difference in sound when good capacitors go bad..........

[PrestonTom]

As far as advice about caps, I am not a big believer in spending too much money on caps. This is my bias, so I simply use Solens and am done with it. I could probably be talked into to trying bypass caps (a second smaller cap in parallel). Bypass caps are typically about 10% of the value of the main cap and frequently are one of the film types. For specific advice on caps I would take BEC's opinion and DeanG's opinion and then average the two. I say this in jest. They each have an individual perspective and they each have a good deal of experience. Their opinions will probably differ, but they should be considered.

As far as crossover point, you really want to get a point where the two drivers have a comparable dispersion and amount of distortion, and where the drivers are well-behaved in terms of changes in phase and amplitude (as a function of the local range of frequency). So really it depends what cabinets you are referring to. If you have substituted non-stock drivers or horns, then we need some measurements. Don't try it "by ear".

I am biased toward electronic crossovers because they allow time-alignment and to a lesser degree they can also allow steeper filters (which can hide a multitude of sins). Time-alignment is more important for some designs than for others (K-Horns would be a good candidate). Bi-amping can help in a number of situations and it need not be all that expensive (a smallish amp for the mids and the highs can be nicely supplemented with a giiger amp for the lows).

Good Luck,

-Tom

[PrestonTom]

You can't biamp a three-way speaker -- you need to triamp. So you need three amplifiers, along with the processor to apply the crossover and EQ parameters. Those parameters are derived from the data accumulated after careful measuring, using instruments a majority aren't in possession of -- and wouldn't know how to use and/or intrepret the data correctly without some training. Hobbyists, or "audio enthusiasts" are not electrical or acoustic engineers. Without a way to test and verify, there is simply no way to know when it's right. Properly designed passives sound every bit as good, and in some cases -- better. Related but unrelated, with the majority of commercial recordings, less revealing can also mean more musical or listenable, and who wants to build a case against that?

In the case of old Heritage, many of those caps are exhibiting breakdown of the dielectric, and this causes the resistive element to climb -- causing a reduction in the signal and reduces sensitivity of the speaker. Before a cap replacement, the highs will sound muffled and/or subdued, and the midrange sounds a little dirty. What most notice after a capacitor changeout is increased tweeter output, though if they're paying attention they should also notice not having to turn the attenuator on their preamp as high to get the same perceived volume level (as previous). The second thing they notice is a much cleaner presentation. You don't need an A/B test to notice any of this.

Sorry Tom, but it effects imaging and soundstaging in a big way. Maybe you shouldn't have used those crappy Solens.

Dean, Dean, Dean,

You used words like "cleaner" and "increased output" and I wholeheartedly agree. Those are consistent with my experience also.

"Imaging" and "sound staging", well we will need to disagree on those. I interpret these words as indicating the "left-right" dimension of the stereophonic spread (and it stability, diffuseness or "pin-pointedness") and the apparent depth (or perhaps diffuseness) of the image. These two aspects of the precept are really mostly governed by decisions in the recording process and the speaker's setup within the room (and its room interaction). Perhaps if these are obscured when the driver's output is decreased (due to an old cap), then there might be an impact, but that is a "maybe".

If some one had a problem with imaging and sound stage, they should first look at the recording and the speaker design and particularly its set up in the room. If the channel separation were very very poor in the electronics it might also hurt the imaging. But first things first and crossover caps in a decently designed crossover are not going to be the weak link.

I am not down on passive crossovers (in fact, some of my best friends .....). It is just that with currently available electronics, some horn loaded cabinets can be easily improved with time-alignment. Passives can not do time alignment (with the magnitude of the delays involved). However the benefit of time alignment is less if the room is sufficiently reverberant and to a lesser degree if the crossover slopes are fairly steep.

I also agree with all the above comments that properly setting up a crossover point (and slope) does require measurement equipment, but some of these measures are already available.

Good Luck,

-Tom

[Erik Mandaville]

"I also agree with all the above comments that properly setting up a crossover point (and slope) does require measurement equipment, but some of these measures are already available."

We should note, too, that many of the older Heritage networks (the type A is always a good example) were designed around established criteria that is available to anyone. The fundamentals of designing a crossover include the impedance of the drivers being used, the desired crossover frequencies, and attenuation networks (where needed) for adjusting gain. Certainly compensation circuits such as impedance equalization, notch filters, and so forth require measurement, but the type A (among some of the other Heritage designs) do not use those special circuits. This can be very simply verified by the fact that the value of the single series capacitor in the squawker branch of the type A is selected according to nothing more than the (reflected) impedance of the driver (which happens to also include a sort of built-in attenuation network) and cut-off frequency. On-line calculators will provide a virtually identical value, as do design tables in books on crossover design. The same is true for the woofer, which uses a single series inductor, the value for which, once again, is derived from known figures -- impedance and frequency. In many cases, driver manufacturers will specify the precise network to make -- such as 18dB/octave @ 3,500kHz for a tweeter. And that's it. No mention of Zobel network, notch filter, or anything else. Given that information, there is a very specific way of designing an 18/dB network for that crossover frequency and driver impedance, and the information needed to create that type of network need not be made more complex than it is. In the case of the tweeter I'm using, it involves two capacitors in series with the driver, with a shunt element (an inductor) between the two capacitors. That's all, and in doing this one would be following the design suggested by the company that designed the driver. What if you were to find that the tweeter was too loud compared to the other drivers? There are formulas for making very simple L-pad circuits -- or these days, very conveniently, on-line calculators where all one has to do is punch in a couple of numbers. These work by the way, and I have tested them against formulas I have in books on the subject. Loudness doesn't need to be measured with anything more than one's ears. If it's too loud, turn it down until it sounds better to you.

Just to be clear: Obviously very high order passive crossovers (which IMO may, depending on associated equipment, bring about other compromises) MUST be thoroughly tested and fine tuned. Using loudspeakers drivers with unknown behavior, as was indicated above, is another case where measurement would be important. My point has to do with the fact that the type A (which also happens to be a favorite of many) does not contain compensation networks or circuits of any kind. It was designed around nothing more than simple and readily available information: driver impedances, reflected impedance, desired crossover frequencies, and desired amount of attenuation. The squawker doesn't even use a true band-pass. It's an extraordinarily simple design, which IMO is one of the reasons why it may sound as good as it does, particularly with lower power amplification.

I agree that new capacitors might be an improvement for you, and also that it's (possibly) not necessary to spend overy large amounts of money on expensive capacitors. If you can afford the likes of Hovland and others, absolutely let your own ears decide what works best for you. In many cases, mylars can be very satisfactory, but in this case I'm speaking of some of my own experiences in building speakers and crossovers. As others also indicated, you may not know what you're missing until you try for yourself. The autoformer and inductors do not wear out with time in the same way, and so a simple cap switch might do the trick. Solens are good for the price, IMO, and I have experimented with all types, cheap and not cheap, for a long time. ALK makes beautiful crossovers, and the ES versions border on electronic sculpture. Because of the complexity and number of passive components, they may not be an ideal choice for really low-pwer SET amplifiers (I'm talking about a couple to few watts max), but then again, maybe there are those who use them very happily with single-ended 45 and 2A3 triodes. I prefer not to tell someone what they should or should not like.

Have fun!

edit: Bob C. also sells capacitor upgrade kits, which may have been mentioned, and that will make things more straightforward.

Erik

[mark electro]

I have read and digested the wealth of information you all have bestowed upon me.

I happen to have enough back up amps to make tri amplification happen; however, I would have to purchase electronic crossovers and probably quite a lot of test equipment to make this happen.

While pondering; the voice of reason walked in; (my girlfreind LOL), and said " you have a set of speakers that you only have $300. invested in, so why don't you just restore the crossovers instead of re-engineering an all ready excellent design; at great cost? "

I think she might be right on this one.

So for now I am going to stay with passive networks in my heresy's, and dream of the K-horns I will get one day.

So the way I see it now: Which capacitors should I get and what would the differences be?

I know Bob Crites uses solens. Many people use Audio Cap Theta PPT's. Yet a lot of others use Hoveland Musicaps.

So let me start at the Hovelands. I notice they seem to be very robust in build quality. Will the 16ga. braided leads help to lower the ESR even further? Also what sound inprovement could I expect? Will the sound be smoother?

Now the Audio Cap Theta's: look like they are well built. Will the gold plated leads bennifit me.What can I expect from their sound and smoothness?

Solens definatly the best value; but it seems a lot of people don't care for them. Why? And I guess the next question would be: Will they approach the sound quality of the OEM Aerovox, paper in oil, capacitors?

Will the Audio Cap's or the Hoveland's meet or beat the quality of the of the OEM caps?

Maybe some of you have some experience with this delema I now face?

I would like to restore my speakers to OEM or a bit better; on a buget.

[BEC]

"I know Bob Crites uses solens."

Actually, Bob Crites uses Sonicaps. I just checked with him.

Bob Crites

[mark electro]

"I know Bob Crites uses solens."

Actually, Bob Crites uses Sonicaps. I just checked with him.

Bob Crites

Truely sorry Bob; I always mix those two names up.

So do the Sonicaps equal or exceed the OEM caps??

Mark

[PrestonTom]

".......

So for now I am going to stay with passive networks in my heresy's, and dream of the K-horns I will get one day.

....."

Well, Mark that is actually the best plan I have heard all week. A Heresy cabinet is fine, but it is no match for a Klipschorn. Even in its stock form, a Klipschorn will sound very good. Save those dollars for the upgrade to a Klipschorn.

Good luck and I hope your dream comes true,

-Tom

[BEC]

"I know Bob Crites uses solens."

Actually, Bob Crites uses Sonicaps. I just checked with him.

Bob Crites

Truely sorry Bob; I always mix those two names up.

So do the Sonicaps equal or exceed the OEM caps??

Mark

Sonicaps are better than the original ones in your Heresys. They are also better than any other caps I have been able to find and test. You know, (Dean can verify this), I tried hard not to like Sonicaps because the guy that owns that place they come from is a real audiophylic nutcase. He called me up to convince me to use Sonicaps and used every one of the audiophile terms I hate the most. After getting through all the marketing hype and "veils being lifted" and perhaps even "skirts being lifted" and other meaningless babble about soundstages and images, he did indicate that they are manufactured to minimize ESR and that their manfacturing process was of high quality to produce a consistent product.

Anyway, by now I have bought a few thousand of the Sonicaps in the last two years I have been using them. I am still pleased with them and don't think you can find better caps than the Sonicaps.

Bob Crites

[PrestonTom]

Bob, when I refreshed the caps on my Klipschorn I was not able to accurately measure the ESR of the old caps or the Solens that I replaced them with.

What I did however was to replace the the 13uF cap with two caps (7 & 6 uF in parallel). My thinking was that this might be a poor man's version of a "low ESR" cap. Have you ever bothered making that comparison? Also, what actually is the major contributor to the series resistance in a capacitor. I thought I had read somewhere that it was (mostly) the internal connection of the leads. I recall that I was not convinced when I read that however.

Your thoughts, sir?

[BEC]

Tom,

I have played around a bit with that a couple of years ago. I thought at first what you have said about paralleling would be right for lower ESR, but found one pair that came up worse paralleled than they did measured separately.

I think where this idea gets off the mark is the fact that the ESR is measured as an AC parameter. We put a pulse of known voltage into a capacitor and measure what we see back out of the cap. So, ESR is a measurement of all losses in the cap expressed as an equivalent value of resistance. I think the major losses are lead connection, inductance, and material impurities.

But, since good caps have very low values of ESR, trying to reduce it further is likely pointless. A good Sonicap measures at around 0.02 ohms ESR. Other caps like Dayton or Solen also measure pretty good, but vary around more inside a group being tested. So, seems to me that we can use measured ESR as a pointer to quality. If all caps of a type measure very close to the same, I think we could infer that the manufacturing process is very well controlled.

Then, we could talk about how much ESR is too much. I think, personally, I can hear when ESR is about 1/2 ohm in a tweeter circuit. The typical 20 to 30 year old crossovers I rebuild usually have about 1.5 ohms of ESR in the tweeter circuit for Type AA crossovers and about 0.8 ohms or so in something like a Type B crossover. I think just about any polypropylene caps will make an audible difference in those crossovers. But, I think, (relating ESR to quality again), rebuilding with caps that have very low and consistent ESR means they are likely to be good for longer than other caps.

Bob Crites

[mark electro]

Thanks Bob, I think I will be contacting you soon for a set of caps for my Heresy's.

While I am at it can you let me know about my Heresy II crossovers? Maybe a schematic? Or should I plan to send them in to you?

[BEC]

Heresy II schematic.

Bob Crites

[mark electro]

And again: Thanks Bob.

I know you are anticipating my next Question(s).

It looks like any good quality cap is considerably larger; (approx. 1" x 2.5" on ave.), in a 68 uf value. Looks like all four caps together are like trying to stuff 10 pounds of *#@* in a 5 pound bag on the back of a binding post cup.

Do you have a solution to this?

I don't know if we are allowed to talk pricing on this site; but maybe you could e-mail me???

If you can: price for Heresy kit and possably Ct 125 Pair too.

They are all on my wish list.

Thanks again,

Mark

[BEC]

Mark,

Let's talk pricing by email.

bobcrites@mac.com

Now, in my opinion, the one cap you mention, the 68 uF is not really a problem. That woofer roll off cap can be a non-polarized electrolytic like Klipsch used originally. But those little mylar caps need to go and be replaced with good polypropylene caps. It is possible to rebuild the Heresy II crossover using the same method as the original. I prefer, though, to relocate the crossover to a separate board mounted in the bottom of the cabinet. That allows more room for the components and gets the heavy autotransformer off the plastic input terminal cup. I have seen several of those that had just broken off the plastic and were hanging by the wires. This new board can have a terminal strip and has plenty of room to either mount the reused old autotransformer and inductors, or it will also accomodate all new parts.

Bob Crites

[Erik Mandaville]

It's interesting to see the evolution of (Klipsch) crossover design from the earlier to later, somewhat more contemporary approach. Note the tweeter connection.

Just FWIW: This is not to be automatically contrary, but rather to offer the idea that mylar capacitors aren't inherently bad and in need of immediate replacement. That might be another subjective sort of call -- just as the NP electrolytic shunt element has been 'discussed' here by some to be more appropriately replaced by a plastic film type (though I'm in the same camp that would use the MUCH cheaper and MUCH smaller NP electrolytic). A passive network for the Jubilee had been posted here that in fact specifies mylar. That doesn't mean a poly film type might not sound better to some, but I think could indicate that if they were so obviously inferior they would be avoided altogether by engineers. I've had the opportunity to change the mylars in mine on many occasions, and it's work that would take me several minutes. They sound fine to me.

The new tweeters are a fabulous upgrade! ....and still, even with that there are those who think the original alnico or mud-mag K77s are more musical. A very large portion of all of these changes is subjective. The alnico tweeters sounded comparatively dull and uninvolving. I want a tweeter that will help make a sizzle cymbal sound like what it is rather than an ambiguous hiss somewhere in the background. The C125s were a big improvement without sounding like they put a microscope on the recording.

Good luck on your Heresy freshening up!

Erik