Jump to content

Capacitor Myths


Pete B

Recommended Posts

I've said that listening tests are unreliable unless they're controlled, this is a fact that is difficult to convey or teach even to some professionals.

I normally stay out of capacitor, amplifier, and wire discussions on the net so let me make a few comments. Sure caps that are at end of life or completely failed will obviously sound bad, this is not what I'm talking about. And certainly it's possible for a company to make a very poor product, this is not the subject either. I'm talking about competently designed components, but not necessarily expensive.

I wrote in response to Vern's idea that it would be interesting for someone to do a capacitor switch test:

This is an excellent idea, yes it has been done, and the results are interesting. One such test was done among a group of highly opinionated speaker builders. The group had a culture, they had common beliefs, most heard differences in caps. Solens, Daytons, Electros, the boutique caps sounded so much better, they were friends with a designer of some boutique caps, he would advise them on which to use. Many in this group were advanced, they had and used measurement equipment, understood much of the theory. Then they did the test - uh oh. Perhaps I'll tell the rest of the story some time.

I want to add a few comments about what I saw in this experiment/discussion. The gentleman who did this switched capacitor test is a PHD economist with a good understanding of science in general but not in depth in electrical engineering. Yet he knows how to use measurement tools and has done some of the best home brew designs on the net. He wins speaker design contests against some of the other bests out there and their designs would give many commercial designs a run for their money.

It is interesting to watch discussions and see how many members chime in with how the differences between caps are so obvious, blatantly obvious.

This fellow added relays to one of his excellent designs so that he could listen to 3 different caps, an electrolytic, a low cost poly, and a boutique poly. He described when he first tried it that he thought the switching was not working, little if any difference. He then replaced one cap with a very different value and what do you know it *was* working. This is what I've found and it can be an eye opener. What more convincing does one need, it seems like the switch isn't even hooked up.

These builders have meets around the country and they decided to do a comparison test. This was what one of the organizers had to say "The bottom line on the whole day was that nobody did better than flipping a coin."

Here's the story summarized by that organizer:

http://www.members.aol.com/pjay99site/captest2.htm

This is why you see mylars, electrolytics, etc. in some of the most expensive designs out there.

Pete B.

Link to comment
Share on other sites

  • 1 month later...
  • Replies 108
  • Created
  • Last Reply

I'm not trying to bash Marsh here, in fact I thought his early paper with Jung was well done - I've not followed his later work.

Anyway I just thought I'd point this out, not an error but their certainly was prior work that was not referenced. Note figure 7 here:

http://www.capacitors.com/picking_capacitors/consider.htm

It is the way their MutiCap is constructed, you would think it was revolutionary the way it's hyped.

Note figure 2B on page 1 of the images for this 1966 patent:

http://patft.uspto.gov/netacgi/nph-Parser?...9&RS=PN/3287789

Pete B.

Link to comment
Share on other sites

Two thoughts immediately come to mind. First is the size of the parasitic elements compared to the nominal value of the capacitor and their effect when considering frequencies of interest. Audiophiles often talk in generalities but when these parasitic elements are very low at audio frequencies, they would be expected to be of inaudible consequence. Secondly is the fact that they are ALL LINEAR elements in the circuit model. There are no non linear parasitic elements therefore, their effect is to alter the frequency response of the circuit. Since I am a strong believer in using graphic and parametric equalizers to correct the overall system frequency response to achieve the desired system energy transfer function within the passband of interest, these are merely a few more of the many elements to be compensated for simultaneously. Therefore it doesn't matter what their sonic alteration is. It is interesting that most audiophiles are obsessed with the quality of capacitors and inductors in a high level frequency selective circuit ie. a crossover network where control and consistancy is most difficult but are adamant against their use in a low level frequency selective circuit ie. an equalizer where they are most effective, controllable, and predictable in their results. Curious creatures audiophiles.

Link to comment
Share on other sites

  • 4 weeks later...

The reasons that there are no non-linear elements in those models is that they are linear models. There are some non-linear effects in capacitors but it is rare when they have to be modelled.

It is interesting that the diagrams on that page have no labels on the axis, typical of marketing hype. Capacitor performance that is talked about on that page is outside of the audio range, applying to video and RF frequencies. This is important for power supply decoupling, but not devices used in a crossover network.

Pete B.

>Two thoughts immediately come to mind. First is the size of

>the parasitic elements compared to the nominal value of the

>capacitor and their effect when considering frequencies of

>interest. Audiophiles often talk in generalities but when

>these parasitic elements are very low at audio frequencies,

>they would be expected to be of inaudible consequence.

>Secondly is the fact that they are ALL LINEAR elements in the

>circuit model. There are no non linear parasitic elements

>therefore, their effect is to alter the frequency response of

>the circuit. Since I am a strong believer in using graphic

>and parametric equalizers to correct the overall system

>frequency response to achieve the desired system energy

>transfer function within the passband of interest, these are

>merely a few more of the many elements to be compensated for

>simultaneously. Therefore it doesn't matter what their sonic

>alteration is. It is interesting that most audiophiles are

>obsessed with the quality of capacitors and inductors in a

>high level frequency selective circuit ie. a crossover network

>where control and consistancy is most difficult but are

>adamant against their use in a low level frequency selective

>circuit ie. an equalizer where they are most effective,

>controllable, and predictable in their results. Curious

>creatures audiophiles.

Link to comment
Share on other sites

I want to point out that even with the above test I do recommend replacing old electrolytics with low cost film caps whenever possible and reasonable, usually for values less than 150 uF. I do this so that I won't have to worry about a new set of electrolytics going bad in 5 or 10 years.

I've put a lot of thought into cap replacement and considering that it is established that electrolytic caps do not age well, and after measuring several that did not meet their voltage specs and in one case had highly increased capacitance, I now suggest always replacing electrolytic caps that are more than 10 or 20 years old. The exception to be considered on a case by case basis is the large computer grade caps found in some systems, and perhaps some of the more exotic types. But always the electrolytic caps to the mid and tweeters which are more likely to fail due to a leaky cap:

http://www.classicspeakerpages.net/dc/dcbo..._id=7195&page=4

http://www.classicspeakerpages.net/dc/dcbo..._id=&page=#7710

http://www.classicspeakerpages.net/dc/dcbo..._id=&page=#7728

http://www.classicspeakerpages.net/dc/dcbo..._id=&page=#7713

Link to comment
Share on other sites

I've used many different caps over the years, ASC, Dayton, GE, Panasonic, Solen, Sprague Hermetically Sealed surplus, Motor start caps, Bennics, various mylars, various 1% precision caps, and even paper caps in old tube equipment.

I most often use Solens and there have been many comments from others about harshness with these caps. I have not had this experience, ever, and neither has a trusted friend. I would go on using them however, since I have found a lower cost poly I'm going to use these in some upcoming projects, the Axon MP 250V:

http://zalytron.com/caps.html

If I were rebuilding the AR-11 in the most economical way, I would purchase a large number (to get the quantity discount) of the 10 uF GE caps here:

http://madisound.com/gecaps.html

One 10uF GE would replace the 10 uF tweeter cap, and 4 in parallel would replace the 40 uF midrange cap. I would use series resistors to match the ESR of the electrolytics being replaced.

The 120uF woofer cap would be replaced with three 40 uF/100V non-polar electrolytic Bennics caps in parallel. I've used these often for large values and they measure well, even after 10 years.

A low cost upgrade would be the Axon 120 uF poly.

The owner of the AR-11s that I'm working on has asked for all Solens, which is a bit more expensive and I have no problem with this selection as long as series resistors are used to match ESR.

Just wanted to point out some low cost options mainly the Axons and GEs, I do not have a strong preference for Solens.

Pete B.

Link to comment
Share on other sites

Pete,

Thank you for giving me an excuse to set the record straight on my capacitor experience.

I have had it positively suggested to me that my comments might have been interpreted as bad-mouthing a brand. I would like to clear this up, as that was never my intention.

The Solens I was talking about were the 400v version. These were more expensive and "hardier" than the 250v version. My thinking at the time was that the more overkill the better. Many times that simply isn't true. I was possibly mistaken in believing that these would be better than the less expensive 250v version.

It is important that readers understand that what I called "harshness" and complained about at length has cleared itself over time for whatever reason. It may not happen for others, but some time playing seems to have helped mine quite a bit. (Theoretically, this shouldn't happen.)

Perhaps the difference between the 400v and 250v versions is a contributor to why some people swear to this effect and others do not.

My comments about the experienced Solen "harshness" were based on a total sample size of fewer than a dozen capacitors bought all at one time.

So while I was *initially* horrified with *those* capacitors in *that* application, it may not happen next-time. A sample-size of a dozen is a little unfair as a judge of a brand.

Pete, I wondered if you would be good enough to share with us how you would back-into or estimate the series resistence necessary to compensate for the lower ESR of a poly capacitor. I find this task daunting as I can only very roughly estimate the original ESR of the original electrolytic capacitor based on an average of other electrolytic capacitors.

While a tenth of an ohm might seem nitpicking, if the total ESR of the original electrolytic was somewhere between one tenth and probably no more than a third of an ohm, a tenth of an ohm is a major percentage of the total resistence.

Bret

Link to comment
Share on other sites

I had a similar experience with the Solen Caps, and I used the 400 volt series. I also agree that what I heard shouldn't happen, but it did and I've heard a similar change in other speakers that have had new caps installed. My only guess is it has something to do with the dialectric forming over time, or our ears adjusting to the change.

Link to comment
Share on other sites

>I had a similar experience with the Solen Caps, and I used

>the 400 volt series. I also agree that what I heard shouldn't

>happen, but it did and I've heard a similar change in other

>speakers that have had new caps installed. My only guess is it

>has something to do with the dialectric forming over time, or

>our ears adjusting to the change.

Well, here's two posters (Bret and yourself) that have reported new PP caps getting better over time. This is exactly why professional audio reviewers who write for many of the audio periodicals today 'break in' new speakers for many hours before actually assessing the speaker's quality attributes. When I upgrade a crossover for a customer with more modern caps like those you have discussed, one of my bonus services is 16-24 hours of playing a repeating pink noise track thru the speakers at about 90 dB. This is to the customer's benefit as they will not experience that harshness typical of brand new caps.

Have any more viewers of this post experienced that same affect? Please respond pos or neg on this.

There may be no scientific explanation at this time as to why new caps 'get better' over time. Perhaps it would make a good basis for a EE Masters degree student's thesis?

Carl

Carl's Custom Loudspeakers

Link to comment
Share on other sites

>Perhaps it would make a good

>basis for a EE Masters degree student's thesis?

>Carl

>Carl's Custom Loudspeakers

I don't think most degreed electrical engineers take this seriously along with many other audiophile myths which fly in the face of knowlege gained through formal training and experience backed up by scientific research such bench measurements and as double blind tests. I know I don't. Pete's initial posting demonstrates the fact that many people who believe what they think they heard didn't really hear anything more than what they thought they SHOULD be hearing. It would not be hard for either capacitor manufacturers or speaker manufacturers to debunk this one way or the other once and for all, at least for their own products. All they'd have to do is "break in" their product and make measured comparisons or DBTs with identical production units which have not been broken in. One very disturbing aspect of the theory that electronic equipment must be broken in is that it is a frank admission that the product is not stable and will undergo a change which is neither predictible in its progress or its outcome and essentially beyond the manufacturer's control. IMO, except where this is unavoidable such as in the inital use of a new automobile engine where valves and piston rings self machine the last few ten thousands of an inch to seat themselves because such precision would be too expensive at the factory, this is unacceptable. The product should reach the consumer ready to use as intended. The notion of break in is also a direct contradiction of the philosophy behind ISO 9000 which expects that production units match the prototype as closely as is humanly possible which means that if a so called break in is required, it should be performed at the factory and then tested again prior to shipment to the customer.

Much of what is called break in is IMO becoming accostomed to what is initially unfamiliar. If you want to know how good or bad your sound system really performs, don't listen to it for a month or two and go out and listen to lots of live unamplified music to get the sound of the real thing in your head. Then when you go back to the facsimile your equipment creates, you can lament what will become its obvious shortcomings you don't usually hear. The good news is that your ears will probably become re-broken-in to its sound again very quickly.

Link to comment
Share on other sites

>Much of what is called break in is IMO becoming accostomed to what is initially unfamiliar.<

I concede this happens. It's a given. (and one reason A/B tests on unfamiliar equipment fail to prove anything beyond coin-toss odds) In my case, thankfully, I was given the opportunity to have mine "broken-in" (which I think is a horrible misnomer) while I wasn't there.

So I walked in after about a week and it had changed. Substantially. What I'm communicating inadequately is that the effect wasn't an "audiophile subtle impression," I'm talkin' pucker-factor. I'm talking "YEeeeooow, that's awful." I'm talking about distortion so plain that you couldn't hear the cymbal past the noise "shssszzzzzz." I don't mean you couldn't tell what type of cymbal it was, or maybe it wasn't ringing just exactly right - I mean it barely resembled a cymbal. Voice was incredibly awful. It almost sounded like it shifted "up" in pitch. It was awful.

I got to listen elsewhere a few hours a few days spread over many weeks, while my "reference" stayed the same at home.

I think there probably is a perfectly scientific and measurable (if we knew for sure what to measure) "event" going-on here. I suspect John has already told us what this is, we just don't know enough to connect the two things.

I've read the study with the preamplifier. I'm not sure that this is a good model for crossover behavior. And even it if were, would we not assume the preamplifier would be used some before the test was run?

The reason we keep arguing about this is a matter of faith. I have faith that you can walk into a showroom and hear problems in crossover regions and count peaks and troughs in an expensive pair of speakers. But if *I* hear something (blatant), you cannot force yourself to believe it since it can't possibly be true. My hearing or mental health seems to be in question, but not yours.

Just accept that this was at least as obvious to me without graphs and charts as your counting anomalies in really expensive speaker response was without graphs and charts and a labful of equipment. We don't have to disbelieve each other.

Bret

Link to comment
Share on other sites

>I think there probably is a perfectly scientific and

>measurable (if we knew for sure what to measure) "event"

>going-on here. I suspect John has already told us what this

>is, we just don't know enough to connect the two things.

If there were something to measure, believe me the reviewers and manufacturers would know exactly what it is. The manufacturers know a lot more about their products than you'd think including things you'd never think of such as the toxicity of the finish and the antidote for a baby who teethed on one or the smoke and flamability rating of the grillcloth.

>The reason we keep arguing about this is a matter of faith.

Having seen a lifetime of audio scams and other scams and having lived in a couple of the scam capitals of the world New York City and California, I'm fresh out of faith. In my mind, I'm from the "show me" state.

>I

>have faith that you can walk into a showroom and hear problems

>in crossover regions and count peaks and troughs in an

>expensive pair of speakers. But if *I* hear something

>(blatant), you cannot force yourself to believe it since it

>can't possibly be true. My hearing or mental health seems to

>be in question, but not yours.

>

>Just accept that this was at least as obvious to me without

>graphs and charts as your counting anomalies in really

>expensive speaker response was without graphs and charts and a

>labful of equipment. We don't have to disbelieve each other.

By experimenting with equalization and all types of frequency response errors over a period of years, I learned to do what the guy from Revel taught his trained listeners to do, get a feel for what areas of the frequency range typical errors sound like. It helps in adjusting equipment and judging how bad things are. BTW, one problem EVERY speaker system designer/sound system owner faces is the inconsistancy of spectral balance from one recording to another, especially IMO on vinyl phonograph records although there are certainly plenty of them on cds too. Classical music is IMO the most critical for precise accuracy and happens to be the kind of music I like the best by far. With musical instruments in the house, I can get a feel for so much that is lacking. Tsk, tsk.

Link to comment
Share on other sites

>With musical instruments in the house, I can get a feel for so much that is lacking<

Yes, I know what you mean. Violin, clarinet, flute, piano, electric and acoustic guitars here at the moment.

I usually chalk what's missing up to the recording, unless it's Telarc, then I chalk it up to the mixing (Telarc's Saint-Seans Symphony in D is preposterously mixed -> it's fun, but it's wrong).

Just because I listen to YES, Genesis, and the ocassional Whitesnake doesn't mean I don't listen to Andre Previn, Andrea Bocelli, and Heifetz.

BTW - I've concluded why I like electronic music so much. Acoustic instruments sound different depending on the volume at which they are played. Typically, mixing is done such that a cymbal crash may not be as loud as the first violin. If I bring the cymbal to "realistic" volume, the violin is louder than a jet taking-off. You don't have that problem with synthesizers because there is no "reference" to tell you they are wrong.

Take something like, oh. . . Fanfare For the Common Man, Copeland. Typically the trumpets at the beginning are being played very loudly. If you adjust your volume to equal several horns' volume played loudly enough to get that brassy edge, the rest of the piece will simply destroy your speakers. If you turn it down to the volume you would experience in a hall, then there is too much "edge" in the brass for a 20th-row, center, seat at that volume.

Compound that, over and over again. . . you either can hear the "edge" on a cymbal but not the rasp of a bow on a violin, or more-commonly, you can hear the rasp of the bow, but not the edge of the low-brass.

So while I dislike the quality of most "live" recordings taken during a performance with most kinds of music, that is the only kind of symphonic music I can listen-to in order to judge whether my system is being faithful, or not. Of course, that's missing almost all edge from almost everything. Even at that I realize that the skill of the engineer and the hall has much to do with how "real" it sounds on this end.

Much easier to get Bocelli and Brightman, or Alpert, right in the studio.

Canadian Brass isn't bad, nor is chamber music. Typically you can get the SPL in your room about right for the apparent "detail" in the recording. I find that very pleasant.

Deep Purple, Accept, or Dream Theater? Well, that just can't be loud enough.

I have enjoyed several movie soundtracks for the quality of the recording; yet even they are "out-of-balance" for anything you could experience live.

I concede that you can, without doubt, define the problem frequencies more readily than I can; yet I'm not entirely without that skill, myself. So I put-forward the following proposition: I'll believe your story if you'll believe mine.

;-)

Bret

PS - the edit is a change of "you" to "you'll"

Link to comment
Share on other sites

Higher voltage film types are usually "better", however, remember that both the film and the metalization are very thin. We don't want the metalization to get so thin that it raises the ESR as a tradeoff to thicker film for increased breakdown voltage that we don't need.

I believe that most of what I've used have been the 250V version.

I should probably point out that when I see/hear an anomoly, I usually try to put controls on the experiment when I want to learn more about it, usually I just chalk it up to a bad day, unless it repeats in other situations. If something like this is reported to others it is very likely that they might see/hear it through the "power of suggestion". We should keep in mind that most differences wire, caps and more "go away" when the comparison is done with controls. Now here I'm getting back into it, and I really don't want to, I don't enjoy this type of discussion. So take my comments for what they're worth, they're my view and opinion.

I measure the old cap keeping in mind that the ESR might have gone up due to their age, I draw from my experience having seen ESR values for many caps, and I draw from all the historical recorded measurments that give an indication as to the cap ESR values.

By the way, I'd like to see test reports or published input impedance and frequency response tests for the AR-11 if anyone has them.

Keep in mind that there certainly are many cases where ESR is not critical. The tweeter cap feeding the pot, with no inductor, in the AR-3a is not critical (assuming an ESR <<1 ohm) since the ESR is compared to the load (about 4 ohms) and not the losses in the inductor (less than .5 ohms in the AR-11). Note also that the "AR-3a Limited" where the tweeter cap is 4 uF and the inductor .16 mH is a much lower Q circuit, due to the smaller cap and larger inductor, than the AR-11's 10 uF and .105 mH inductor, and would therefore also be much less sensitive to ESR.

It is interesting that a crossover can be designed to be much less sensitive to ESR by having small value series resistors set the Q of the networks. This can be seen, for example, in the AR-44bx schematics where the woofer shunt cap has a 1 ohm resistor in series, the mid shunt cap a 3 ohm, and the tweeter series cap a 1 ohm, much better design practice. It is interesting that when these fixed loses are not in place it is possible for the XO Q to be very sensitive to source impedance - cable and amplifier. Some might call this a revealing speaker, it would in reality be a finicky - poorly designed speaker where the frequency response is highly sensitive to source impedance.

Finaly, I've mentioned values for electrolytic ESR usually below 1 ohm, which is what's usually seen in better quality electros for crossovers, there are electros with much higher ESR.

The Compulytic type are a different breed where there are many attachment points to the plates, this brings the ESR back down to very low, I've seen adds claiming 10 milli-0hm (.01 Ohm) ESR values however this was a highly optimistic "marketing" number. More common values were .05 to .1 ohm. It is near zero in these caps, as it is with films, and therefore some other loss in the circuit (should) sets the Q when they're used, so the exact value is not critical as long as it is low.

Circuits can be designed to be less sensitive to ESR and this is why it is not usually spec'ed in caps.

Pete B.

>Pete,

>

>Thank you for giving me an excuse to set the record straight

>on my capacitor experience.

>

>I have had it positively suggested to me that my comments

>might have been interpreted as bad-mouthing a brand. I would

>like to clear this up, as that was never my intention.

>

>The Solens I was talking about were the 400v version. These

>were more expensive and "hardier" than the 250v version. My

>thinking at the time was that the more overkill the better.

>Many times that simply isn't true. I was possibly mistaken in

>believing that these would be better than the less expensive

>250v version.

>

>It is important that readers understand that what I called

>"harshness" and complained about at length has cleared itself

>over time for whatever reason. It may not happen for others,

>but some time playing seems to have helped mine quite a bit.

>(Theoretically, this shouldn't happen.)

>

>Perhaps the difference between the 400v and 250v versions is a

>contributor to why some people swear to this effect and others

>do not.

>

>My comments about the experienced Solen "harshness" were based

>on a total sample size of fewer than a dozen capacitors bought

>all at one time.

>

>So while I was *initially* horrified with *those* capacitors

>in *that* application, it may not happen next-time. A

>sample-size of a dozen is a little unfair as a judge of a

>brand.

>

>Pete, I wondered if you would be good enough to share with us

>how you would back-into or estimate the series resistence

>necessary to compensate for the lower ESR of a poly capacitor.

> I find this task daunting as I can only very roughly estimate

>the original ESR of the original electrolytic capacitor based

>on an average of other electrolytic capacitors.

>

>While a tenth of an ohm might seem nitpicking, if the total

>ESR of the original electrolytic was somewhere between one

>tenth and probably no more than a third of an ohm, a tenth of

>an ohm is a major percentage of the total resistence.

>

>Bret

Link to comment
Share on other sites

Pete wrote:

>The Compulytic type are a different breed where there are many attachment points to the plates, this brings the ESR back down to very low, I've seen adds claiming 10 milli-0hm (.01 Ohm) ESR values however this was a highly optimistic "marketing" number. More common values were .05 to .1 ohm. It is near zero in these caps, as it is with films, and therefore some other loss in the circuit (should) sets the Q when they're used, so the exact value is not critical as long as it is low.<

Okay, so we're saying that replacing cap ESR in a crossover which was designed with Compulytic caps is unnecessary despite the crossover's being sensitive to ESR.

>Circuits can be designed to be less sensitive to ESR and this is why it is not usually spec'ed in caps.<

Just to be sure I understand the weekend hobbyist's need to understand this: Compulytic capacitors are close enough to poly film capacitors where ESR is concerned that nobody should worry about losing ESR when replacing their Compulytics with poly.

Is that right?

Bret

Link to comment
Share on other sites

Several here ask about how to determine the ESR for the original caps. There is some data here on the site but it is expressed as DF:

http://www.classicspeakerpages.net/dc/dcbo...&mode=full#6176

First we must understand that the name used - Dissipation Factor or DF is equal to tan(delta), see here Figure 3: http://www.capacitors.com/picking_capacitors/consider.htm

DF is the term used below about 1000 Hz more by power supply designers, and tan(delta) at high frequency by RF people. These are just conventions. One might ask why not just spec ESR? DF and tan(delta) are ways of expressing ESR in comparison to the capacitive reactance at the frequency of interest. It is a relative value rather than absolute.

Recall that tangent (remember the t-o-a saying) equals opposite over adjacent. Opposite in that figure is R (really ESR), this is what we want to solve for, adjacent is Xc along the imaginary j axis.

Recall that an ideal cap has no resistive component so R being in the numerator of DF indicates that a zero DF is an ideal cap. We want low DF in power applications so that the cap does not get hot when high currents pass through the cap.

Some speak of Quality Factor (Q or tan(theta)) in capacitors which is simply the ratio of the desired impedance Xc over the non-ideal loss factor R. Thus Q = 1/DF, just two different ways of looking at the same thing. Simply, Q equals what you want over what you don't want. An ideal cap has infinite Q, since R the loss in the denominator goes to zero. RF designers want high Q caps so that the Q of tuning circuits is defined by the resistor values in the circuit and not the brand or type of cap that they use. Often if they keep the design losses low and use a high Q cap they end up with the best selectivity in tuned circuits. This was the old days, now they use 1/Q or tan(delta) rather than Q or tan(theta) as shown in Figure 3.

Xc is easy to calculate from the expression for ideal capacitive reactance:

Xc = 1/j(2*pi*f*C)

where: f is the test frequency

C is the capacitance at the test frequency

then:

R (ESR) = DF * Xc

plugging in:

R (ESR) = DF/(2*pi*f*C)

This could easily be done in, or added to the spread sheet. Note that ESR increases with age, this must be kept in mind when measuring old caps. ESR is also different at different test frequencies, the important value is around the crossover frequency or sometimes in the passband.

I've not checked the math above, another look would be helpful.

Pete B.

Link to comment
Share on other sites

  • 2 months later...

Here is an interesting and good read about capacitors, I do not agree with everything said as there are some errors in my opinion, but there are also some wise points made:

http://sound.westhost.com/articles/capacitors.htm

The author writes:

"Claims have been made that most capacitors must be inductive, because they are made from a wound sandwich of film and foil, or metallised film. Because it is usually wound (in a flat coil), logically, this leads to inductance. The problem with that theory is that it assumes that the termination is made to the foil at the end only, but a quick check of manufacturer data will show that this is generally not the case. The vast majority of capacitors are made so that the foil or metallisation projects from each side (one 'plate' on one side, the other 'plate' on the other side). Each end is then connected so that all sections of the plate are joined together. There is no longer a 'length' associated with the plate, and only its width becomes significant for inductance."

The above is entirely false for common electrolytic capacitors, most do have the lead contact attachment point at one end of the foil contrary to what is stated. In fact the "trick" in the low ESR Panasonics, that everyone liked for so many years, was to move the attachment point about 2/3's of the way down the foil which shortens the longest path from the contact to either end of the foil. I know this is absolutely true because I opened one up and took a look. There are some electrolytic caps of special construction, such as "Computer" sometimes called "Power Supply" grade caps that have many attachment points along the foil to lower the ESR, and ESL. I have called the common end contact electrolytic a transmission line due to the long parallel plates, and high capacitance which will increase the delay time down the network. A capacitor expert is referenced at that site who has come to the same conclusion.

His statement above that "width becomse significant for inductance" is true for cap constructed the way he describes. But actually, for a common electrolytic a wide and shorter, plate is better that a less wide and longer plate, the inductance is lower with increased width in this case.

Pete B.

Link to comment
Share on other sites

  • 9 months later...

I've heard for years that there are many counterfeit transistors and other semiconductors being sold and wondered why those would be the only type of part.

There is discussion and a picture here, of a capacitor:

http://www.diyaudio.com/forums/showthread....0227#post390227

http://www.diyaudio.com/forums/attachment....tamp=1084091465

See here for transistors:

http://sound.westhost.com/fake/counterfeit-p3.htm#mjl21193

Link to comment
Share on other sites

Hi Pete;

Decades ago I remember seeing photos of counterfit American cigarettes in Asia.

The packaging was as close to the original artwork as the naked eye could see.

You can buy counterfit Rolex watches on the street for elcheapo.

I see counterfit logos currently being sold.

By counterfit, I mean look-a-like, feel-a-like copies of the original OEM logos.

The brandname is not important, the fact that they are currently manufactured attempts at duplicating the originals.

Their availablity is not so much for museum archival status but for financial gain.

Great caution must also be taken when buying owners manuals and reviews.

Sometimes there is very, very limited words to indicate it is a photocopy or digital copy rather than the oringinal documents.

This is particularly bad for those persons where English in not the buyers first language.

Some sellers state the fact that it is a photocopy or digital copy.

I was suckered in to buy a few brand new sealed videos at a swap meet.

First I did not realize the movies were still playing in theatres.

It was not until I got home and noticed the actors names were mis-spelled and in one case the list of actors was for another movie.

When I went to return the defective one the next Saturday, I had missed a police raid and they were gone for good, so was my money.

This is the free enterprise system at work here.

Link to comment
Share on other sites

Guest matty g

Lots of counterfeit components out there. Amperex "Bugle Boy" tubes, Mullards, Telefunkens with the diamond logo stamped in the glass on the bottom (still can't figure that one out), "Marantz" stamped output transistors, you name it. Let the buyer beware - too bad all that energy can't be focused into making a genuinely good product!

Matt

Link to comment
Share on other sites

  • 2 months later...

Some interesting links:

Not suggesting these very expensive electrolytic caps for use in

speakers, just that standard caps are mentioned in this document:

http://www.cde.com/new/tech/capability.pdf

http://www.elna-america.com/tech_al_reliability.php

Standard 85 deg C aluminum electrolytics have a relatively short

expected life.

Link to comment
Share on other sites

  • 2 weeks later...

Hi Ken,

I happen to be recapping some of my equipment and noticed that there are some new low cost, 8 to 10,000 hr long life electrolytic caps that seem to be a big improvement over the 3K hour types. Came across the links in the process.

I think this thread started before we had a mods and tweaks section.

Link to comment
Share on other sites

  • 2 months later...

I read a post on the Forum by Bret claiming that I say:

"I want to believe Pete when he says that capacitors are all alike, and I want to believe Ken when he says that changing capacitor types will change the crossover"

From: http://www.classicspeakerpages.net/dc/dcbo...mesg_id=126#120

I don't like being misquoted, in fact I've not said anything remotely like the imagined statement that Bret wrote. Indeed, I've gone into the details of ESR and differences in capacitors, and did my own capacitor test:

http://www.classicspeakerpages.net/dc/dcbo...ing_type=search

People were not able to hear a difference in a switched capacitor test as stated at the start of this thread. However, they used an 8 ohm speaker where the ESR of the caps was insignificant, even for the electrolytic. My point is that the difference is mainly capacitance value, and ESR, and that for "good" caps matching ESR is all that's required. Often, in well designed crossovers even matching ESR is not important, it depends on the details of the design.

I have been measuring and reporting ESR on several different designs that I've looked at so that people can choose a reasonable resistor to use in series with film caps. This goes completely against Bret's statement "I want to believe Pete when he says that capacitors are all alike" and I find it offensive that he makes such statements in support of his emotional need to discredit me. The spin doctor. If his spun interpretation of what I've stated here was at all true there would be no basis for me to suggest using a series resistor when replacing an NPE with a film type. Indeed, it would seem that I am in agreement with Ken contrary to Bret's spun interpretation.

I also see that people are quoting me on stating that a .33 ohm to .47 ohm resistor should be used with films in place of electrolytics. This is correct for what I measure of electrolytic caps that I estimate as having a dissipation factor of about 5% and for the smaller values used on tweeters and mids. Please don't generalize my statements and jump to conclusions.

From what I've seen most of the early classic New England speakers, AR, Advent, Genesis, have used 5% dissipation factor capacitors. However, The later Jensen Advents used what I estimate are 10 % dissipation factor caps.

Another classic design, the Spica TC-50 also seems to use cheaper 10% DF caps. The TC-50 has several economy non-polar electrolytics with an ESR of over 1 ohm, and the .33 to .47 value would not be correct in that case. One of these caps is used in a tuned notch filter and it's ESR has a strong influence on the Q of that filter. It is not so simple boys.

One can also conclude that expensive caps are not necessarily better from a sonic standpoint as many claim.

>I've said that listening tests are unreliable unless they're

>controlled, this is a fact that is difficult to convey or

>teach even to some professionals.

>

>I normally stay out of capacitor, amplifier, and wire

>discussions on the net so let me make a few comments. Sure

>caps that are at end of life or completely failed will

>obviously sound bad, this is not what I'm talking about. And

>certainly it's possible for a company to make a very poor

>product, this is not the subject either. I'm talking about

>competently designed components, but not necessarily

>expensive.

>

>I wrote in response to Vern's idea that it would be

>interesting for someone to do a capacitor switch test:

>

>This is an excellent idea, yes it has been done, and the

>results are interesting. One such test was done among a group

>of highly opinionated speaker builders. The group had a

>culture, they had common beliefs, most heard differences in

>caps. Solens, Daytons, Electros, the boutique caps sounded so

>much better, they were friends with a designer of some

>boutique caps, he would advise them on which to use. Many in

>this group were advanced, they had and used measurement

>equipment, understood much of the theory. Then they did the

>test - uh oh. Perhaps I'll tell the rest of the story some

>time.

>

>I want to add a few comments about what I saw in this

>experiment/discussion. The gentleman who did this switched

>capacitor test is a PHD economist with a good understanding of

>science in general but not in depth in electrical engineering.

> Yet he knows how to use measurement tools and has done some

>of the best home brew designs on the net. He wins speaker

>design contests against some of the other bests out there and

>their designs would give many commercial designs a run for

>their money.

>It is interesting to watch discussions and see how many

>members chime in with how the differences between caps are so

>obvious, blatantly obvious.

>This fellow added relays to one of his excellent designs so

>that he could listen to 3 different caps, an electrolytic, a

>low cost poly, and a boutique poly. He described when he

>first tried it that he thought the switching was not working,

>little if any difference. He then replaced one cap with a

>very different value and what do you know it *was* working.

>This is what I've found and it can be an eye opener. What

>more convincing does one need, it seems like the switch isn't

>even hooked up.

>These builders have meets around the country and they decided

>to do a comparison test. This was what one of the organizers

>had to say "The bottom line on the whole day was that

>nobody did better than flipping a coin."

>Here's the story summarized by that organizer:

>http://www.members.aol.com/pjay99site/captest2.htm

>

>This is why you see mylars, electrolytics, etc. in some of the

>most expensive designs out there.

>

>Pete B.

Link to comment
Share on other sites

People here often mention capacitor and speaker break-in, here is what Paul Barton founder of PSB speakers has to say about it. I don't know where this was first printed, otherwise I would give credit:

Finally, and perhaps most controversially, Barton talks about the supposed break-in effect of components that has become so popular in audio today. Break-in refers to running components for a long time (sometimes hundreds of hours) to the point where their components "settle" into their proper operating mode. Barton doesn’t doubt that some components do change subtly, but he thinks that the major improvements people think they’re hearing aren’t in the components at all. Barton doesn’t doubt that people are hearing these changes, but thinks that what they’re hearing is actually brain break-in.

Barton has examined his own speakers to test this. He has taken a Stratus Gold loudspeaker, built and measured some ten years ago, and re-measured it today. The deviation is slight, perhaps 1/4dB at most. Although that deviation can possibly be heard, it is certainly not a huge difference that one may attest to hearing. Instead, Barton surmises that the difference in sound that people are hearing over time is conditioning of the brain. He cites experiments done with sight that indicate the brain can accommodate for enormous changes fairly quickly and certainly within the hundreds of hours that audiophiles claim changes occur in. Could this apply to hearing, too? Barton thinks that more often than not, what happens is that the changes in perceived sound that are attributed to component break-in are simply the brain becoming accustomed to the sound. He warns listeners not to fool themselves.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.


×
×
  • Create New...