Are high-end crossovers worth it?

[djk]

Ever looked at the schematic on the Behringer DSP crossover (DCX2496)?

15 electrolytics and 17 opamps in the signal path.

https://groups.yahoo.com/neo/groups/DCX2496/files

Edited June 13, 2014 by djk

[John Warren]

There's only so much you can do with the Klipsch designs.

Though not Klipschhorn the response plots shown in link can be demonstrated in the K400 and K77M.

The network design and driver positions work together to locate nulls above and below the listening location.

The biggest "improvement" that can be made in a Klipschorn is getting all the speakers to "speak" correctly relative to each other, i.e phasing.

http://www.northreadingeng.com/Forums/index.php?topic=5.msg14#msg14

Edited June 14, 2014 by John Warren

[SWL]

Normal sound to a producer who sits in a mixing room is going to be different than normal sound to the HiFi enthusiast. Normal to the guy who mostly hears music at clubs and public venues is different than normal to the HiFi enthusiast who doesn't regularly hear sound reinforced music. Normal is a very powerful factor

Couldn't agree more.

[DrWho]

I just don't get his weird recollection of the events in Michigan that night.

......

Mike, there was the year after that, with the two sets of Jubilees in the room. Are you sure that it isn't that what you're remembering - because that was pretty bad. Opamps and a $50 CD player sure didn't carry the night - at least not for me.

Ummmm - the night referenced in this thread happened at a pilgrimage in Indy, not Michigan.

I'm familiar with the Jubilee setup that came much later - I don't think I ever said that was anywhere near the epitome of sound quality. The CD Player wasn't to blame either.....those hotel rooms just sound bad.

[DrWho]

P.S. Immediately I can see all sorts of other "normal" conditions that are relevant. Normal sound to a producer who sits in a mixing room is going to be different than normal sound to the HiFi enthusiast. Normal to the guy who mostly hears music at clubs and public venues is different than normal to the HiFi enthusiast who doesn't regularly hear sound reinforced music. Normal is a very powerful factor.

This is why I like PWK's approach of calibrating ourselves to live acoustic music... I have this (naive?) belief that there is some level of universal commonality of preference. Certainly there is different subjective weighting, but I firmly believe digital can be done in a way that the analog purist could appreciate - it's just crazy hard to pull off, and I don't think it's actually been implemented yet in the audiophile realm.

Edited June 18, 2014 by DrWho

[Deang]

You're right, it was Indy.

Hotels rooms may suck, but so do $50 DVD/CD players from Walmart.

I've spoken with at least a half dozen people who were there that night that have no idea what you're talking about, and remember the events the same way I do.

I've repeated this experiment with the same results so many times, that I can't even keep the locations straight in my head.

Michigan was the AKfest. I should let Craig tell that one.

Edited June 19, 2014 by DeanG

[Fjd]

Great history on the ALK Universal revisions. I'm curious if anyone has any insight into the impact of change in the ALK universal inductor from 2.4mH to 1.3mH?

For example, since I'm thinking that the inductance of the voice coil is not constant, but will vary with the position of the voice coil, does either value of inductor in relation to the voice coil inductance significantly change the upper roll-off requency of the woofer driver for the better or worse in a Khorn or La Scala or Belle? Is there some other effect that I'm not aware of?

I'm also curious as to why I see builds using 7.5uF capacitors and builds using 6.8uF capacitors in the same spot?

3. Revision 1 to the original ALK: an adjustment to the high pass and low pass values were performed, driven by the idea that the VC of the K-33 is part of the network, which constitutes 6 ohms in series with 1mH of inductance (VC inductance of K-33). The 1mH of the VC is subtracted from the original low pass value (went from 2.4mH to 1.3mH). Easily identified by the use of Sonicaps.

There are now five versions of the "Universal".

What I think is that these are the original versions, which is what was on Al's DIY section before he removed it. The coil looks like a 2.4 instead of a 1.3 or 1.5, I see a 3619, not a 3619-ET, I see the old way of wiring the autoformer, and I see a bank of caps that probably add up to 40uF. This would correspond to a 500Hz crossover point. This network is not compatible with a Cornwall, since the crossover point is below the Fc of the horn. It might be compatible with a CornScala, depending on the horn that is used. One would need to attenuate the tweeter.

The five versions:

1. The original ALK

2. The original Cornwall ALK

3. Revision 1 to the original ALK: an adjustment to the high pass and low pass values were performed, driven by the idea that the VC of the K-33 is part of the network, which constitutes 6 ohms in series with 1mH of inductance (VC inductance of K-33). The 1mH of the VC is subtracted from the original low pass value (went from 2.4mH to 1.3mH). Easily identified by the use of Sonicaps.

4. Revision 2 to the original ALK: Built under license by me. Incorporated ClarityCap ESA and Mundorf Supremes. I later went to "double wrapped" AudynPlus capacitors in place of the Mundorfs. Also used the new 3619-ET, and the network was rewired to allow for 1dB steps in attenuation. I also built a version for the Cornwall.

5. Same as Revision 2, but built using entry level capacitors and called " Economy CornScala-wall Universal". Parts should be scaled based on the loudspeaker the networks are being used in.

[Deang]

I don't speak for Al anymore, it just gets me into trouble. We are worlds apart on some things and this is one of them. We finally decided it was just best to do what we do independently of one another.

My opinion is that the 1.3mH coil is undersized, which is why I stay with the 2.5mH value in all of my builds (I don't build AK or AL networks).

After initially insisting that the VC was part of the filter, some years later he agreed that it wasn't -- but since it impacted the impedance of the other section of the filter as seen by the amplifier, he "synthesizes" the inductance as part of the filter anyways, in other words, he still thinks it's better to back the inductance of the VC out, so the network's impedance seen by the amplifier is smoother.

I've mentioned to Al many times that the inductance and resistance values of the VC are not static, and are constantly changing. I often brought up power compression.

Al would be quick to point out that in the Type AB for the Belle, PK didn't even use a coil, and that the inductance of the VC itself was enough to get it done. I myself listened to my own Klipschorns like this for many years. Since the horn's determine the actual crossover point, there's very little summing error. Al believes the first order filter is so "sloppy", that it basically doesn't matter.

I think that just because first order filters are forgiving, this doesn't mean they aren't worthy of having mathematical integrity applied.

I know people listen louder than they openly admit, and in my experimenting over the years, I discovered that it just sounds better with the 2.5mH in the network. Admittedly, the difference is not jaw dropping, but it is audible, and the bass takes on a bit of a tubby quality.

7.5uF is the correct value, but it's not in a critical position. I pushed for the change to 6.8uF since it's a common value. Al said the impact on the impedance was minuscule and since it was within 10% in a non-critical position -- he would let me slide. : )

Edited July 6, 2014 by DeanG