump in performance comes from going to the polypropylene dielectic, primarily due to the drop in ESR. Certainly going from a polyester type to a polypropylene dielectric drops the resistance a little, but it's like I said to Al -- I don't think it's enough to account for the large increase in performance I'm hearing when going to these capacitors. I also think a significant drop in resistance would have a major impact on the the sonic signature of the speaker -- but that's not what I'm hearing.

What I'm hearing is a "cleanliness" to the output that I don't hear with the stock stuff."

As I have said before, the differences in capacitors comes down to two things:

lead attachment

dielectric absorption

Think about what happens when the dielectric absorbs some charge on a positive going signal, now the signal swings through zero and goes negative. At zero there should be dead silence, no signal. The charge absorbed by the dielectric now discharges. The discharge is very nonlinear, random, it sounds like noise/hash. Since there is no signal to cover it up, it stands out like a sore thumb.

The better the dielectric, the 'cleaner' the cap sounds.

---

12-30-2004, 2:47 AM

djk




Joined on 09-01-2000
Posts 1,562

Re: what makes a capacitor sound better?

What makes a cap sound good?

Low dielectric absorption (polypropylene)
Non-ferrous leads (copper or silver)
Soldered or welded leads (most cheap caps just mash the leads into the end with some tin paste)
Foil and film (cheaper caps are metalized film)

Old paper in oil caps sounded good as they tended to have foil construction and soldered or welded leads. Jensen is the only current company making them like this AFAIK.

Current oil caps from GE and ASC are metalized film with less than the best lead attachments and as a result do not sound the same as the old stuff.

Klipsch used WWII surplus PIO until stocks ran out in the early 70s, switched to mashed end film, and then back to custom made PIO in the later 70s. About 1985 they went to film again, using some better quality parts this time around.

09-10-2005, 8:37 PM

djk




Joined on 09-01-2000
Posts 1,562

Re: Capacitors - Bypassing vs. Film & Foil in Crossover

"but they are polyproplyne, not film & foil types"

A foil cap has to have a dielectric layer, usually film, usually polypropylene. Most film & foil types are therefore polypropylene.

The difference between a 'regular' polypropylene and a 'film & foil' polypropylene is the metal sheet in the F&F type. A regular cap has a metalized film instead (like those mirrored balloons you see).

The big difference in sound comes from the way the leads are attached to the ends of the caps. With a F&F type they can be soldered, with a regular type cap they are just pressed into the ends with some tin paste. This makes a difference with a lot of current flowing, like in a speaker crossover.

11-09-2004, 11:52 PM

djk




Joined on 09-01-2000
Posts 1,562

Re: Interconnects - Twisted pair vs. coaxial

I find Mark's comments to be interesting.

Most Teflon wire available surplus is silver plated stranded copper. This does seem to be 'brighter' sounding than solid silver or copper wire.

Solid does seem to sound better than stranded (excepting Litz construction).

The cheap no metal RS plugs seem to sound better than big heavy gold plated jobs.

I've used the three twisted wire (no shield or braid) construction with the unshielded plastic handle RS plugs less than 1/4 mile from a 50KW AM/FM/TV station with no RF problems.

A friend of mine bought a 10M pair of Audio Quest Diamond ($5K).

I made a 10M pair of teflon twisted pair (overall braid shield) with the RS ends, and added a 2.2K termination resistor to the amplifier side. Cost me about $30 total.

My friend sold the Audio Quest and sent me a check for $200.

[Al Klappenberger]

Mike,

I did a little experiment in the computer by modifying the "super AA" tweeter filter to ad .1 Uf and 500 Ohms in parallel with each capacitor. The analysis shows no difference. The experiment really ins't worth much though since I have no idea what the actual values of the extra R and C would be in reality. If you ever find more specify info on that I'd like to do it right sometime.

AL K.

[Erik Mandaville]

I think what matters is what an individual 'likes' or 'dislikes' about the perceived sound a given type of capacitor imparts. As Dean indicated, this has been discussed for years, and many of us have contributed over and over again. What one reviewer or author publishes about what physical characteristics and/or types of capacitor construction (including dielectric material, foil composition, etc.) in terms of sounding 'good' are ultimately completely subjective! One builder (including me) has had outstanding results with motor run oil capacitors in both tube equipment and in loudspeaker crossovers. Others have quickly and strongly dismissed them as being inferior to film and foil types costing much more. IMO, just because one kind of capacitor may have lower ESR specs does not necessarily mean that the resulting sound will be mutually agreed upon by everyone as BEST.

It's like saying in an absolute sense that filet' mignon, because of its very fine texture and tenderness, is THE absolute BEST cut of beef money can buy. It also happens to be among the most expensive. What can I say, but as an admitted carnivore, I much prefer the flavor of flank steak. The filet is among the elite cuts, and the flank often isn't. What does it matter, though, when the deciding factor should (hopefully) be one's personal preference without regard to cost, texture of meat, dielectric absorption, self-healing properties (or lack thereof), silver or copper leads, stranded or solid, high power or low, SET or PP, etc., etc., etc. the unfortunate thing, I think, is when someone may be reluctant to overtly state his or her preference for the 'cheaper' or less 'glorified' product simply out of fear of criticism......or in fact rejection. Who has the right to reject another person's right to an opinion???

How about this?! Instead, lets be thankful for what we have, and appreciate the fact that we have as many to choices as we do from which to choose.

Erik

[Deang]

Some stuff sounds grainy and etched out to me and it comes off like distortion. I don't care if it's flank or filet -- if it's tainted it tastes bad. Ha! When in doubt throw it out!

[Erik Mandaville]

Mike:

You said, "I know some will disagree especially about wire but simple measurements of resistance, capacitance and inductance definitly doesn't tell us all we need to know."

I feel very much the same. Specifications may have nothing whatsoever to do with what different people may prefer. Specifications may shed some light on what people hear in terms of perceiving similar characteristics of whatever it is that's being listened to; however, hearing the same thing and forming personal value judgements concerning the quality of that sound are IME two completely different factors. It's like sharing Brand X (great band, BTW!) of high sugar content candy with 10 different people. The chances that all people involved in the test will agree that the candy is SWEET are probably very high. However, the fact that all test subjects detect sweetness does not necessarily mean that they will all like the candy equally well.

"What?! you don't like this candy? Do you, with your obvious lack of experience, have any idea of how much this stuff costs to make? The sugar is from cane grown in the far reaches of Borneo, and cost a considerable fortune to produce. The flavor is not only derived from vanilla beans, it comes from vanilla plants grown in soil collected in Northern India, and watered by a Perrier-carrying aqueduct built by the ancient Romans.

The Emporer's New Clothes are also smashing!

Erik

[Erik Mandaville]

"Some stuff sounds grainy and etched out to me and it comes off like distortion. I don't care if it's flank or filet -- if it's tainted it tastes bad. Ha! When in doubt throw it out!"

Sure, that's partly my point. The other element has to do with the word 'tainted' (remember, Dean, I was elected as President of the Anal Analyzer Society -- I still laugh when I think about that![])

What's 'tainted' or unsavory to you, or me, or anyone should probably be avoided. My thinking here is that our perception of what tastes or sounds good or bad shouldn't be imposed on anyone other than ourselves. IMO, this applies to amplification power, solid state or not, oil or film-and-foil, fully horn-loaded or not, loud SPLs or less loud ones, classical music or non-classical music, electric or acoustic. The list is too long to continue.

....I'm supposed to be out shopping for Thanksgiving groceries, so I'd better go before I get in trouble!

Erik

[Deang]

And so many people much prefer the ride of a Chevy to a Cadillac.

[Erik Mandaville]

Cadillacs make me motion sick. I'll never have one, but lots of people like them because of the soft plush ride. The reputation also seems to add to the cost, but my feeling is that there are many less expensive cars that are built just as well (if not much better) at a fraction of the cost. If reputation is a deciding factor for someone, that's that person's business.

I received a Music Direct catalog yesterday in the mail. Inside were a pair of 6ft long interconnects that cost $8,900. Each additional foot was another $1,300 (something like that). Why?

Erik

[Spkrdctr]

I was actually thinking about the capacitor issue until wires were mentioned........

You are just baiting me.[] I will not take the bait on the wire issue this time....[]

[BEC]

Of course we can just make things easy and measure the ESR. That will take into account all these losses.

"ESR is a performance characteristic representing the energy losses of the equivalent series resistance of a capacitor. It takes into account all of the losses lead resistance, termination resistance and plate resistance, as well as the inherent losses of the dielectric material.

Where:

DFTotal=Dissipation Factor

RIR=Losses resulting from leakage currents and resistance (IR) in the capacitor

RS=Series losses due to leads, terminations, plates, etc. (metal losses)

DFDie=Inherent losses in the dielectric caused by molecular polarization dielectric absorption factors of the particular material."

Bob Crites

[mikebse2a3]

Of course we can just make things easy and measure the ESR. That will take into account all these losses. "ESR is a performance characteristic representing the energy losses of the equivalent series resistance of a capacitor. It takes into account all of the losses lead resistance, termination resistance and plate resistance, as well as the inherent losses of the dielectric material. Where: DFTotal=Dissipation Factor RIR=Losses resulting from leakage currents and resistance (IR) in the capacitor RS=Series losses due to leads, terminations, plates, etc. (metal losses) DFDie=Inherent losses in the dielectric caused by molecular polarization dielectric absorption factors of the particular material." Bob Crites

Bob seems to me when you measure the ESR of a capacitor (as is normally displayed on a Meter) you are only seeing one aspect of that capacitor's quality. What you don't see is what is happening to the signal over Time as the Dielectric Absorption(Dielectric Memory) is imposing itself at a later time on the signal that is passing through the Capacitor. Its like making a standard Frequency Response Measurement on a Loudspeaker that doesn't take into account the Time Factor of that measurement. To get a complete picture of the frequency response you need to show frequency response based on a Time Scale to see a more complete picture of what is actually happening to a signal in time.

You know Bob you made me think of something. If we had a perfect capacitor that showed Zero ESR then I think yes there would be no need to measure any other factor(excepting capacitance value) because to pass the signal completely with no deterioation everything would have to be perfect(ie: no delay/interferance to the signal as it passes through the capacitor). Its like the Loudspeaker example in If we measured a loudspeaker with a perfect frequency response then all the other variables(ie: Energy Storage, Disapation and Timing Errors) could not exist. Its kind of reminds me of reading something PWK said about if he could produce a 100% Efficient Speaker. I think if a speaker was 100% efficient all things(ie:amplitude, frequency & phase would have to be correct.

Eric and Dean and Others I'm sorry if this is a subject your tired of and I sure don't want to cause any arguing because everyone has a wright to their opinion and should choose to spend their money on whatever seems to be of value to them whether its Speakers, Amplifiers, Wire or Whatever. I just believe their is a reason why I believe Capacitors, Wires and Other Things make a difference in sound and eventually measurements that corelate with how we hear will be found that will show why we have a perception of these differences in sound.

Happy Thanksgiving Everyone!

mike[]

[Al Klappenberger]

Mike,

I tend to agree. I don't think ESR or "Q" will measure dielectric absorption. I would not know how to measure it other than the charge and short scheme while looking at it with an oscilloscope capable of transient capture. I do have one of those, by the way!

AL K.

[BEC]

Mike,

That is just the thing I am talking about. ESR is really the sum of all the losses. All of the good polypropylene caps have ESR so low that the total of all the losses can be considered insignificant. It is in the hundredths of an ohm or thousanths of an ohm. So they are perfect capacitors (depending on where you put the decimal point)

Bob

[efzauner]

Here is a good explanation of Dielectric absorption and some models. Should be easy enough to put into a xover simulation to see what kind of output distortion (linear of course!) it causes!

http://www.national.com/rap/Application/0,1570,28,00.html

Note that it really only applies to caps charged with DC for a long time. I dont think it will affect ac very much. The additional caps and resistors models seem very small. I may try to simulate it.

Bob? Any comments?

[BEC]

I took a look at the link and believe the guy testing there is seeing the same thing most people have settled on. Polypropylene caps were much better than others like mylar. I have a new 2.2 mF mylar cap here at my desk right now. It measures 2.316 ohms ESR. A good Polypropylene cap typically measures in the hundredths of an ohm on ESR. I believe that an ESR measurement picks out the bad ones easily when talking about dielectric absorbtion.

Bob Crites

[Al Klappenberger]

efzauner,

That paper shows some exact numbers. That's usefull. Judging by the values give I am quite sure that total circuit in place of each cap in a crossover will make no difference at all in the computer sumulated response. As to if it's audable or not, I doubt it, but theres no way to know until you listen it somehow comparing tow caps in an A-B test. But what two caps and what makes you believe the difference between the two that you MIGHT hear is due to dielectric absorption?

Bob,

If you think about how an LCR meter measures ESR I think you will conclude that it does not include dielectric abosrbtion as part of it. All it does is compare the amplitued and phase between voltage a current through the cap. It does that at several fixed frequencies, 100,120, 1000, Etc. How can it possibley see something that happens so slowly and at zero level? Also with an impedance bridge, you have virtually the same thing. You are equalizing an unknown impedance in one leg of the bridge with known impedance in another leg. Again, nothing but one frequency.

Al K.

[BEC]

Al,

I think ESR measurement at least "infers" dielectric absorption. At least in my testing, those with poor dielectric absorption also test high in ESR. Also what I posted earlier was a definition of those parameters included in the ESR measurement. One of those is dielectric absorption.

Bob Crites