Jump to content

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


Chris A

Recommended Posts

Okay, before we get on to the Hornresp parameters, I need to say the following to the readers here. Bentz already understands the following, I believe.

 

There were only two decisions that I had to make that I used Hornresp to help me with:

 

1) the position of the through holes into the K-402 horn (i.e., NOT the size of the holes nor their shape, length or lateral position on the horn walls equidistant from the throat), and,

 

2) which woofers that I should use--including the nominal woofer diameters and approximate T/S parameters in order to get reasonable performance.  The choices were either: a) two 12" woofers, or b ) two 15" diameter woofers.  T/S parameters came with the choices in woofers considered from places like Parts Express and this forum (Crites and other K-33 models).  

 

That's it. 

 

I got reasonable answers to those two questions using Hornresp, and found how insensitive those answers were to tweaking other parameters (including the number of woofers used, etc., etc.), so I stopped spending more time on the "sim" at that point.

 

My design as envisioned worked.  I'd say that was fairly spectacular...perhaps luck, perhaps not.  I currently don't plan on doing many design tweaks because there are few tweaks that have much effect on the performance of the current design.  I'm sure that over time, there will be design changes that might be warranted, but thus far I haven't found any of those.

 

I saw the somewhat severe limitations of Hornresp to do much else, but getting approximate answers to the two questions above were critical.  I got those two answers, albeit with a little surprise that Danley never really advertised the amount of EQ he used below the 1/4 wavelength frequency of his horn designs.  Now I know...and I also found that discussion existed in somewhat disguised form elsewhere, mentioned cryptically in a diyAudio thread about 7 years ago.  No matter as it turns out. 

 

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. 

 

The same thing rings true in this case.  Having spent about half my career functioning in a role of how to "bound the problem" to check whether or not my own self-written simulations were giving me reasonable answers.  If they were, I typically spent no more time on simulations to get "higher resolution answers".  That is the case in this instance.

 

"Horses for courses"...as they say.

 

Now...to begin the process of addressing Hornresp modeling questions--which I believe is useful to the community here...I want to stress that any conversation about using Hornresp to get more detailed answers than the two above is likely not something that I'm currently interested in spending a lot of my time chasing. 

 

I posted my abbreviated Hornresp model for others to play with and perhaps fashion for their own needs.  I didn't have to do that, BTW.  So let's all be mutually supporting, if you don't mind terribly.  There is a reason why I am saying what I am at this point, and that reason is to not detract from the design implementation itself discussed in this thread.  If anyone wants to start a new thread on "blue skying" new multiple-entry horn designs in general--then go for it...but elsewhere, please.

 

Now for answering Mike's questions....

  • Like 2
Link to comment
Share on other sites

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?

 

When I get a chance, I'll take a look and let you know what the values should be.

 

 

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?

 

No, for the reasons stated above. I don't know if I'm interested in getting into sim verification and validation activities at this point, especially one whose source code I do not have access to.  I'd much rather spend time on a sim that gives better answers to the design of ports and their location, since that is the next item of optimization.  Hornresp did its job for the simple questions that I asked of it. 

 

There is yet another simulation that I have in mind (for a later date) that models boundary value problems that can give me much better answers.

 

...I think your extra output right around 300Hz is due to the directivity of the horn.

 

I believe that I knew that in advance of starting the fabrication, since I had pretty good KPT-305 data already.  Note that all the data that you see within Hornresp is power - NOT SPL.

 

adding a phase plug to your particular design won't be advantageous

 

Thanks for that -- I already found that any improvements in that area were of lower value than the areas I mentioned above.

 

This is probably totally off topic, but I think it is possible to design a bass reflex that doesn't have that flabbiness sound.

 

Perhaps start a new thread?

 

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.

 

Then I'm clearly not hearing IMD, if what you say is true.  I believe that the answer to this discussion is not IMD in this particular instance, but transient (time-based) behavior.  I use technical terms sometimes to describe what my ears are telling me: that is the case on this particular subject.  

 

I'd rather defer discussions on possible vented cabinets to another thread, since I've never heard one that I really liked yet.

 

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...

 

This is a valid thought, but when you consider the relative distortion measurements between the New Center, corner-located Jubs/Tad 4002s on either side, and the JuBelle that preceded it in its current place, I think your opinion might change--in fact I know it will. This is especially true of the subjective listening trial that I've been conducting over the past week or so.  I know that you'll be amazed by what your ears tell you over the typical engineering measurements discussed here.  The are interesting and revealing differences that have caused me to go back and look more closely at the other loudspeaker measurements in my setup.  I'm typically not the type of guy to make these kind of statements without significant findings.

 

I'm currently not interested in posting distortion measurements here...for reasons that might become apparent later.

 

Thanks for the QA on the Hornresp stuff, Mike.  I'll try to update the model for any future tweaks to further the current design and communicate those back to you and the forum members here.

 

Chris

Edited by Chris A
Link to comment
Share on other sites

Chris, do you think it was less expensive (parts-wise) to buy the K402 horns, and mod them than to just buy the SH-50s?

 

I missed this question from last Wednesday.  Sorry about that.

 

Clearly the answers depend on availability of the Danley SH50 loudspeaker and/or the Klipsch  K-402 horn on the used market,  I haven't seen SH50s advertised on the web used since about four years ago.  Perhaps Claude can give you tips on where to look.  I have seen a few K-402s advertised in ones and twos, used.  In general, I believe that K-402s are generally much cheaper. 

 

The new prices make your question much easier to answer: the K-402s are cheaper--about $1K including a stand and driver attached--or you might can buy just the horn.  However, there is a fair amount of fabrication/modification required.

 

If you buy it in a KPT-305 configuration, then the price will be about $1600 (each) plus $300 for the woofers (two per loudspeaker) and about that also for a reasonable 2" compression driver. I've seen used TAD 4002s for about $600, which is a screaming deal. 

 

You'll also need a digital crossover.  Those go for about $400 (a typical used Dx38 price), and work extremely well. The Xilica XP2040 would be even better (24/96), sometimes available used.

 

The real objective it seems is to have a source for the horn with woofer mounting pads and through holes, beefed up to make it much stiffer--as a single-piece part.  You need something that is the correct geometry and extremely stiff,  and second, something that might or might not be "pretty".  I wouldn't put money into a horn that isn't much stiffer than a stock K-402--made in pretty wood.  I believe that would be a large mistake--acoustically speaking.  The box is an easy shape to make--it's just a box that needs to be fairly stiff and air tight. 

 

I'd much rather have a K-402 based horn than a Danley SH50 or SH60 because the coverage angle differences are a pretty big deal, IMO.  You'd need a pair of Danley SH96s to match the coverage of two horns of the current type here.  That's a lot of money (i.e., $22K US).  I think that a stereo pair of K-402-based units would be a lot cheaper, even with the best drivers that money can buy.  In fact, I believe that a pair can be easily had for significantly less cost than a new pair of La Scala IIs, depending on drivers used...using all-new stuff.

 

If you're ever through the D/FW area, you have an open invite to come hear.  Perhaps I'll have more than just one unit set up by then. 

 

Chris

Edited by Chris A
  • Like 1
Link to comment
Share on other sites

Chris,

Would this idea work on a tractrix horn, one designed to have a low Fc? I was working with Erik Forker's spreadsheet at a larger horn. His, like Al's and Dave Harris' have two flat sides (top and bottom) which might make it easy to mount the woofers.

 

If this should be in another thread I will start one.

 

Bruce

Link to comment
Share on other sites

 

Chris, do you think it was less expensive (parts-wise) to buy the K402 horns, and mod them than to just buy the SH-50s?

 

I missed this question from last Wednesday.  Sorry about that.

 

Clearly the answers depend on availability of the Danley SH50 loudspeaker and/or the Klipsch  K-402 horn on the used market,  I haven't seen SH50s advertised on the web used since about four years ago.  Perhaps Claude can give you tips on where to look.  I have seen a few K-402s advertised in ones and twos, used.  In general, I believe that K-402s are generally much cheaper. 

 

The new prices make your question much easier to answer: the K-402s are cheaper--about $1K including a stand and driver attached--or you might can buy just the horn.  However, there is a fair amount of fabrication/modification required.

 

If you buy it in a KPT-305 configuration, then the price will be about $1600 (each) plus $300 for the woofers (two per loudspeaker) and about that also for a reasonable 2" compression driver. I've seen used TAD 4002s for about $600, which is a screaming deal. 

 

You'll also need a digital crossover.  Those go for about $400 (a typical used Dx38 price), and work extremely well. The Xilica XP2040 would be even better (24/96), sometimes available used.

 

The real objective it seems is to have a source for the horn with woofer mounting pads and through holes, beefed up to make it much stiffer--as a single-piece part.  You need something that is the correct geometry and extremely stiff,  and second, something that might or might not be "pretty".  I wouldn't put money into a horn that isn't much stiffer than a stock K-402--made in pretty wood.  I believe that would be a large mistake--acoustically speaking.  The box is an easy shape to make--it's just a box that needs to be fairly stiff and air tight. 

 

I'd much rather have a K-402 based horn than a Danley SH50 or SH60 because the coverage angle differences are a pretty big deal, IMO.  You'd need a pair of Danley SH96s to match the coverage of two horns of the current type here.  That's a lot of money (i.e., $22K US).  I think that a stereo pair of K-402-based units would be a lot cheaper, even with the best drivers that money can buy.  In fact, I believe that a pair can be easily had for significantly less cost than a new pair of La Scala IIs, depending on drivers used...using all-new stuff.

 

If you're ever through the D/FW area, you have an open invite to come hear.  Perhaps I'll have more than just one unit set up by then. 

 

Chris

 

Who told you the SH96's were $22K LOL there not much more than what Klipsch advertises Klipschorns for on there website .

  • Like 1
Link to comment
Share on other sites

Since you bring it up, I assume that it would take you about 4-5 characters.  What is the price? 

 

Unfortunately, it means very little to me. I'm pretty sure that wouldn't want SH96s for other reasons related to their acoustic performance in a hi-fi home environment.

 

:rolleyes:

 

Chris

Link to comment
Share on other sites

Chris,

Would this idea work on a tractrix horn, one designed to have a low Fc? I was working with Erik Forker's spreadsheet at a larger horn. His, like Al's and Dave Harris' have two flat sides (top and bottom) which might make it easy to mount the woofers.

 

If this should be in another thread I will start one.

 

Bruce

 

It's okay to talk about why I made the choices that I did, I believe, among other related discussions.  I'm not trying to curtail discussion but perhaps only those discussions to subjects that really aren't discussing the subject of this thread.

 

The real secret behind the multiple-entry horn is the conical horn section itself--not tractrix: tractrix is really only useful for the mouth area expansion.  If you had a K-402 in your hands, and measured it, then constructed  a tractrix curve of the same mouth dimensions (it's pretty easy to do that - the word "tractrix" means "to pull" in Latin), you'd see something fairly startling.  I'll let you munch on that for a moment.

 

So if a multiple entry horn requires a large area of the horn expansion to be conical (Roy calls this "straight-sided"), for reasons having to do with its polars vs. virtually limitless frequency response above the quarter wavelength frequency (something that distinguishes the K-402 from the rest of the crowd of horns and waveguides), then having a true tractrix horn would actually mess that up. 

 

The reason for that is that high frequencies act like light rays--they don't go around the curvature of tractrix, exponential, or hyperbolic horn walls, etc.  That's the real contribution that Geddes makes clear: while the Webster area expansion horn 1P equation says that geometry doesn't matter, the more correct equations governing waveguides (read: horns) says that horn "line of sight" geometry does matter, not just area expansion rates.  Roy knows this, too.

 

If you were to construct a tractrix horn with the same mouth dimensions of the K-402, it would be more than three times longer.  I don't know about you, but I'll never have room for that or any other midbass horn using a full non-conical expansion profile.  It's as simple as that. 

 

The K-402 works because it works in its extremely short length, and actually has better acoustic properties than a true tractrix curve in terms of coverage angles vs. wide frequency, and distortion properties at higher SPL due to rapid throat expansion rate.  It's the same thing as PWK's "rubber throat" discussion.

 

By the way, having flat parallel walls in a horn is a no-no.  Read about higher order modes (HOMs) in the literature to understand why that is.  I wouldn't like to use a horn that has any parallel flat surfaces, nearly flat parallel surfaces, or internal slots for reasons related to HOM generation and propagation.

 

Chris

  • Like 2
Link to comment
Share on other sites

Since you bring it up, I assume that it would take you about 4-5 characters.  What is the price? 

 

Unfortunately, it means very little to me. I'm pretty sure that wouldn't want SH96s for other reasons related to their acoustic performance in a hi-fi home environment.

 

:rolleyes:

 

Chris

Sorry can't advertise prices on the web also there are plenty of DIY smaller Synergy horn builds available to build that the work is all done for you so if you wouldn't want a large Synergy horn like the SH96 why even go this route. :D

I can tell you for not much more than the price you put above you could get a pair of SH96's and a pair of DBH218's shipped to your door.

Sorry if this is a disruption I'll stay out of the thread after this. 

Edited by kg4guy
  • Like 1
Link to comment
Share on other sites

That's a lot more than 4-5 characters. 

 

Perhaps you should talk to your employer about not being able to talk about your product's critical information.

 

Chris

Link to comment
Share on other sites

By the way, Bruce, if you constructed a horn that had a mouth of 70 x 42 inches, and was 34 inches long, with the same scaled geometry as the K-402 (albeit with the same 2" throat), you could have just about everything that you'd want in a full-range multiple entry horn with spectacular performance in almost every way that you can measure or hear subjectively--down to 87 Hz without EQ leveling on the woofers, with corner loading below that point and baffles that extended to the side walls and probably toward the ceiling, too. 

 

Getting it shipped and through your doorways might be an issue, however.  Who wants a pair?

 

Chris

Edited by Chris A
  • Like 2
Link to comment
Share on other sites

KPT_942T_-_Angled_635284153385055000_med

 

The above configuration is quite similar to the New Center configuration, albeit in a Cinema-Professional package and with what I'd call "flabbier' reflex bass. 

 

I've heard the above configuration A-B'ed against a Jubilee, and I've heard the Jubilee A-Bed against the New Center configuration.  There is a lot of difference in the sound of the New Center.  I intend to build at least two more of the New Center configuration as a result of those listening trials. 

 

Chris

  • Like 1
Link to comment
Share on other sites

By the way, Bruce, if you constructed a horn that had a mouth of 70 x 42 inches, and was 34 inches long, with the same scaled geometry as the K-402 (albeit with the same 2" throat), you could have just about everything that you'd want in a full-range multiple entry horn with spectacular performance in almost every way that you can measure or hear subjectively--down to 87 Hz without EQ leveling on the woofers, with corner loading below that point and baffles that extended to the side walls and probably toward the ceiling, too. 

 

Getting it shipped and through your doorways might be an issue, however.  Who wants a pair?

 

Chris

If it could go down to 50Hz, and the shortest dimension was 30" (I could even handle it being more than 40" on the one dimension, but 70" on the other is max for me), I would like 3. :)

  • Like 2
Link to comment
Share on other sites

That's essentially what a Jubilee bass bin and a Khorn bass bin do, except that they're folded exponential horns.  Their mouths are more undersized than the K-402, which is the major reason why their frequency response has ripples--due to undersized mouths.  Of all the things that you can do to a horn, undersizing the mouth has the greatest effect on its performance (above its cutoff frequency). 

 

So if you want a longer horn with undersized mouth, I recommend a folded horn instead - like the Jubilee bass bin.  Straight horns of these wavelengths don't mix well with interior decorators--or movers.

 

If you instead want a horn without folding or bending the horn, point source of the entire loudspeaker, in a much smaller package (i.e., able to be used in as center channel easily) with Jubilee imaging performance (better imaging, really...), the New Center design makes more sense.  It also costs a lot less.

 

Chris

  • Like 1
Link to comment
Share on other sites

 

Chris,

Would this idea work on a tractrix horn, one designed to have a low Fc? I was working with Erik Forker's spreadsheet at a larger horn. His, like Al's and Dave Harris' have two flat sides (top and bottom) which might make it easy to mount the woofers.

 

If this should be in another thread I will start one.

 

Bruce

 

By the way, having flat parallel walls in a horn is a no-no.  Read about higher order modes (HOMs) in the literature to understand why that is.  I wouldn't like to use a horn that has any parallel flat surfaces, nearly flat parallel surfaces, or internal slots for reasons related to HOM generation and propagation.

 

Chris

 

 

I didn't say it had parallel walls. I said 'flat'. Only the side walls are curved.

 

I assume the formula for conical horns readily available?

 

Regarding the tractrix formula and name, I've had this paper for a long time. It gives some of the history.

 

Bruce

Steyn.pdf

Link to comment
Share on other sites

Thanks Bruce - that's an interesting article on the tractrix curve that I don't believe that I've seen before. 

 

So the construction of a true tractrix curve isn't very difficult to do.  Once you do one that matches half the horizontal mouth size of the K-402, then look at the K-402 midline-to-horizontal curve from measurements, it's quite easy to see that the horn can be called a "modified tractrix curve", since it really is a truncated tractrix construction, that makes about half of the length of the horn conical.  (The conical section walls are about as flat as flat gets.)

 

So you CAN call the K-402 a conical-tractrix horn, but there are other "mid-bass" DIY horn designs that use that same term to mean two walls are straight and two wall are tractrix.  Those are the type of mid-bass horns that I personally would avoid, and is the reason way I described the flat walls as problematic.

 

If you look at the TH subs that Danley has designed, you'll probably note that their horn expansion type is exponential, with a very, very low flare rate corresponding to the 1/4 wavelength cutoff frequency of the horn's length.  I've often wondered if the THD measurements that you see out of TH subs is partially due to the nearly parallel walls in its expansion, leading to HOM generation in the specific horn segments between the horn folds.  That's an example of nearly parallel horn walls that I was referring to.

 

[OT: I've often wondered what would happen to the harmonic distortion performance of a TH sub if it were constructed using continuous spiral expanding tubing anchored between two plywood sheets and using foam to constrain the tube-plywood matrix together, but anchored at the driver port from the throat-to-near mouth penetration points, thus eliminating the multiple hard folds in the horn path length.  Will the harmonics be lower order, and will the HOM generation be much lower due to the elimination of the horn fold discontinuities?]

 

Chris

  • Like 1
Link to comment
Share on other sites

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.

 

 

Remember that Danley advertises that coupling between the drivers in the conical multiple-entry horn as an advantage, not as a disadvantage.  I've seen that coupling in the crossover region with regard to phase.  It's actually quite nice to have, since I don't have to worry about using FIR filters to flatten the phase of the total loudspeaker -- I can cheat and add more delay to the HF section, using a 4th order L-R crossover filter (i.e., 360 degrees of phase lag is induced into the LF drivers relative to the HF compression driver).  I can get less than 360 degree of phase delay across the entire loudspeaker passband without using RePhase or something like that to apply the more expensive FIR filtering processing-which would require a much better digital processor capability in the active crossover.  I believe that Danley is doing the same thing with his Synergy designs and getting away with it. This is really interesting stuff.

 

 

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.

 

Okay, I may be having second thoughts on this subject, but I'd have to say that adding reflex ports would only be useful to those folks that don't have something to crossover to on the low end.  I've got a lot of LF capability as-is that I don't need the reflex ports.  But using a pair of these units along a wall might require a fair amount of bass channel "umph", so some people might not mind the extra sound of reflex porting that I'm personally most sensitive to.

 

Perhaps you could share your thoughts on that subject that I so quickly dismissed, above...?  My current thoughts include simply porting the current box design on the side or top with whatever current enclosure volume that they have--and take what they can give in terms of performance. 

 

Bill Waslo's "CoSyne II" design actually uses a very involved bass reflex box below the dual-flare conical horn, braced as if it were going to be doing subwoofer duty, but ported through the horn, as the current New Center design has done to get the coaxial midbass performance...which is a really, really big deal here I've found.  I think that Waslo's design is a bit overboard to my current thinking, but it also addresses another need: to get the horn axes up the ear level without having to use baffles to the floor--which was my current thinking in terms of using the current design in corners of the room without subs, etc. Thoughts?

 

CoSyne%20II%20reduced.jpg

Bill Waslo's CoSyne II design with bass reflex porting

 

BTW: I'm waiting for the wind to calm down and the temperature to rise to more comfortable 50s-60s F before I attempt outside testing of the polars of the horn.  It looks like Wednesday or perhaps later is when that weather will arrive to facilitate that.  I'm pretty sure what I'm going to see in the polars, so I've not been terribly motivated to bust my rear to do that testing.   However, I may spend a bit more time on the dual-tone testing until that time. The IMD levels are really what I've been sensitive to, not THD, I've found.  There seems to be a poor correlation to THD and perception of bass distortion, but total cone motion at higher midbass frequencies IS really, really important, as is the simultaneous controlled polars found across the woofer-compression driver crossover passband.

 

It's really been difficult to stop listening to this unit playing multi-channel recordings: I'm still amazed by the difference in what I'm hearing--even relative to the Jubilee performance.  More on that later.

 

Chris

Edited by Chris A
  • Like 1
Link to comment
Share on other sites

It is rather obvious that the conical with the flat walls would be easier to build than the tractrix with the curves. Plus, as you have said, it works better anyway. For  totally new project, I think I am all for trying to build something like this. First I have to finish a rebuild on my LS and perhaps move them on to someone else. <_<  

 

Again, I appreciate what you've been working on. What are the actual dimensions (height, width and depth) of the K-402? I thought the conical wasn't very good at controlling low frequencies. Or perhaps it isn't as critical in a home setting, or not nearly as critical as high db p.a. work.

 

Bruce

 

ps Are you now retired? You said you had lots of free time...

Link to comment
Share on other sites

The current outside dimensions, including the box, are

 

27.25" H (69.92 cm) x 42.25" W(107.32 cm) x 23.63" D (60 cm) 

 

The horn itself is somewhat smaller, but you'll need a box of at least those dimensions anyway.  Of course, If you have relatively high ceilings, carpet on the floor and a narrow room, the speakers can be turned 90 degrees to cover 60 degrees horizontally, so their form factor is fairly close to that of a La Scala II--with (IMO) better than Jubilee imaging performance.

 

The La Scala dimensions (original La Scala, not the La Scala II) are:

 

35.5" H (90.17cm) x 23.75" W (60.33cm) x 24.5" D (62.23cm)

 

The conical horn doesn't have as high of an acoustic resistance (the key parameter for "horn loading efficiency") at low frequencies just above Fc that an exponential or hyperbolic profile will have, however, the flip side is that the conical actually has some resistance/reactance below Fc that the other profiles do not, and that is key to the current horn design performance.  Refer to Beranek for more information of conical vs. exponential vs. hyperbolic.  Its actually a fairly key part of why this design works the way that it does.

 

I'm retired only in the sense that I'm not on someone's payroll nowadays (by choice) and have a lot of latitude in spending my time.  The current New Center project is the fruit of some of that, as is the unmastering stuff and other audio adventures, in addition to work doing pro bono related to my prior chosen profession.   The audio portion has been very easy for me to focus on of late.

 

Chris

  • Like 2
Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
×
×
  • Create New...