A K-402-Based Full-Range Multiple-Entry Horn

[Chris A]

I've thus far refrained from descriptions that are difficult to quantify or describe.  My experiences with this loudspeaker have identified the "point source, controlled directivity" factors as the source of those differences that you describe, Claude.

 

I've found that the design itself integrates the phase and directivity of its output in the crossover region--a single crossover frequency at 475 Hz: in other words, I hear very little difference when I make changes to the crossover filter or the delay. This leads me to think that some of those comments on the esoteric crossover design of the DSL Synergies is perhaps more for marketing than something that actually makes a significant difference in listener preference.  I can see that the measurements in the 470 Hz crossover region are almost unaffected (relatively speaking to multiple-way vertically-separated-horn designs) with the crossover filter tweaks and HF channel delay. 

 

Perhaps this is because there is no higher frequency crossover in the upper midrange for this current configuration.  Not having that extra crossover apparently is a big deal.  The sound is coherent all the way up and down. 

 

In addition, the tonality that you refer to is what I've identified as full mid-bass point source and directivity performance (down to 175 Hz) which is extremely important for the listening effect that you describe.  This is something that is not present in any other loudspeaker that I've heard--except perhaps planars (...but planars have their own set of issues that get in the way like issues with coverage, polars, and compression that Toole has described that have shown to be unfavorable by listeners in blind listening trials).

 

Chris

Edited January 26, 2016 by Chris A

[Khornukopia]

During the weekend I decided that the reported price on the K-402 was to good to pass up, so I would order a pair and store them until I could start a build later this year, but on Monday I received an e-reply from Spencer Chao that the K-402 was no longer available for sale as a horn only. Although disappointed that I missed out on the deal, I was also slightly relieved that my Klipschorns were not to be moved yet, because they have been sounding very good since adding the digital electronic crossover and the tri-amps last year.

 

Still, the Full Range design is so logical that is very easy to be distracted with the thought of building a pair someday and then re-arranging the placement of several large speakers in my listening room, so please keep writing about how good they sound.

[ClaudeJ1]

In addition, the tonality that you refer to is what I've identified as full mid-bass point source and directivity performance (down to 175 Hz) which is extremely important for the listening effect that you describe.  This is something that is not present in any other loudspeaker that I've heard--except perhaps planars (...but planars have their own set of issues that get in the way like issues with coverage, polars, and compression that Toole has described that have shown to be unfavorable by listeners in blind listening trials).

 

Before I bit the bullet and simply purchased Sh-50's for my LCR, Tom Danley had me measure each driver section separately in my TD-1's and provided me with some active PEQ settings that would make the older boxes sound as good as his new Synergy ones. He claims he was still learning how to do better Xovers 15 years ago. I decided not to go that route because I didn't want to install 16 more Neutrik connectors in the back of  all my TD-1's and rewire them for Tri Amping, since I was using them in pairs per side.

[tromprof]

During the weekend I decided that the reported price on the K-402 was to good to pass up, so I would order a pair and store them until I could start a build later this year, but on Monday I received an e-reply from Spencer Chao that the K-402 was no longer available for sale as a horn only. 

 

Couldn't afford a pair right now any way after build QPs six and seven... but still.

Edited January 26, 2016 by tromprof

[ClaudeJ1]

 

During the weekend I decided that the reported price on the K-402 was to good to pass up, so I would order a pair and store them until I could start a build later this year, but on Monday I received an e-reply from Spencer Chao that the K-402 was no longer available for sale as a horn only. 

 

Couldn't afford a pair right now any way after build QPs six and seven... but still.

 

Well, it's not like your system sucks or anything like that, geez. Yes, folks, I heard that one too, but I"m biased!

[ClaudeJ1]

SL Synergies is perhaps more for marketing than something that actually makes a significant difference in listener preference.  I can see that the measurements in the 470 Hz crossover region are almost unaffected (relatively speaking to multiple-way vertically-separated-horn designs) with the crossover filter tweaks and HF channel delay. 

 

 

Well I do have measurements somewhere, but not of both at the same time.  Here's are Tom's SH-50's after using the Mini DSP with a convolver. 

[DrWho]

If you can't describe something with an equation, then how do you know you understand the behavior?

 

That's not to imply that having an equation implies understanding (ala Geddes, ha) - but how can you have a complete understanding without any equation? Partial understanding or a gut feel, sure - but a fundamental understanding?

 

Chris made this point earlier in the thread:

I think that a lot of people make the mistake of believing simulations over real life a great deal. In my experiences I've seen otherwise bright engineers make that mistake periodically over my engineering career, often to the derision of the more experienced engineers that knew better (...from the school of hard knocks).  Fortunately, the enterprises that I worked in were guided by these most experienced engineers and those issues of over-belief in simulation were mitigated quite early before other less recoverable decisions were made from those simulation results.

I totally understand and agree with the overall point....but I think a more complete philosophical approach is to simply improve upon the models where the real world performance doesn't match our expectations. The bench should absolutely be our reality, but that doesn't mean we can't more precisely predict what the bench will show us. And by precise, I don't mean calculating unnecessary precision, but providing metrics by which we can make informed decisions about the inevitable compromises in every system.

 

Empirical approaches are certainly valid, but analytical solutions (verified by the bench) are far more repeatable and allow for way higher levels of precision. I also think much deeper insight is gained when deriving the analytical solution - which is the point I wanted to make at the beginning of this post.

 

One other thing to consider about an empirical approach: what happens if something is built wrong? Do you have enough intuition to determine that the unexpected results were in fact unexpected? Or do you conclude some unknown behavior is responsible and leave performance on the table? To date, I've never seen a system that couldn't be described analytically, and I've never seen it to be impossible to match a proper analytical expectation. I think it too often that people want to believe in magic, or want to believe something is completely unexplainable.

 

Anyways, just some philosophical meanderings since it's been brought up a few times in this thread.

[DrWho]

I hear very little difference when I make changes to the crossover filter or the delay.

 

Can you describe the magnitude of some of these changes? Are we talking 100us, or 3ms?

 

I would be interested to see a plot showing the final EQ + xover of the LF section, the final EQ + xover of the HF section, both overlayed on the combined system response. Like this:

 

You've posted a lot of plots and there seems to be consistently a dip right around the xover range (or peaks surrounding the xover - however you want to think of it). I'm curious if this is because of how you EQ'd the individual driver responses before implementing the xover?

 

And have you read this article?

http://www.rane.com/note160.html

 

Do you see the expected changes in phase response when you change the xover types? Going to a butterworth with the same alignment should give you +3dB at your xover frequency (475Hz, right?).

 

Btw, I know it's minutia stuff....just trying to understand how you don't see any changes when changing the filter. I've never seen that behavior before.

Edited January 27, 2016 by DrWho

[Marvel]

Thanks Mike, I actually understood what you posted.

Doesn't the acoustic crossover point (caused by the ports and port placement) cause another 3db or so drop at the crossover point along with the electrical? Trying to wrap my head around all this.

Bruce

Edited January 27, 2016 by Marvel

[DrWho]

Thanks Mike, I actually understood what you posted.

Doesn't the acoustic crossover point (caused by the ports and port placement) cause another 3db or so drop at the crossover point along with the electrical? Trying to wrap my head around all this.

 

I think the best answer to that question is "it could". The thing is, if you EQ the response back to flat before implementing the xover, then the acoustic xover effects get nulled out and you're back to dealing with a classic filter response.

 

You can also design the ports such that they're aren't introducing an acoustic filter near the xover frequency.

 

Here's another interesting article:

https://www.trueaudio.com/st_mr2.htm

The last section talks about "constant-voltage" filters - definitely an interesting idea.

Edited January 27, 2016 by DrWho

[Chris A]

Can you describe the magnitude of some of these changes? Are we talking 100us, or 3ms?

 

I inserted 1.3 ms of delay into the HF channel, ostensibly to offset the 24 db/octave (L-R) crossover filter phase delay - HF to LF. 

 

I should have seen a null at 475 Hz due to the 180 degrees of phase shift at the crossover frequency (1/475 is about 2.1 ms).  I didn't see that at all.  In fact, I saw an overall phase flattening of the entire HF channel response wrt the woofer channel (which is good), and a very uncharacteristic flattening of the phase and amplitude response around 475 Hz.  So I left it alone until can I return to investigate the crossover settings a bit more.  I can't hear the changes in crossover settings due to these changes, however, like I can with the Jubs on either side.  I believe that I have a reasonable ability to hear those things, but I didn't. 

 

I can only assume that there is some acoustic integration going on inside the horn itself (something that Danley and others have claimed) that you simply don't get with a vertical stack of horn mouths playing into free air in front of the loudspeaker, lobes and all.  I don't have any evidence that there are lobes being formed outside the mouth of the horn. 

 

The polar plots that I made above indicate to me that this horn has extremely stable polars vs. frequency once you control for microphone and loudspeaker ground bounce nulls in the data.  My ears also tell me that this loudspeaker arrangement has extremely consistent polars/directivity during sweeps and as a result of the now-extensive listening trials that I've been through with it.  It seems to behave a lot more like a small box direct radiator in that regard (like a Yamaha NS10 or its predecessors) with very close packing of its drivers - within 1/4 wavelength at crossover frequency. 

 

I'll try to run the three plots (LF, HF, summed response) of the crossover region (like you show above) the next time I do any more REW measurements, and post the results. 

 

Chris

Edited January 27, 2016 by Chris A

[ClaudeJ1]

some acoustic integration going on inside the horn itself

 

That was the whole point of the original TD-1 USA USA patent. Unity Summation Aperture. EVen though the xover was much simpler then.

[Chris A]

That was the whole point of the original TD-1 USA USA patent. Unity Summation Aperture.

 

See http://www.google.com/patents/US6411718 (now expired). 

 

That's what clued me into my conclusions--about what is happening at the crossover passband.  I don't believe that most people have experienced this effect.  I assume also that this is the saving grace of using multiple compression driver manifolds to increase SPL (like the TSCM has, among others).  What happens inside the horn or its manifolds leading to the horn simply get stripped away - they don't propagate. 

 

This effect is talked about in several places, but until you experience it, it doesn't occur to you what is actually happening.  "What happens inside the horn tends to stay inside the horn".  The only exceptions seem to be harmonic distortion products and higher order modes (HOMs) above the so-called "cut-on frequency" (related to the frequency corresponding to the horn's throat diameter).  The effect of having off-axis ports on the horn with the compression driver throat-located port do not show up outside the horn.  You can apparently even have asymmetric port locations in the conical/flat-sided horn, and those asymmetries don't show up outside the horn (see "A Bookshelf Multi-Way Point-Source Horn" thread on diyAudio for more evidence on that subject).

 

Chris

Edited January 28, 2016 by Chris A

[ClaudeJ1]

You can apparently even have asymmetric port locations in the conical/flat-sided horn, and those asymmetries don't show up outside the horn

 

That would be subject to further listening or measurement for sure. I prefer symmetry when possible. 

[NBPK402]

 

In addition, the tonality that you refer to is what I've identified as full mid-bass point source and directivity performance (down to 175 Hz) which is extremely important for the listening effect that you describe.  This is something that is not present in any other loudspeaker that I've heard--except perhaps planars (...but planars have their own set of issues that get in the way like issues with coverage, polars, and compression that Toole has described that have shown to be unfavorable by listeners in blind listening trials).

 

Before I bit the bullet and simply purchased Sh-50's for my LCR, Tom Danley had me measure each driver section separately in my TD-1's and provided me with some active PEQ settings that would make the older boxes sound as good as his new Synergy ones. He claims he was still learning how to do better Xovers 15 years ago. I decided not to go that route because I didn't want to install 16 more Neutrik connectors in the back of  all my TD-1's and rewire them for Tri Amping, since I was using them in pairs per side.

 

According to this review they are already setup for active or passive...an I missing something here?

 

http://www.livesoundint.com/lab/lab/messages/archive4/12463.html

[ClaudeJ1]

According to this review they are already setup for active or passive...an I missing something here?

 

Yes you are. Active version was only 2-way and still had passive inside. I have the original schematics (there were 3) and one of those actually had a small error that I discovered. The active tri-amping I speak of was from private emails with T.D. and he worked out the custom settings for me. I may still do the necessary mods at some point, but for now, my 4 TD-1's are strictly for HT surround use, bounced off a flat wall for wider dispersion.

 

I love OVERKILL, don't you?

[NBPK402]

 

According to this review they are already setup for active or passive...an I missing something here?

 

Yes you are. Active version was only 2-way and still had passive inside. I have the original schematics (there were 3) and one of those actually had a small error that I discovered. The active tri-amping I speak of was from private emails with T.D. and he worked out the custom settings for me. I may still do the necessary mods at some point, but for now, my 4 TD-1's are strictly for HT surround use, bounced off a flat wall for wider dispersion.

 

I love OVERKILL, don't you?

 

I am debating on which way to go (later this year)....Klipsch K402 (like the thread starter), Danley SH50, or Danley TD1 for my front 3 channels.

[Chris A]

One of the issues that I believe existed with the early Unity horn designs was that they had "nulls" or dropouts in their passive-crossover-designed passbands, essentially because of the acoustic low-pass effect that wasn't fully recognized until later when the FR of the speakers (non-smoothed response, that is) was posted by someone else.

 

At that point, it sounds like Danley looked at positioning the midrange and woofer ports closer to the throat to raise the low-pass frequencies of both the midranges and woofer--but especially the midranges, where the FR dropout nulls were audible during upsweeps.  Once that was done, and the passive crossovers redesigned to (...my guess here...) flatten the phase and redistribute the electrical crossover points to something that matched the internal acoustic crossover points, then suddenly, they're "synergy horns".  No change in the concept itself itself except perhaps recognizing first invention's acoustic low pass effects more carefully. 

 

By that time, Danley had changed employers.  The problem is--the second patent apparently overlaps the first patent (i.e., will not stand up in a court of law), so the only thing that DSL did was to try to patent their "improved" crossover design (...which they curiously didn't add to the patent or describe in the patent...)

 

Good luck with that...

 

What I'm saying is, "much ado about very little".  Assuming that you can set up your unity horn to match the as-implemented low-pass frequencies of the unity horn midranges and woofers, then you've still got a unity horn...that sounds a little better, but not a whole lot better.  You can probably hear the drop outs in the original Unity horns during an upsweep.  If you CAN hear them, then you can also hear the time delay nulls during upsweeps in passively crossed Khorn, La Scala, Belle, Cornwall, and Heresy, too, because they also have crossover nulls.

 

One more point:

 

Note that this thread is about the New Center design--not something else.  It also isn't really about patents: it's about the outstanding performance that I'm seeing and hearing with the DIY multiple-entry K-402-based full-range horn. 

 

I invite you to start a separate Danley thread...if that's the focus that interests you, so we can keep this thread focused on the DIY aspects of the present design and clear of aforementioned Klipsch forum-hosted marketing efforts by other manufacturers.

 

TIA.

 

Chris

Edited January 27, 2016 by Chris A

[ClaudeJ1]

I invite you to start a separate Danley thread...if that's the focus that interests you, so we can keep this thread focused on the DIY aspects of the present design and clear of aforementioned Klipsch forum-hosted marketing efforts by other manufacturers.

 

I agree with this notion, as all threads have a tendency to get derailed, although, since Danley's evolution of ideas borne out in a large choice of products, those ideas are still a huge part of your implementation in a K-402, it's pretty hard to determine when to hit the reply button or not. That being said, I'm very interested in realizing YOUR well researched, simulated (and actually BUILT!) implementation in my wooden K-402's when they become a reality. So I will attempt to not tread on this thread.

Edited January 27, 2016 by ClaudeJ1

[Chris A]

No, no, don't misunderstand, I'm not calling foul.  I'm just saying that we were getting close to the edge of a cliff, so to speak, and I didn't want to call foul in the near future.  The discussions on the history of the multiple-entry horn design is definitely in-bounds, and I'm enjoying that history.

 

Chris