K-400 Horn

[mike5555]

Hi All I was wondering if someone who owns a K-400 Horn could trace the outer profiles as I have shown on the pic below. You see I built a set some years back and I really loved the sound of the wood version a lot. I am now in the process of building another set of LaScala's and could really use the sketch. sorry about the quality of the drawing the M.S. is acting up a bit. Thank you for looking and have a Great day! wardm40@yahoo.com

Mike.

K-400.pdf

[WMcD]

It seems an interesting project.  Could you send pictures of what you built.

 

Of course I can trace one for you.  Would you rather borrow one for a few months?

 

Addition by edit.  It would be interesting to fill one with plaster to see what the internal dimensions really are.

 

WMcD

Edited August 13, 2015 by William F. Gil McDermott

[mike5555]

Hi William! Thank you for the response. I have just started buying the material and parts but I have to go with wood on the horns They give a less tinning sound I can post picks of the horns as I build them as I use a bit of outside bracing. I think I can get what I want by the tracing but if you want me to go as far as an ID cast and or a outside cast it would not be a problem. I build large RC boats and make molds out of fiberglass they work excellent. Let me know and We can proceed that way. I would send you the speaker and molds back.

Thank you!

Mike

[WMcD]

Whatever you wish.  I was more making an observation on what would be interesting to do, rather than asking you to do it.

 

I can make the tracing for you of course.  Otherwise if you want to borrow a K-400 I can send it to you on loan for a few months.  I assume you'll pay shipping, of course.

 

You can give me your address via the forum.

 

WMcD

[tigerwoodKhorns]

Why not build a tractrix horn instead? 

[mike5555]

I just like the sound of the wood k-400.

[Arash]

I can't live with K-400 since I heard Edgar Tractrix horn. far better in term of sound quality. K-400 has a better wow factor but it will cause me a headache after 30 minutes of listening or so

[WMcD]

Mike contacted me and would like the tracing for now.  I'm pleased to do so.

 

But let me offer the attached spreadsheet to Mike and all here.  It is self-explanatory, even if in a complicated way.

 

WMcD

Dimensions of inside of a wood version of a K-400 varient.zip

[WMcD]

If you are used to a spreadsheet at all, you can figure out how the values in the cells are calculated. 

 

One technique might be a bit foreign, and I anticipate questions.

 

We are used to using an exponential expression to figure areas based on either the length of the horn (or area someplace down the horn) from the flare constant m or the distance x.  For example:

 

Area1/Area2 = e^mx  were e is 2.718

 

The issue is what do we do when we want to get mx isolated so we can figure out one of them?

 

The answer lies in the fact that the function of raising e to a power is the inverse of taking a natural logarithm or in spreadsheet terms ln(...).  Because they are inverse functions, ln(e^y) = y.  To say when we multiply something by its inverse, we get 1.  Just like 2x1/2 =1

 

Note we're talking ln(..) which is the natural log and not log(. . .) which is the common log which is based on 10 and not e.

 

So we can go back to

 

Area1/Area2 = e^mx  (And we do have to get the e^mx all by itself on the right side to do things.)

 

Then we write a new equation by performing the same operation on both sides (taking the natural log), which preserves the equality. Like the teacher said, you have to do the same thing to both sides of the equation.  Smile.

 

ln(Area1/Area2) = ln(e^mx) = mx

 

We usually have Area1 and Area2 and we can take the natural log of the number resulting from the division.  And then that number is equal to mx and we usually have one of those, and we can find the other with simple algebra.

 

- - - - -

 

We're used to working with areas.  In the spreadsheet I've determined the equation of the offset of the wall from the centerline, and that is not an area, but it is on a line determined by a similar exponential equation.   We have not (Area1/Area2) but (Offset1/Offset2) = e^nx.  The n is a flare constant for the displacement.  I just use n here to distinguish from m.

 

- - - -

 

The final thing is that we're used to working with the theory that you set an area for the throat (small end), figure the flare constant m, and then calculate how the horn area grows as we increase the distance x toward the mouth (big end).

 

But, you can also start with the mouth (big end) and a flare constant m, and treat the distance as a -x (or m as -m I suppose).  Then as we increase x in magnitude, the area (or displacement) shrinks.  It is all the same thing with a change of sign.

 

- - - -

 

I noodled with these issues for a long time before getting an understanding.  It doesn't seem to be explained anywhere.  So I'm pleased to pass this items along to you so you can have some fun with the design without getting frustrated.

 

Smile,

 

WMcD

Edited August 13, 2015 by William F. Gil McDermott

[mike5555]

Hi William as I can design most types of boats I am not familiar with spread sheets and do not own Office or any other spreadsheet software. I own AutoCAD and Photoshop but no Office. I really do better with a picture or schematic. Thank you William!

Mike

[WMcD]

Attached is the spreadsheet display including the graph printed in .pdf.  You should be able to view it without any expensive software.  Just the free Acrobat Reader.

 

I think the information is everything you'll need.  But let me know.

 

WMcD

Print Out of inside a wood version of a K-400 varient.pdf

Edited August 14, 2015 by William F. Gil McDermott

[mike5555]

Thank you William, Thank you for the spread sheet, I do understand it and it will work perfectly! Thank you again and have a Great day!

Mike

[JohnA]

Hi William as I can design most types of boats I am not familiar with spread sheets and do not own Office or any other spreadsheet software. I own AutoCAD and Photoshop but no Office. I really do better with a picture or schematic. Thank you William!

Mike

You can download Apache Open Office free. Its spreadsheet, Calc, is comparable and compatible with MS Excel.

[Mykick]

Attached File   Dimensions of inside of a wood version of a K-400 varient.zip

Dear William, thank you very much for valuable data! I'm trying to get deeper in the subject. Your spreadsheet is very helpful! Thanks again!

 

The only thing I can't get is why the overall lenth of the "horn width line" (blue colour in your spreadsheet) and the "height line" (red) are equally measured in respect of the x axis. It appears to me that the lines should be of the same lenth by themselves to form a correct horn side. And in your case it's shorter!

Shall I make any corrections to the calculations? Or am I getting something wrong?

Thank you. Mike

[Welborne]

Sharing my roller coaster experience here and may be useful for you guys.

In the past I read A LOT about how bad and headache-causing the K400/401 could be even before I acquired a pair of La Scala. With such belief in mind I even went directly ordering a pair of Le Cleac'h horn to use on top of my LS with 1" throat. I tried everything stock originally (AL-3 crossover, k77m tweeter, k401 horn and k55M mid drivers) initially with two of my Single ended tube amp one being 6c33c the other being 300B all stuffed witg good parts and proper driver and operating points etc.

Initially I tried with 16ohm tap, and the sound? Horrible the least to speak. Bright, honky, resonance here and there, no image no soundstage to speak of, bass sounds boomy and seriously boxy. Everything you can read about "bad ugly horn sound" you name it. Yeah the La Scala k401 horn and tweeter are as bad as I have read.

Immediately I have dismissed the stock components and installed the Le Cleac'h flare and made a Type A crossover using solen coil and Russian K40-9y PIO capacitors. Hooked these to my tube amps again and yeah they sound so much nicer, so much less honkiness, things get much smoother, and now it started to sound more normal. But there is still some brightness and honkiness that I could not completely get rid of.

Some weeks later, I converted my tube amps tap to 8ohm, then the sound has taken another major leap forward, its so much more involving. "Yeah it's the impedance matching things again" so I thought. But there was still something there that kept bothering me. The sound was too relax and lost a lot of the charm of the famous Klipsch 3-way horn that I first heard at a friend's home 10 years ago. The sound I get now has nothing too wrong but also lack lustre. I kept questioning myself "Is this what the Klipsch is famous for 60 years? Just this?"

Then one day, out of the blue, I hook the LS (still with Le Cleac'h horn, Type A, B&C De120 tweeter) up with a Nuprime Sta-9 class D my friend lent me to try with my open baffle speakers. I did not think class d would work with Klipsch LS but bearing the "no harm to try" attitude we hooked it up anyway. OMG! The treble is so silky and smooth, violin sounds so airy and rich, mid is palpable and warm, bass is tight and dynamic, Nothing stand out to ask for attention, and nothing nasty to distract me from enjoying music. Inspired by this, I removed the Le Cleac'h horn, put back the k401, hooked up the stock Al-3 crossover and k77m tweeter, then I listened to everything stock with Nuprime class D....

God! This is shockingly amazing! I have goosebumps all over my body, violin is as silky as the B&C DE120, perhaps a little bit less extended but only a little bit if any. The biggest improvement is the MID! Human voice is hanging in the middle of the front wall as if musician is here, soundstage is deeper, image is much MORE solid and realistic. Bass, mid and treble all coherently come from the same piece of cloth. I played 50 recordings ranging from chamber baroque, violin concerto, solo, big choral works, jazz big band, wind symphony, saxophone solo, complicated symphoc works, Ella Fitzgerald, Louis Armstrong, Duke Ellington, Sophie Mutter..... Etc etc and there is absolutely ZERO brightness, glare and ZERO honkiness to speak of. Where are all those nasties I read about horn and old Klipsch classic models? They disappear in front of this lowly class D Nuprime, completely. I deliberately turn up the volume to test to see if I was just listening to them too softly. No, even at higher volume that would have invited some complaints from neighbours those brightness, honkiness, harshness are not here, with k401 and stock crossover.

Ok, is this all the magic of Nuprime? I try another much cheaper Class D based on a Texas Tpa3118d2 chip called Wiener that was on a grouo buy thread from diyaudio.com. Bingo! The sound of LS is just as delicate, refine, smooth, romantic and silky as when hooked to Nuprime, and perhaps even more refine (due to tpa3118's even higher switching frequency).

Puzzled by this unbelievable and unanticipated journey, moving from boutique DHT tube amps, boutique horn flare, all PIO crossover to everything stock and some modern class D amp, I investigated further into the logic behind all these.

My conclusion: Klipsch classic horn are amazing design, can sound absolutely MINDBOGGLING head and shoulder above many other speaker design on planet earth, even in stock form which many people have written badly about. All you need is a low distortion amp with very low output impedance that can "drive through" the wild impedance roller coaster inherent in all Klipsch classic models (LS/Belle/Klipschorn/Cornwall etc).

You only need a few watts, but you need good few watts, and you need amp with high damping factor. Needn't be class D, can be any topology SS, tube PP, Otl etc.

This is my experience. And a big lesson. It's a game more than just the "efficiency" number.

[Welborne]

Now after all these, I tend to form the opinion that much of the honkiness, shouty, harshness that many reported are not the problem of the horn, the drivers nor even the crossover. The popular belief has been that high efficiency speakers, and in particular HORN for that matter, SHOULD be matched with flea power TUBE amps with singled end, zero negative feedback and Directly heated topology, and so a lot of us including myself followed that belief and pair our KH/LS/Belle with high output impedance/low damping factor gear. When use with Klipsch's wild impedance curves (from as low as 4.5ohm to something like 43ohm) serious tonal coloration and abnomalies are to be resulted. Surely you can drive the Klipsch to hearing imparing level with barely 1w to 2w but we also need the high damping part so that we will see as little db variation along the frequency response as possible.

[Welborne]

I had much the same results playing with the crossover networks along with multiple amps. My experience with Class D amps it only limited to the TPA 3116 that some swear by as sounding great. My personal best sounding amplifier among the 7 or 8 I'm using at the present time is a SET. It is one of 3 I own and everyone of them sounds different, some better than others just like SS amps. Just because you have a SET does not guarantee great sound not that your SET's are not good amps. Amplifiers and speakers are a team with the room also being a big part of the equation. The single best sounding change to my speakers was using a simple 1st order crossover, ditching the autotransformer with Russian PIO caps. No harshness at any volume and tremendous sound stage and with my favorite SET pinpoint center vocals and instruments in their perspective space in the sound stage with no smearing of each other, everything in it's place. Many of my amps do not have the high damping factors and they still sound better than amps with the high damping factor. I have not listened to a good Class A/B amplifier in a few years, Most of my amplifiers are SS Class A or SET's but there are many happy with the better Class A/B amps as well. My understanding of the damping factor is better control of the bass speaker. You definitely want some decent damping factor in an amp but extremely high numbers will not guarantee great sound either.

I personally had no problem with the K400/K401 horn the many years I listened with it. If you have a some harshness or honkiness with your speaker I would look for some other solution to the problem. PWK did not make or design a bad horn when he developed the K400 horn for his bass bin. If you try and say the K400 horn is a bad design you are saying that the great PWK did not know what he was doing. I had a problem with a Firstwatt F6 amplifier as sounding bright when I first started listening with it. I blamed it on the amp but after some discussion with Nelson Pass and some changes to my system I found it was my setup not the amp than needed improving to reach the level of excellency of the F6 amp. The F6 showed the deficiency of my system. A better piece of equipment can do this and if you do not improve you system you can be quick to blame the new equipment as not sounding good. You hear this all the time. "I bought so and so amplifier and it sounded like crap." When I made the changes in my crossovers all my other amps along with the F6 sound much much better. The harsh brightness was gone.

Welborne is like me, he kept trying and changing things until he got it right for him. We have different opinions and results but we reached the sound we liked. We reached the line together, I am not going to say the finish line because I am not through yet. I have 3 new amps in the works ri

Hi Wdecho,

You are spot on. My 6c33c and 300B sound horrible at 16ohm with stock LS, they sound decent enough with stock LS at 8ohm to the extent that I don't miss a better horn and better crossover. The class D (Nuprime v4 generation for this matter) took a major step forward from my two tube amps in nearly ALL aspect, approaching the best sound I have ever achieved in my 30 years in this hobby. To my ears they are much better but I would not call it end of the journey yet.

My last speaker which I am still rolling with LS is a Bastanis Mandala Chrystal with 18" active ripole bass, driven by the same two SET tube amps and they sounded gorgeous.....until now when LS meets high damping factor amp.

Attached is the impedance curve of Klipschorn for reference. I am very sure a low damping amp dring these will cause quite serious tonal anomalies, but then it could still produce some wonder "tube" tones and harmonics that please our ears sufficiently that make us overlook problems elsewhere.

Edited February 15, 2016 by Welborne

[Welborne]

My conclusion is like yours: there is nothing wrong with the K400/401. They just show problems more readily (gear matching, harmonics distortion etc) and are less forgiving than other offerings. PWK has done a lot of work and he knows what sounded right to his ears.

[Deang]

"PWK discussed this in some of his papers and in his opinion it was not necessary a bad thing."

PK did not like resistors. In what paper did he say using them "was not necessarily a bad thing"?

The plot is from the Heyser test report. Regarding the complex impedance, he stated, "Fortunately, from the standpoint of amplifier drive requirements, none of these deviations can cause any problems whatsoever, as long as the power amplifier can drive four ohms at modest power. Because of this, and the fact that up to 25 average watts there is absolutely no change of admittance with drive level, I chose to omit the admittance plot for the Klipschorn. In this case, we do not need it."

Demands on the amplifier will be the greatest at the lowest frequencies. The current demand decreases with increasing frequency, where the impedance increases as well.

Good amplifiers are designed to deliver their output with varying impedance - that's what amplifiers do. Designers are fully aware that loudspeakers do not present a resistive load. The idea that amplifiers want to see a resistive load is wrong. In reality, well designed amplifiers with a moderate amount of feedback simply don't care.

An autotransformer does not turn power into heat in order to attenuate, so why add a resistor? In some cases, amplifier distortion might actually be reduced, as PK demonstrated.

Swamping the network back provides convenience if you are trying different drivers and/or horns - but if you know the sensitivity and impedance of the drivers, you can scale the parts accordingly and dispense with the swamping resistor.

Now we come to the exception to the general rule. Zero feedback amplifier designs have a high source impedance, and the frequency response of these amplifiers will track the impedance of the loudspeaker. So, with the networks that use an autotransformer, the response will sound tipped up in the area where the impedance is peaking - the midrange. However, it isn't necessary to swamp back the entire network, excellent results can be achieved by putting a 15 ohm resistor in parallel with the K-55, and dropping from tap 4 to tap 3 (capacitor values remain the same).

The K-400 is fine, but it doesn't sound as good as Tractrix, which sounds much more relaxed.

[Marvel]

I used the DHA2 crossover in my LS (pulled the AL crossover...), and everything else stock, it sounded amazing with my Welborne Labs 2A3 amps. Finally switched out the tweeter, but the basic system sounded very good.

 

Bruce