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

[Chris A]

The rationale for posting the picture is to illustrate only that a woofer compression volume "phase plug" isn't always used.  However, since I placed it on my list of things to do, I intend to address that issue at some point.

 

More to read on that subject...with regard to midrange drivers: http://www.diyaudio.com/forums/multi-way/88237-suitable-midrange-cone-bandpass-mid-unity-horn-48.html

 

Here is the front view of an SH96:

 

Chris

[DrWho]

4) I addressed that above. When I get to the point where I have a horn to spare, I'll try cutting ports.  I believe that Danley is using ports that are tuned to the region around 100-200 Hz, just looking at the pictures and doing the simple calculations.

 

Oh my, I wasn't thinking of putting the ports in the horn. I was thinking they could go in the side of the cabinet surrounding the horn. It'd be pretty easy to plug any holes you made, or you could just rebuild the entire enclosure for way cheaper than buying a new K402.

 

Do you think there is an advantage to putting the ports in the horn? I would expect the acoustic loading of the K402 to be nearly negligible below ~60Hz or so (for the same reason the driver excursion unloads so quickly). And for the same air velocity reasons, I'm not sure we want that much port chuffing in the horn? (The port would be much worse than the driver itself).

 

Btw, have you seen this thread on a cardoid 15" driver solution?

http://www.avsforum.com/forum/155-diy-speakers-subs/1764913-forward-radiation-sickness-big-waveguide-cardioid.html

 

There's a camp that doesn't like this approach because you're throwing energy away to accomplish the directivity, but I wonder if the port couldn't be tuned (flared) in such a way as to have more control over the cardoid pattern. I think it'd be interesting to extend polar control below the horn cutoff. I believe EAW is doing this in their Anya cabinets - and it's really as "simple" as putting a hole somewhere in the side/rear of the cabinet. The hard part is knowing how to design those holes.

 

I guess I'm just intrigued by the possibility of a full range system with "perfectly" controlled polar response....and you're showing potential here of doing that with a simple 2-way speaker. I've been looking for a better speaker setup for the sanctuary at my church and this strikes me as a potentially very cost effective solution. I haven't moved forward on it because the size/weight of the matching bass bin has been prohibitive.

[DrWho]

Btw, what is "psy octave smoothing"?

 

It looks like I'm on a really old version of REW....

[Chris A]

There's a discussion on AVS HTS forum somewhere - perhaps Google it.  The newer REW versions with the psy octave smoothing I believe rolled out mid-2015.

Edited January 15, 2016 by Chris A

[DrWho]

How many cycles do you have it set to? The default 15? That would explain why the plot looks weird to me - this is the first time I've seen this, but it's totally something I've wanted for a while. I've been manually doing this for a long time now. So cool.

http://www.hometheatershack.com/forums/rew-forum/99673-feature-request-frequency-dependent-windowing.html

 

Btw, have you made any xover plots? (Three plots total showing the individual response of each driver and the summed response together). I'm analyzing your data more closely and I'm noticing a "dip/spike" right around 500Hz'ish. I've seen this a lot when the delays aren't perfectly matched. What was your method for measuring the delay? Were you using the results calculated by REW? Or maybe it's a dip that wasn't corrected in the LF system? You can see the phase response bumps a bit there too.

[Chris A]

How many cycles do you have it set to? The default 15?

 

I've not changed the default.  That feature (FDW) apparently has no effect on the psychoacoustic windowing.  Currently psy-windowing is either on or some other method is chosen--binary.

 

Btw, have you made any xover plots?

I'm still playing with the delay and crossover filter type/order.  I have to say that there seems to be some sort of summation going on inside this horn--for the phase in the crossover region doesn't behave independently like that from separate horns summing in front of the horn mouths: It's like the phase gets resolved inside the horn.  I was going to vary the order and type of crossover filters, but I find that local SPL amplitude at the crossover band is really the only sensitivity that phase seems connected with.

 

I'm analyzing your data more closely and I'm noticing a "dip/spike" right around 500Hz'ish.

No, but there is a very, very narrow one around 174 Hz--less than 2 Hz wide.  Probably floor bounce or full-length-horn mouth bounce: I can't hear it in sweeps, but I can hear the higher frequency stuff of the K-69-A driver during upsweeps come and go.  It's really, really transparent now.

 

Chris

[ClaudeJ1]

  It's really, really transparent now.

 

I'm still blown away by the micro definition of SH-50's, expecially in the midrange, which is 4X the surface area of a TAD 4002. The bass and treble also have the same definition. I'm planning on going down a similar path as yours with my future wood 402's but it will be tough to beat or even equal the Danleys, for sure!

Edited January 15, 2016 by ClaudeJ1

[Chris A]

Do you think there is an advantage to putting the ports in the horn? I would expect the acoustic loading of the K402 to be nearly negligible below ~60Hz or so (for the same reason the driver excursion unloads so quickly). And for the same air velocity reasons, I'm not sure we want that much port chuffing in the horn? (The port would be much worse than the driver itself).

 

 

I think that it's marginal. I figure that you'll get about an octave of a few dB more output using those reflex ports in the horn, but I still wonder about the increased diffraction of four more really large holes in the horn--especially so close to the mouth.

 

The Danley Synergies do it--clearly because it's expedient to do it in the horn, i.e., they're non-hi-fi-specific loudspeakers (PA loudspeakers) so the customers probably just want coverage and SPL with clean enough performance. If I was selling in that marketplace, I'd put the ports in the horn, too, so that you could array them together in a close-packed configuration.

 

One of the minor reasons why I went to dual 15" woofers instead of dual 12s is that I could avoid using reflex ports by having twice the woofer driver piston area of the SH50s or SH60s, etc..  Given that sufficient HF frequency response and phase performance isn't compromised with the compression ratio used (a little less than 10) the woofers' HF performance is not an issue in the New Center configuration. It's just a cleaner, simpler approach, IMO.  The price of using dual 15s instead of dual 12s is pretty small, indeed. 

 

The problem with the Danley 50- and 60-degree square coverage designs is that they are going for so much more SPL for PA use that they can't use a single compression driver on the throat, and have to use midrange drivers between the woofers and the apex compression driver. That's a lot of clutter and complexity to do that, both in terms of mounting them all on the horn with penetration ports in the right place, and the resulting passive crossovers, Recall that Danley is selling to the PA community, and that customer segment doesn't want active crossovers when mounting their units in remote locations on top of poles and suspended from arena ceiling trusses.  I'm not sure that manifolded compression drivers wouldn't also be viable for them, such that they could get that real estate back on the sides of the horn for bigger woofers and drop one crossover set, and also put the woofers in the right place instead of having to cross them at a lower point due to driver-driver mounting interference.

 

Port chuffing hasn't been an issue in my listening trials with the New Center configuration, perhaps because I'm not trying to use them at 120 dB at my listening position. I've actually put my head inside the horn while playing bass-heavy music, and I've detected no issues.  The effect that Danley talks about--not being able to localize the source of the sound inside the horn except at the apex--that also is true for the New Center configuration.

 

Chris

Edited January 16, 2016 by Chris A

[Chris A]

Btw, have you seen this thread on a cardoid 15" driver solution? http://www.avsforum....e-cardioid.html There's a camp that doesn't like this approach because you're throwing energy away to accomplish the directivity, but I wonder if the port couldn't be tuned (flared) in such a way as to have more control over the cardoid pattern. I think it'd be interesting to extend polar control below the horn cutoff. I believe EAW is doing this in their Anya cabinets - and it's really as "simple" as putting a hole somewhere in the side/rear of the cabinet. The hard part is knowing how to design those holes.

 

This is something that certainly has applicability for smaller loudspeaker designs than full-up unity full-range designs.  I've got to spend a little time plowing through that discussion in a bit more detail.  I've got my hands full right now.

 

Chris

[Chris A]

I guess I'm just intrigued by the possibility of a full range system with "perfectly" controlled polar response....and you're showing potential here of doing that with a simple 2-way speaker. I've been looking for a better speaker setup for the sanctuary at my church and this strikes me as a potentially very cost effective solution. I haven't moved forward on it because the size/weight of the matching bass bin has been prohibitive.

 

Clearly, I understand.  The real question is where do you want your LF crossover or LF cutoff point.  In a large room like a church sanctuary, with a lot of air between the loudspeakers and the pews, having sufficient excess woofer surface area capability is of some interest.  If you're really interested, I believe that it might be possible to use two 18" woofers or four 15" woofers on the current horn, but the enclosure with have to be more rectangular to accommodate them than the KPT-305 case.  I'd actually lean toward four 15" woofers (like Danley did for the SH96, because I think the real-estate is there to do it.  You actually have a lot more latitude on where to put the woofers because you don't have the midrange drivers (...the SH96 uses 6 midrange drivers...) to get in the way.

 

I don't believe that you'll need special high output compression drivers for achieving realistic wide-bandwidth output, especially if you consider the dual-diaphragm BMS 4592ND or something similar  (...unless you hold rock concerts in your church sanctuary ). 

 

Chris

Edited January 16, 2016 by Chris A

[Chris A]

One of the other things that I wonder about is the location of the woofer ports relative to the center of the woofer cones. 

 

The Danley designs typically use the edge of the cones for a lot of their port locations - which at high SPL/cone motions will likely put a tilting moment on the cones themselves, leading to increased harmonic distortion, 

 

The ports on the New Center configuration are much closer to the center of the cone itself, and I can't help but believe that at higher SPL--like kick drum transients--that the performance of the 15" woofers is a bit cleaner due to avoidance of "rocking modes" and required axial motion of the 12" woofer cones used on most of the Danley designs. 

 

YMMV.

 

Chris

Edited January 16, 2016 by Chris A

[ClaudeJ1]

Good stuff, Mike and Chris. I think this a great direction to take. In the hopes that there will be some interest in this related subject, I'd like to add my own experience in my journey with actual production Unity and Synergy horns.

 

The Danley's SH-50's and the 12 year old Sound Physics Labs Tom Danley 1's (SPL TD-1's, patented in 2002, but designed 2-3 years earlier) share the same 4 midrange units, with 50 square inches of area, which is part of what defines their incredible midrange definition on female vocals, piano, saxophone, guitar, etc. from 300-1,000 hz.

 

So, I'm quite surprised to read that that there is a satisfying performance in the all-important midrange from 300-600 Hz. from woofers that have a mass rolloff that only slightly better than K-33 woofers, which roll at about 180 Hz. This is why  I will end up trying out Eminence Kappa 15C's when I get to try this in the future. I guess this is a testimonial to the  K-69 or other 2" exit midrange/tweeter driver to take up most of the slack.

 

My journey to Synergy horns began with just one pair of used TD-1's in August of 2014. I had to drive from Detroit to Buffalo NY, to get them. I was just amazed at the incredible sound from such a small, 135 lb., 103 DB/Watt box when I simply threw them on the floor of a warehouse and hooked up an old Kenwood receiver playing CD's off a $39 DVD player. The imaging from these was fantastic as well as having micro definition such as I had never heard before. When I raised them off the floor, the bass got really "thin" as one would predict. When I found and purchased another pair, butting them together to double the mouth area (like their literature said), the bass was back and the coverage went from 60x40 to 80x60 degrees, HxV. 

 

I have yet to do an A/B conparison between these two speakers. The tweeters are similar 1" design, the midranges are identical, and the woofers are both 12". The SH-50 is ported, while the TD-1 is sealed. The SH is rated at 100 db/W while the TD-1 is rated at 103db/W.

 

I did notice there may be some trace of resonance on some recordings with the ported cab, so I like the idea of keeping your New Center sealed, until I can prove that what I think I'm hearing, months apart, can be atributed to the sealed vs. reflex aspect.

Edited January 16, 2016 by ClaudeJ1

[Chris A]

See the figure below for the effect that I was referring to above.  PWK used that effect on the Khorn during its compression slot change used in the early 1960s.  The woofers used in this New Center design are of essentially the same T/S parameters as those used on the La Scala, Belle, Cornwall, and Khorn.  The compression ratio of the Khorn, IIRC, is about 3.  The New Center design uses a compression ratio of ~10..but with double the total woofer driver area of any of the Klipsch Heritage designs:

 

 

Remember from the comments above that the horn performance itself is insensitive to changes on the back side of the horn in the drivers or their enclosure volumes. 

 

What counts is freedom from harmonics and especially reduced cone travel to reduce non-linear effects of cone movement found in typical woofers.  Overhung voice coil woofers (i.e., low distortion) are not only extremely expensive (about $1K US), but they also have heavier cones due to the increased copper and voice coil support areas, as well as a bit higher inductance due to the increased copper lengths.  But if you can reduce cone motion through any sort of horn loading...however small...you get dramatically reduced modulation distortion benefits, but using conventional drivers at much lower cost.

 

The New Center woofers benefit from the horn loading at frequencies below that of 1/4 wavelength horn length, just looking at the uncorrected FR curve--down to about 60 Hz.  Beyond that point, corner (1/8th space) or quarter space loading reduces the required cone motion significantly.

 

I guess that you could look at the New Center design as a new La Scala or Belle due to this effect--but also using a "corner or a wall loudspeaker" capability to improve its extreme LF performance below 40 Hz and to capitalize on its increased woofer cone area over the Heritage designs, even more than a Khorn uses its 1/8th space loading (either through using a corner or using false corners).  The extra woofer piston area of the New Center is quite useful in that respect.

 

Chris

Edited January 16, 2016 by Chris A

[Quiet_Hollow]

The Danley designs typically use the edge of the cones for a lot of their port locations - which at high SPL/cone motions will likely put a tilting moment on the cones themselves, leading to increased harmonic distortion,

Two things going on there.

1. Not only is it a packaging technique, it's also a filter technique

2. It's a pressure region underneath there within the performance envelope of the design, so the load across the face of the driving element is uniform, even though it doesn't look like it at first glance. Symmetry is not required.

[Chris A]

Symmetry is not required.

For small motion...

[Chris A]

One thing that I find is implied but rarely stated clearly is the consequence of Fletcher-Munson effects on bass performance.  Floyd Toole recently (this year) commented in his video at McGill University (IIRC) about how sensitive loudspeaker performance is at the lowest frequencies due to the effect that small changes in SPL have on perceived loudness levels. 

 

The flip side of this is, and one that all bass players know intuitively, is that they never really play with the same dynamic range as higher frequency instrumentation.  That means that the bass performance of a loudspeaker is usually skewed toward higher SPLs relative to the higher frequencies.  So having a performance at higher LF capabilities is something of real interest when evaluating loudspeaker performance.  This is especially true of transients which often stress the loudspeaker's woofers disproportionately.  Having excess capability to respond to LF kick drum and bass drum transients is an advantage, I've found. 

 

So large signal excursions of woofers happen typically much more often (assuming they haven't been "mastered away" as they are on almost all stereo recordings) than higher frequency SPL transients (...think of a kick drum bass line).  With fully 30% of loudspeaker performance found in its bass performance, it is important to acknowledge that design for large signal behavior is fully supported--as it is in the New Center woofer design.

 

That's why I made the comment about off-center port locations.  Much of the time, the woofers are not behaving under a typical "pressure zone" conditions.  I find that there is unexplored design space for improvement from horn-loaded loudspeaker design practices that are found in the Klipsch Heritage line, i.e., more woofer piston area.

 

Also, it's important to separate concerns when talking about large signal behavior--in terms of bass reflex performance vs. closed box. I first found this out when I first explored boosting the extreme LF end of the Jubilee bass bin below 40 Hz (thereafter suggesting it to Roy)...without apparently audible penalty.  The Jubilee bass bin is closed box direct radiator that is loaded by a horn.  Since the Jubilee bass bin uses an exponential horn expansion, it loses horn gain performance much more rapidly than a conical horn does.  By boosting the Jubilee bass bin EQ below the cutoff frequency, another 10 Hz of LF performance was effectively gained (remember Toole's observation above) without penalty, based on the performance of the woofers in a closed box configuration (i.e., sounds much cleaner), but using room corner gain to support sub-Fc performance.

 

Regardless of many who talk about bass reflex performance, I've found that it detracts from the cleanness and tightness of sound that I experience with both horn-loaded bass and, largely, closed-box bass cabinets--assuming sufficient woofer area to control cone movements below objectionable modulation distortion levels.  If the total woofer piston area is sized in a closed box configuration to compensate for the added "free" SPL that one gets by designing instead of a typical bass reflex box and ports, I find that much of the objectionable bass flabbiness is avoided.  So I separate the concerns of direct radiator performance in closed boxes with those of bass reflex enclosures--within the audible bass regions.  Below 30-40 Hz, the direct audibility of bass (versus tactile feel) starts to become dominant, such that different configurations like TH subs can be used without significant audible penalty.

 

I tend to avoid bass reflex for the reasons cited above and instead rely on eighth-space or quarter-space room loading instead using the performance of closed box woofers below Fc.

 

Chris

[DrWho]

Chris, what are the dimensions of your the 15" ports into the horn? I think somewhere you mentioned a 10:1 compression ratio?

 

I've been going through your hornresp data and think I found a few bugs. For the 15" woofer section, you have the VTC chamber values set to 0. This turns off the effects of the throat port (Ap1 and Lp), which you have set to 46.2 and 10.5.....which is actually a relatively long narrow throat opening.

 

You also are modelling 4 drivers in parallel, but I think you mentioned only using two 15" drivers (The ME1 value should be 2P, not 4P).

 

And is the free volume of air between the K402 and your enclosure really 8cubic feet? I would expect to see a rear chamber volume closer to 5cu ft (144L)

 

I think you should be using the following values:

ME1 = 2P

Vrc = 144

Vtc = 4500

Atc =  590 (average length of 7.63cm to the port)

Ap1 = 178

Lp = 2.20 (3/4" + 1/8")

 

Do these values correlate to what you built?

 

Also, is the first conical length of 10.16cm going to the center of the port opening, or to the edge closest to the throat? Pictures of horns always create optical illusions so it's hard to tell from a picture. Is the K402 really only 18" long?

 

 

The reason I'm digging into the details is to try and understand why the measured response is so drastically different from the predicted response. Or to look at it another way, I'd like to understand why the predicted and and measured results are so different. I've gotten much closer with these adjustments, and along the way I noticed some crazy coupling effects between the LF and HF units. Have you noticed that the Power2 result changes when adjusting the amplifier power sent to Driver 1? You have to crank the voltage on Amplifier 2 and cut the voltage on Amplifier 1 to simulate driving only a single driver - and then the results look very similar to what you're measuring. Likewise, you can raise the output impedance of Amplifier 1 and look at Displacement 1 to get an idea of how much the woofer section moves the tweeter diaphragm.

 

Attached are pictures showing what I'm simulating. It's pretty close to your measurements except for the behavior below 300Hz. I think that's because the aperture of the horn doesn't meet the assumptions hornresp is using in its model....I've frequently seen the bottom octave in hornresp not line up to the real world. This  seems consistent with what I've seen in the past. And I think your extra output right around 300Hz is due to the directivity of the horn.

Edited January 16, 2016 by DrWho

[DrWho]

Also, adding a phase plug to your particular design won't be advantageous....it just creates a larger spike that you'll end up attenuating with EQ (and we'd rather not have that acoustical resonance in the system).

 

If you add a phase plug, then the port opening needs to move closer to the throat.

 

Attached are two simulations showing that.

[kg4guy]

Maybe next week I'll have a chance to measure the woofer response on my SH96HO's if it is a help ..

Edited January 21, 2016 by kg4guy

[DrWho]

Regardless of many who talk about bass reflex performance, I've found that it detracts from the cleanness and tightness of sound that I experience with both horn-loaded bass and, largely, closed-box bass cabinets--assuming sufficient woofer area to control cone movements below objectionable modulation distortion levels.  If the total woofer piston area is sized in a closed box configuration to compensate for the added "free" SPL that one gets by designing instead of a typical bass reflex box and ports, I find that much of the objectionable bass flabbiness is avoided.  So I separate the concerns of direct radiator performance in closed boxes with those of bass reflex enclosures--within the audible bass regions.  Below 30-40 Hz, the direct audibility of bass (versus tactile feel) starts to become dominant, such that different configurations like TH subs can be used without significant audible penalty.

 

This is probably totally off topic, but I think it is possible to design a bass reflex that doesn't have that flabbiness sound. Or maybe we have different thresholds/definitions for flabbiness.

 

The biggest variable from an IMD perspective is an appropriate high-pass filter to address the woofer unloading below the tune frequency. And if you have active xovers in the mix, then you should implement a peaking high pass filter to take advantage of more free output - which generally means you can get away with a lower tune frequency (which also makes it sound more like a sealed cabinet over a wider bandwidth). Then your port needs to be properly sized (most are way undersized), and it should be flared. The last remaining variable is the pipe resonance of the port (which is lower Q with the flare), which can be address with careful placement inside the speaker enclosure (providing 1/4 wavelength cancellations inside the box). Then you need appropriate absorption inside the cabinet to keep high frequency energy from getting out through the port (along with attenuating all them internal standing waves).

 

You just won't find that recipe with the Klipsch pro cabinets.... The Palladium lineup is probably the closest Klipsch gets to that. Many other pro cabinets will intentionally undersize their ports so that at high SPLs, the port compresses a lot helps prevent the woofer from unloading too much (it starts to approximate a sealed enclosure). Not a good sound, but at least it protects the voice coils from slamming against the pole piece. Before this happens, you often get a frequency response that changes with output level - and that change looks like a narrow Q boost, which will tend towards one note sounding louder than others. A lower Q design (low group delay) with a full sized port doesn't tend to exhibit these effects.

 

I only bring it up because I've done a lot of side-by-side vented versus sealed comparisons, and when the two systems have the same frequency response, the vented system has way lower IMD....even when measured in room.

 

I'm not saying you have to port your cabinet here, but 16dB of boost to get 60Hz to match the output at 300Hz is a lot of EQ...

 

 

Btw, I agree that you can double up on drivers and get even better performance all around....but like Roy talks about with horns, you can always add ports to that total system.

Edited January 16, 2016 by DrWho