Are high-end crossovers worth it?

[Deang]

I ended my response with some words from Dr. Edgar, you make it sound as if that was the totality of my response. What you call an "anecdote" contains relevant information that supports my response and explains in part why I build the way I do. I can't help but feel you're being deliberately obtuse, so let me help you out. Do you think I might have an anechoic chamber over here, or the things that I would require in order to present you with "polars of my networks". I could have just said you were baiting and trolling, but wanted to give you the benefit of the doubt, so I responded with the obvious: I'm building crossovers that were used for over 30 years. if you want the polars, contact Jim Hunter. The one network that I build that's not a Klipsch network uses the same principles of operation, so there isn't exactly a lot to report on there.

[Deang]

Right, it's just saying that it's an iron core as opposed to an air core.

"Does replacing all the inductors with much larger wire inductors so there is a lot lower DCR change the ability of the network to do it's intended job?"

Like Mike said, you need to keep the DCR about the same or a little lower or you change the behavior of the circuit. Now, something you can do is change out the iron cores for air cores. To keep the DCR the same using an air core, the wire will get bigger by default. Air cores are also expensive, and there is some debate whether the improvement in some of the measurable parameters translate into audible benefits. Also, since the new inductors will be bigger, you'll have to abandon your PCB and build point-to-point on a piece of plywood, etc.

[cradeldorf]

Thanks Deano & Bruce, I was mislead by a company (north creek) with their 8 and 10 gauge inductors with their claims of less DCR means much better sound. I was thinking about just building new ones with those air cores and huge solen caps and a Mills resistor or a dueland graphite one. but now I think I'll go the other way that we discussed. And I wholeheatedly think he's trying to be obtuse. At least it sounds like he is. I'll be getting ahold of you soon.

Edited May 31, 2014 by cradeldorf

[Marvel]

Mark... I posted to an old thread about JBL L100s in the general questions section, asking you if you ever got the crossover upgrades done. Still curious...

Bruce

[Schu]

I think you've been here as long as Dean has been doing this. Dean never claimed to be designing or inventing new networks. I think I have always understood that he is tweaking the Klipsch designs. Am I wrong about your understanding? In other words, I don't thing Dean has ever misled his public about his work. The paragraph above has a way of coming across as taunting Dean to show something he has never claimed is his objective. I believe he largely touts the products based on their sound in the Klipsch speakers, and that's what I hear people confirming when they buy them. In my own case, I bought his networks for my Belles, and my only basis for the purchase was the sound. Dean's sounded better by being less distorted, cloudy, fuzzy and with better focus. Most likely a function of the much better capacitors he uses compared to the old Klipsch ones in my Belles.

He provided a very nice explanation of what he does and why.

exactly^

very nicely stated, that's why when all this started coming up about data, I clearly stated there is data... a lot of it. It's called Empirical data/experience.

Edited June 1, 2014 by Schu

[Skelt]

I believe it is the Barkhausen effect that causes iron core inductors to sound different than aircore.

http://skullsinthestars.com/2012/10/01/making-magnets-speak-the-barkhausen-effect/

http://video.mit.edu/watch/barkhausen-effect-11295/

If this has this been discussed before please forgive me

[ClaudeJ1]

I believe it is the Barkhausen effect that causes iron core inductors to sound different than aircore.

http://skullsinthestars.com/2012/10/01/making-magnets-speak-the-barkhausen-effect/

http://video.mit.edu/watch/barkhausen-effect-11295/

If this has this been discussed before please forgive me

Ah yes, the coil Barking effect. LOL. Probably a 100db volts of amplification, which on a woofer section would be inaudible.

[Deang]

Something I probably learned before and forgot about. That's pretty cool.

There is a lot of discussion across the net and in books about what the differences are between iron core and air core inductors. What is clear is that air cores aren't plagued by any of the problems iron cores are, and most everyone seems to agree that iron cores should be avoided in favor of air cores if possible, especially in high pass sections.

Thanks Mark, right, you pretty much nailed it. I read that part of Mike's response and thought to myself, "why is he talking about these things and why is he trying to apply them to me?" The discussion was about the events that transpired on a certain evening, and the reactions of those present. I challenged his recollection of the events, because they were inaccurate.

As for crossover design in general, complexity certainly has its advantages, but then again so does simplicity. I have been offered the opportunity to build more complex designs on several occasions, but have declined simply because my personal experience has led me to the opinion that less is more. With that said, I am slowly being influenced by John Warren to be more open-minded. It's a work in progress. I'm stubborn, and often don't understand some of the things that he's trying to teach me, which is somewhat frustrating. However, I've come to realize that one doesn't need to fully understand something in order to benefit from it.

One thing caught my attention, which really exemplifies the rift between the subjectivist and objectivist. Mike said that you could throw $10K worth of crossover improvements at the RF-3, and it would never sound as good as the RF-5. The statement is based on the idea that the drive elements in the RF-5 are so much better than the RF-3, that there is nothing you can do with the crossover to get the RF-3 to sound better. The argument is fallacious. Just because a driver specs out better doesn't mean that it's going to sound better. There are innumerable examples in audio of where a loudspeaker built with lower quality drivers sounds better than a loudspeaker built with higher level stuff - and the reason is tied to the execution, of which the crossover is instrumental.

The drive units in the RF-5 aren't that much better than those found in the RF-3. The cost difference is mostly due to the cabinet size, and that the RF-5 is veneered on both sides of the MDF to reduce the possibility of warpage, while the RF-3 is covered in vinyl. If I were to change out the epoxy coated oval caps in series with the horn, along with the crappy wirewound resistors -- experience tells me that everyone would prefer the sound from the "lesser loudspeaker". I've proved the efficacy of this so many times, I can't believe I still have to defend it, seriously, it's ridiculous. I've used a room full people no less than three times, and I've got at least a gigs worth of feedback on my hard drive - it's piling up on my Facebook page too. The point from the beginning is simply this: people hear these improvements, but for some, the experience has no value unless they can see and understand the math behind it. That's fine, but don't deny or misconstrue events, results, and reactions - just because you can't be supplied with the data you want.

Edited June 2, 2014 by DeanG

[Deang]

I wasn't done ranting, it's tough doing this with a phone.

Mike mentioned Mark Blanchard. Yes. he's brilliant. Like many within Klipsch, he's very approachable. Tired of all the capacitor business a few years ago, I called him up to see what he thought. He told me that the biggest problem with capacitors is their non-linear behavior when they're exposed to high power levels, and this does effect the sound. I got some good advice that day, and decided to share it here. At the time, I didn't reveal my source because I wanted to see how it would be received, and I wasn't disappointed ("bullshit"). Another example of this is when I adopted the concept of DC biasing capacitors in a network to improve their performance and resultant sound. In spite of Greg Timbers of JBL being "adequately credentialed", having an amazing list of accomplishments, and implementing the circuit in his crowning achievements -- I once again had to deal with "bullshititis". I made a decision a long time ago that it's okay to trust those who's claims line up with my subjective experiences, especially if they're well credentialed.

Edited June 1, 2014 by DeanG

[Coytee]

most everyone seems to agree that iron cores should be avoided in favor of iron cores if possible, especially in high pass sections.

[Deang]

lol, nice catch - I wonder if I should leave it.

[beeker]

^{im gonna put this another way}

^{no unless YOU CAN AFFORD THEM!!!}

[cradeldorf]

I've always felt that the erse super Q iron cores gave much better authority in the bass region, but I do agree that air cores go better in the HF secton. but I base that just on what I have actually experienced, So don't ask for any hard measured facts because I don't have them.

[DrWho]

Dean never claimed to be designing or inventing new networks. I think I have always understood that he is tweaking the Klipsch designs. Am I wrong about your understanding?

My understanding of Dean's perspective is that the stock Klipsch xovers are not high quality. It is only when more expensive inductors and capacitors are added that he feels they turn into a high quality xover.

My perspective is that circuit topology is the dominant influencer on sound. I also believe there have been major strides in xover understanding over the last 30 years. In my mind, a "high quality xover" is one that implements the latest and greatest approach - and only because it actually sounds better (lest we be biased by numbers and forget the artistic side of all this). I was attempting to be practical by bringing up some of the major topics of discussion common in what I consider to be the high quality xover world. In fact, the specific topics I brought up were in in light of how the stock networks could be improved far more dramatically.

Btw, Dean was the one saying there were mountains of data - and I was simply calling him out on that. I guess to be fair, he lterally said there was a 747 full of data:

Plots and graphs are in abundance out here, and enough have been posted to fill a 747.

I hope I don't have to be the one to point out the difference between theory and data - Dean has and continues to post a lot of reasons for why he does what he does - nothing wrong with that, but all of that falls into the theory camp. I just know there are insights to be gained when actually digging through the heartless quantifiable nitty gritty of it all. In fact, I know that extremely well today because I've found other resources to actually teach me what is happening and that insight is clearly not being shared anywhere on this forum.

[Zen Traveler]

...I hope I don't have to be the one to point out the difference between theory and data - Dean has and continues to post a lot of reasons for why he does what he does - nothing wrong with that, but all of that falls into the theory camp. I just know there are insights to be gained when actually digging through the heartless quantifiable nitty gritty of it all. In fact, I know that extremely well today because I've found other resources to actually teach me what is happening and that insight is clearly not being shared anywhere on this forum.

Good post doc and it's good to see you pop in here every once in a while.

[DrWho]

Btw, here's a great summary in Q&A format about why different capacitors behave differently:

http://www.analog.com/library/analogdialogue/anniversary/21.html

Not included in that discussion is the mechanical stresses imparted by temperature and voltage (both AC and DC) that affect the physical size of the capacitor (which in turn causes the capacitance to change). This guy covers some of those details, and methods for measuring:

http://www.murata.com/products/capacitor/design/faq/mlcc/images/sokutei.pdf

(Keep in mind that current has an effect, but it really manifests itself in the losses of the cap and the resultant temp rise).

There are also quantum effects at very low frequencies and low currents too.

All of these things are audible, and have their own characteristic sound...

[Deang]

"Plots and graphs are in abundance out here, and enough have been posted to fill a 747."

That quote was from 2005, in what context did I say it in? There has a been a lot of various data posted on this site related to horns, drivers, tweeters, and crossovers. If you missed that, then I guess you were sleeping.

EDIT: said on the second page of this thread. In the context I said it in, it's true, and I stand by the statement.

The old Klipsch networks have parts that are deteriorating, and so they don't sound very good. Some were built with penny junk polyesters and sounded bad to begin with. I don't need a plot to tell me that one thing sounds better than something else, and neither does a room full of people -- which was the point I was making. This is a fact that has consistently been demonstrated at the various Gatherings. I'm primarily in the restoration business, and believe that the original circuits built with good parts sound pretty damn good. I have yet to hear a modern filter built with a truckload of parts sound better. Considering my background in ownership around here, I think I've actually earned the right to have that opinion. The AK-4 network meets your criteria, but I refuse to build it because it doesn't sound as good to me as a simple Type A. No matter, please, continue on with your meanderings.

Since you've "designed hundreds of crossovers", why don't you design one for us. I'd be happy to build it and send it out for the listening tests I do.

Edited June 3, 2014 by DeanG

[Thaddeus Smith]

Since you've "designed hundreds of crossovers", why don't you design one for us. I'd be happy to build it and send it out for the listening tests I do.

Make sure it's 1983 Cornwall compatible please.

[minermark]

This reminds me to get w/Crites, new/old Scala now in my stack needs to be refreshed.

The splits have new Crties last year, so I need to match the stack set up.

[Deang]

Yeah, you wouldn't want one half of the stack sounding better than the other half. : )

Edited June 3, 2014 by DeanG