Chorus II crossover conundrum! leok, deang, alk, mike stehr; share your wisdom!!!!!

[krustyoldsarge]

I've spent the last couple of days trying to digest the numerous posts on the ChorusII crossover. Deferring to the crossover experiences and knowledge of leok, mike stehr, deang and alk, would humbly like to gather opinions on the subject before I go down this road.

I want to upgrade my crossover components, but according to what I've read, the Chorus network defies upgrade! Upgrades from electrolytics to film caps or inductor upgrades change the driver voicing within the speaker. The board is tiny (and fragile).

Nonetheless, some brave attempts have been made to tackle the beast! My questions for those who have attempted upgrading a ChorusII crossover are:

1. Has anyone been been able to completely upgrade the all the caps and inductors (save the autoformer) AND preserve the voicing/balance of the speaker! How?

2. Has anyone used goertz alpha-core inductors in Klipsch speakers with any success? Was the voicing thrown off?

3. Does anyone have a pair of ChorusII crossovers (maybe with a broken board) that they are willing to sell (would rob for parts)

There are already endless threads on caps! I'm not sure I'm any smarter for having read them all, though! I've got some 2uf (1%!) polys on the way.

Last year I did crossover mods to my KG4s, upgrading the caps and coils. The difference was remarkable (but not cheap) Will post pics later when I get back to home station. Of course, I finished this project just in time to buy the ChorusII's that I had been searching for last 7 years. So right now, the KGs just sit.

I also screwed up by originally posting this to HT

It never ends

[Al Klappenberger]

Sarge,

I sympathize with your frustration over crossover upgrades! It is not a

simple problem especially with all the conflicting advice you read here on

the forums! I have looked at the schematic of the Chorus II (I don't know

about the original one though). Anyhow, it's a typical Klipsch heritage type

design. It's impedance goes all over the place! It also uses both a resistor

and a transformer to reduce the level to the squawker. I suspect that was

done to utilize the transformer model that they had tons of on hand that

didn't have the tap in quite the right place! I personally think it needs to

be redesigned, but there's nothing in it for me to do it for the same reason

I stopped making the Cornwall network, it would be too expensive for the

speaker! All you can do is upgrade the part quality. You should get rid of

the resistor in the transformer circuit, but that would require instrument

testing of the acoustic output and changing the parts values to compensate

for impedance levels. It's also a 2nd order filter which complicates things

even more!

I will suggest some basic things again that I always suggest:

1 - Do NOT use foil type inductors. These are expensive and have their lowest

loss far to high in frequency. You want the peak "Q" at the filter cutoff

frequency. For a tweeter filter, that's around 6 KHz. The best inductors are

Solen Litz by a factor of three to one of ANY gauge solid wire. For woofer

filters, use solid wire. Litz wire only helps you above 4 KHz or so.

2 - Use polypropylene caps. The film and foil types (Hovland music caps) are

best but are very expensive I do not come in large values. Metalized

polypropylene are nearly as good and cost far less. Paper in oil caps are

very poor and are equivalent to a cap in series with a resistor. Don't use

them! If you want the mellow sound they give you by killing the highs, just

put in an attenuator pad and you will be able to adjust it to your taste!

3 - To decrease losses in caps, parallel up two or more to total the capacity

you need.

Also....

4 - The tweeter filter looks very much like the one in the "AA". It MIGHT

benefit from the 1 uFd bridging cap mod that was posted a while ago. My guess

is that it WOULD.

Al K.

[krustyoldsarge]

Thanks Al!

After my original post, I did some more digging in the archives and saw your measurements on the Litz coils versus the foils. Was too late to save me from bidding for some foils on ebay, but with any luck I'll get outbid! Will follow your advice, there.

Also want to try conditioning the new caps, per leok I believe.

The ChorusII is a magnificent speaker, too bad klipsch took what seems to be a less than ideal solution for balancing the crossover voicing; slapping in a resistor. Can certainly understand your reluctance to tinker with a CII crossover. On the other hand, maybe the same approach was taken with the Forte. There might be enough of them out there that lessons learned on one could be applied to another.

I've got an opportunity to scrounge some Aerovox PIO 6uf motor starter caps for the squawker but will probably pass on that, too.

By bypassing with a 1uf, I'm guessing you're placing it before the first 2uf and after the second 2uf immediately before the (-) line to the tweeter.

[Al Klappenberger]

Sarge,

I'm not sure what you mean by "conditioning" caps, but I wouldn't bother. Parts and networks do NOT "break in" as a lot of people claim. It's your ears that break in! So people might like to put a DC voltage on a cap and test for leakage, but that's usually a "go, no-go" test. I do that with old unused electrolytics that have been sitting in the junk box for a while, but that the only time I do it.

The 1 uF cap goes right around the tweeter filter, from input to output. I think the description you made is correct.

Al K.

[J.4knee]

Al,

You seem to have the innate ability to answer the very question in your routine posts that I was going to ask without me asking. You did it this time with the comment on the woofer inductor and this is about the third or fourth time you have done it. Hmmmmm.....

Sorry Sarge didn't mean to hi-jack your thread.

[Al Klappenberger]

J.4,

I do have a tendency to repeat myself, don't I? Old age I guess!

Al K.

[leok]

I have tried Hovlands, sonicap, and Jensen PIO in RF-7 and Chorus-II crossovers. I have tried the Hovlands conditioned at about 70V (+ & -).

The Hovlands need conditioning. I don't know where Al gets his information or why he says polypropylene caps don't need conditioning. I am an electrical engineer and have worked on audio and medical instrumentation for years. Capacitor conditioning is an industry standard if high quality and reliability is desired. Polypropylene cap manufacturers refer to the "clearing" or "conditioning" properties in technical papers. I find that conditioned Hovlands sound very good. Unconditioned Hovlands sound terrible.

I prefer Jensen PIOs to Hovlands, by far. But, I can't make any real technical sense out of why that should be so, except that the PIO may have lower dielectric absorbtion, taking less energy to repolarize when crossing 0 Volts. My original choice, based on engineering, was the conditioned Hovland. I just tried the Jensens because so many people recommended them.

In my Chorus-IIs I replaced only the caps in the mid and tweeter circuits. I didn't change anything else. On a set of Forte-IIs I had also replaced inductors and woofer caps, and didn't think it was really worth the effort.

The crossovers are a little difficult to work on, but the improvement in sound is pretty dramatic.

Leo

[JohnA]

O.K. Leo, describe how to condition capacitors. I think I can imagine how, but that is not the same as knowing.

[leok]

Apply a DC voltage, substantially higher than the cap will ever see in use, but not higher than the cap's rating, to the cap, for about 20 seconds, in both polarities (one polarity if the cap is a polar electrolytic*). Having a resistor of several KOhms (4.7KOhms for example) in series with the voltage source is a good idea.

* I know there wouldn't be a polar electrolytic in a crossover, but this process is also used on that style of cap, especially tantalums.

I use about 70 Volts on 100V caps like the Hovland "Musicaps."

I made a little device that uses rectified house power. But Dean snapped a bunch of 9V batteries in series, which I think is a better idea.

You have to look at a crossover circuit to figure out which caps can be conditioned in the circuit and which have to be removed. Always disconnect one lead to each speaker so you don't blow the speaker. Of course, if installing new caps, condition them before installation.

Leo

[BEC]

Might I suggest a test method here that will eliminate the audio memory problem and any bias that may be present in the cap choice?

Go to Radio Shack or similar electronics parts place and purchase a double pole double throw (on-on) switch. Make sure it is a 2 position switch not a 3 position (on-off-on) Place a short across the two center terminals. Connect a pair of wires to each of the other terminal pairs on the switch. Connect the two caps you want to test in series in the circuit where they operate. Connect one pair of the switch wires across each of the two caps. Now when you operate the switch, you will place a short across one of the caps and remove the short from the other each time the switch is operated.

Test part 1:

Play music and listen to the sound. Have another person operate the switch at random while you listen. You should not be able to see the person operate the switch and and that person should not tell you when he has operated the switch. Can you identify a difference when the switch is operated? If you can't, the test is over and there is no audible difference between the two caps.

Test part 2:

If you can identify a change in sound from part 1 of the test, continue to compare and determine which sounds best to you. After many test cycles, the person operating the switch for you should be able to tell you if you consistently picked the same cap as sounding best and which cap you picked.

This test method should work well for several component tests. Inductors for crossovers would be tested exactly the same way that the caps are tested. The test is easier, of course, on crossovers that are easily accessible.

Bob Crites

[Deang]

I could of swore I posted into this thread, I guess not.

"I want to upgrade my crossover components, but according to what I've read, the Chorus network defies upgrade! Upgrades from electrolytics to film caps or inductor upgrades change the driver voicing within the speaker. The board is tiny (and fragile)."

Nothing "defies" upgrading. You say you are concerned about the new parts changing the "voicing -- yet you stated you upgraded your KG4s with good results. What's the difference? I guess any change in the sound changes the voice, but it's not like it becomes some speaker you don't recognize anymore. Hell, if we are going to apply that strict of a definition to the word, then tone controls and gear change-outs change the "voice" too. After all, who here listens with the same gear Klipsch used to do the original voicing? The goal is to use quality parts which do a better job of letting the music through. We can discuss what is responsible for the improvements until the cows come home, but the bottom line is that improvements are there, and in the end -- that's all that really matters.

"Has anyone been been able to completely upgrade the all the caps and inductors (save the autoformer) AND preserve the voicing/balance of the speaker! How?"

You're working with PCBs, and so there are inherent limitations to what you can do because of size. Leave the electrolytic and iron core inductor in the woofer ciruit. The board simply doesn't have the room to accommodate a polypropylene capacitor or higher quality inductor. The good news is that the ears don't seem to notice/respond as much to changes in the LF circuit as they do to changes in the HF circuit. If you want to capitalize on any improvements good parts would bring to the bass -- you are going to have to ditch the PCBs, and rebuild the crossovers on separate boards using point-to-point wiring.

"Has anyone used goertz alpha-core inductors in Klipsch speakers with any success? Was the voicing thrown off?"

I've used them before in the LF section with my RF-7s, as well as a couple of rebuilds for others -- primarily because they had the correct DCR and they fit nicely on top of the PCB. I've never used them in HF circuits. I would go with Al's recommendation and use the Solen Litz air cores -- you should have no trouble fitting those on the boards. Be forewarned though -- they are a pain in the butt to work with. Solen does not send these with tinned ends, and it sure is a lot of fun scraping all of the enamel off of the stranded ends.

Cap "conditioning": It makes sense that it would work for a metallized "self healing" type, but I still can't figure out how it would do anything for a film/foil or PIO.

Bob, I don't trust your laptop.

[Al Klappenberger]

Leo,

My response to your points is simple. I disagree 100%.

All my type "A" networks are made with UNconditiond matched set of Hovlands. I guess that's why everybody says they sound so "terrible"!

If conditioning was so important, it would be done at the factory, or maybe it IS!

Al K.

[leok]

Al,

I know you disagree 100%. I would have a similar statement: Since conditioning is so important it should be done at the factory.

Leo

[Al Klappenberger]

Lets see... We are going to "condition" a capacitor. What shall we

"condition" it for? Let's "condition" it for the "conditions" it is to be

operating under. Every crossover network I have ever seen operates under the

"condition" of 100 VDc on it for X seconds then it reverses to 100VDc in the

other direction for x more seconds. RIGHT! NOT! This is AC at a frequency of

1/X (half that actually)! If we speed that up we have audio! Rather, let's

"condition" it for the "conditions" it will see in normal continuous

operation. That would be to let it run under normal "conditions". That's real

hard to do. You build the network and let it "break in". BTW: That doesn't

happen either! These effects a minuscule, if real at all, compared to human

perception changes from day to day! THIS is what is responsible for "break

in".

AC had no polarity. It goes both ways. In an instant all the DC you have

applied is gone along with any changes it made at an atomic level in the

dielectric! ALL THIS IS BULL S__T! The caps in a crossover network see a

range of frequencies and levels in totally random patters (called music).

This is a formula for ERASING patters, not forming them! Have you ever DE-

magnitized a screwdriver or a wristwatch? You do it with an alternating

magnetic filed. Another example is a bulk recording tape eraser.

I have tested Hovland caps with DC bias on them. Their capacitance or loss

factor does not change, AT ALL! Besides, I suppose that a year or so down the

road the capacitors all remember that it once saw 100 VDC across it for a few

seconds, or even minutes! Come on!

The field of stereo is loaded with old wives tales and B.S. This is just

another example! Even manufactures promote this sort of thing, so it's not

hard to see why people believe it!

Al K.

[sfogg]

"Might I suggest a test method here that will eliminate the audio memory problem and any bias that may be present in the cap choice?"

I have an ABX Comparator that could be used for this sort of test. Double blind and no need for anyone else to swap capacitors around without the listener knowing which is which.

Shawn

[Number 9]

Has anyone read about (seen and heard would be even better) the new JBL top-of-the-line K2 loudspeaker? They use a battery in the crossover network to supposedly provide some type of constant current flow through the caps when the speaker is not in use. This makes me think that obviously something is going on, but what it is, I don't know.I doubt they would do this as a "marketing gimmic".

[leok]

Sorry Al,

As much as this seems to upset you, I find that applying a substantial DC voltage, in both polarities, to a Hovland Musicap results in a cap that performs much better in the mid and tweeter circuits of Klipsch crossovers. That a new cap may have conductive flaws that are "cleared" by the process doesn't pose a problem for me, except that I would prefer that Hovland perform the operation.

Most commerial grade caps suffer from similar problems with manufacturing flaws. Initial application of power usually clears them. Some suffer catastrophic failure. What I have found in crossover circuits is that some, without sufficient power in the application itself, may never clear.

Sorry Al. It may just be a slightly more complicated world than you thought.

Leo

[leok]

I felt this thread was due a little research into possibilities of self-healing in a film/foil style cap, since I can't imagine a modern film/foil design to have no such properties at all. I found the following.

I don't expect Hovland, or any other cap manufacturers to explain the secrets of their designs, or to say that of course their caps are likely to leak a little until conditioned.

Following from http://my.execpc.com/~endlr/construction.html paragraph "Metallized film vs. film/foil" halfway through the paragraph:

" .. Other variations include foil layers separated by metallized layers to get reduced ESR yet still allow for self-healing, and even foil and metallizing in the same electrode layer. In most cases, the idea behind these various constructions is to get the good self-healing of a metallized-film capacitor with the low ESR/high current capacity of film/foil."

Leo

[Al Klappenberger]

Leo,

Your quote speaks of "healing". Healing what? Healing implies some injury. What is going to "injure" the capacitor dielectric in a crossover network or in any other circuit where the capacitor is operated within its ratings? I doesn't matter anyhow. This is small potatoes even if it is so. It's like trying to stop global warming by poring a few ice cubes on the sidewalk. computer analysis shows the voltage levels in the networks I build to be well below the caps ratings even at extreme levels. The caps have got it easy!

AL K.

[Deang]

Al, I'm sure you've noticed how most of the caps out there are advertised as "self healing". So yeah, it makes you wonder -- "healed from what"?

My understanding is that the raw polypropylene and polyester is delivered to the manufacturers with small defects in the sheets resulting from the manufacturing process. Most notably, very small pinholes. The plastic film is then attached to the metal foil by some adhesion process, and then wound. While being wound, the plastic film stretches, and the holes open up, exposing the metal foil underneath. When voltage is applied to the cap there is some low level arcing due to the exposed foil -- which translates to noise. If and when the voltage reaches a high enough level, the arcing becomes powerful enough to blow the exposed metal around the hole away, hence -- "self healing". Of course, in Klipsch crossovers, the necessary voltages are rarely if ever reached, so the self healing aspect of the caps is never realized. Film and foils, if not double wound, are also prone to this.