Jean Michel on LeCleac'h horns

[angeloitacare]

hello Jean Michel

i think your horns deserve a exclusive thread at this forum , where all related issues can be discussed, questions made and be answered, and contribute to common knowledge about your horns.

maiby someone else wants to start with the first questions.

Angelo

[panomaniac]

I'll start!

Perhaps Jean Michel would like to explain something about his horn profile. What was the begining design idea?

How is it different from other profiles such as:
Conical
Exponential
Hyperbolic
Tractrix
Iwata

What advantages does the LeCleach'h horn have over any of these?

A vous, Jean Michel!

.

.

(I have heard and/or built all of the above, but have not heard the LeCleac'h)

[tinitus]

Tried to calculate fore a 10" midbass going down to 150-200hz
I want it to be about 350mm hight and about a meter wide

But the calculation shows a length og around 850mm, which is fine, but suggests a mouth width of 10 meter

[tinitus]

Ahh, I reckon I should just look at the area at a shorter length
This is exciting

[Jmmlc]

Hello Panomaniac,


While it has been eventually said that the Le Cléac'h is based on Webster's equation, in fact the expansion law is secondary when considering the horn known as Le Cléac'h horn since David Mc Bean used to introduce this profile into Hornresp and since Bjørn Kolbrek call it that way in the second part of his paper in Audioxpress

part 1 :
http://www.audioxpress.com/magsdirx/...olbrek2884.pdf

part 2:
http://www.audioxpress.com/magsdirx/...olbrek2885.pdf

In fact my contribution should be looked more as a method to calculate horns than rather a new expansion.
If one of those days Angelo should invent a new expansion law for a horn then my method could be applied to calculate it's profile.

If we start from a given expansion law for the evolution of the area of the wavefronts versus their distance, by example catenoidal or hypex or exponential, what is the difference between my method and the anterior methods?

For commodity I prefer to define the wavefront as an equiphase surface described by the ensemble of points reached at the same time by a single wave emitted at the throat.

All anterior methods rely on an hypothesis for the shape of the wavefronts. Common and false hypothesis taken for the calculation of horns is that the wavefronts are planar. Voigt took for his Tractrix horn the hypothesis of spherical cap wavefronts having a fixed radius. Kugelwellen horns are based on the same hypothesis but with a doubled radius compared to the Tractrix.

During the 70's and the beginning of the 80's I was looking at the rare published pressure fields maps inside horns ( measurements by Morse in Mac Lachlan's book, by Hitachi Labs in Jean Hiraga's book, ... ). This readings lead me to think that all the anterior methods to calculate horns were eroneous as the mesured wavefronts where neither planar neither spherical.

I had the idea that if a single wave was propagating in the case the speed of sound was constant inside the horn, the above mentioned equiphase surfaces should be parallel ( = orthogonally equidistant ) each one from the other, a feature than you don't find, by example, in Voigt's hypothesis for the design of the tractrix horn for which the wavefronts cannot be orthogonaly equidistant (because they are translated at constant “speed” along a common direction). Another requirement is that the wavefronts should reach the walls of the horn at 90°.

At the time I used to program a first profile in Basic on one of the first computer I had in my lab (a Hewlet Packard 9625 if I correctly remind). I used for that a method based on more than 1 million of discrete geometrical elements the assembly of which possess the required parallelism I was looking for. No hypothesis was done on the shape of the wavefronts, the only hypothesis was their equidistance (and a known expansion law...).
It is only after personnal computers were available that I wanted to wrote a first QuattroPro spreadsheet and many years after an Excel spreadsheet... I built the first horn using that method in 1982.

Fortunately I could replace the calculation of millions of discrete elements of the original software written in Basic into a recursive formula. Around 1995 that horn was discussed on Joenet (Sound Practices ‘s discussion list) and on the french speaking group [son-qc] and I was asked to commit a paper which one was published in issue N°6 of “Musique & Technique” a publication of the Belgian Lowther Club. In 1998 Marco Henry (Musique Concrete) used to read that paper and some time later he begun to build in France his Jerzual horns calculated by my method . Nearly at the same time, Martin Seddon in Australia begun building a large format Le Cléac’h horn (now known under reference Azura Horn AH160) and the first batch came in USA where the horn received good fame within the horns/tubes crowd. One of the first user was famous tube electronics builder Dave Slagle ( see http://www.arduman.com/aa/Sayfalar/slagle/slagle.htm )

One thing that schocked many persons when they first saw the profile calculated by this method for an hypex or an exponential horn was the rolled back mouth. Even, many people were convinced at this time that the rolled back mouth was an empirical choice I made. The truth is that it is fully “natural” and results from the combined parallelism and expansion law. Unfortunately today many persons still think (wrongly) that is is useless to curve back the mouth and that the horn can be stopped when the opening angle reaches 180°.

Another question that was often asked was why the horn is wider than all commercial horns having the same acoustical cut-off. It is because in order to reduce at maxium the reflection of waves from mouth to throat we need to open the horn at more than 180° (I recommand 360°). Doing this we can consider the horn as quasi-infinite (measurements confirmed the simulations done with that hypothesis).

There was several versions of the spreadsheet, the first ones rely on the simplified hypothesis that the curved width of a given wavefront could be estimated from the diameter of a disk having the calculated area. The last versions use a step by step estimation and are much more precise, specially near the mouth.


Best regards from Paris, France

Jean-Michel Le Cléac'h

Originally posted by panomaniac
I'll start!

Perhaps Jean Michel would like to explain something about his horn profile. What was the begining design idea?

How is it different from other profiles such as:
Conical
Exponential
Hyperbolic
Tractrix
Iwata

What advantages does the LeCleach'h horn have over any of these?

A vous, Jean Michel!

.

.

(I have heard and/or built all of the above, but have not heard the LeCleac'h)

[tinitus]

Hi Jean Michel

I understand what you say about mouth curves, that they should prolong all the way, and not stop at 180gr

But surely you cant demand that at low frequencies, where we are bound to make compromises

Is there a frequency where it gets important or where its less important

Im mostly thinking of 150-1khz

edit,
I have been trying to find your previously posted layout of an idialised elliptical horn, but cant find it

[angeloitacare]

If one of those days Angelo should invent a new expansion law for a horn then my method could be applied to calculate it's profile.

ha ha, i think that would be a little over my mind and abilities....

but, a curiosity of mine, and i guess, never asked you , Jean Michel :

I guess you experimented over the years a series of drivers : which ones do performe best in your opinion with your horns :

- in the mid-bass region ( 200hz - 1000hz )
- in the midrange region ( 1000hz - 10khz )

if i remember right, you mentioned you use Yamaha drivers. What is your opinion abouth them ? i can get a pair with 1" exit, and 3" aluminium diaphragm. Suspention seems to be the same as from the JA 6681B. Do you think these are worth a try, and a contender with S2 ?

Angelo


---------------------------------------------------------
http://www.audiovoice-acoustics.com/forum/

[soongsc]

Jean-Michel
If I want the initial angle to be a certain angle, how can I do that?

[Jmmlc]

Hello Angelo,

The best sounding low mid compression drivers I used to listen were:

1) Goto
2) WE555 (I have the Alnico French modification by LMT)
2) Onken
3) Yamaha JA6681 (I have 2 pairs of them)
4) TAD TD2001 (yes, a bit modified and on the Sato horn above 250Hz)
5) Vitavox S2
6) Philips

The best sounding mid compression drivers I used to listen were:

1) Onken
2) TAD TD2001 (I have 3 pairs of them with old or recent diaphragms)
3) Goto
4) Altec 288


Best regards from Paris, France

Jean-Michel Le Cléac'h

Originally posted by angeloitacare


ha ha, i think that would be a little over my mind and abilities....

but, a curiosity of mine, and i guess, never asked you , Jean Michel :

I guess you experimented over the years a series of drivers : which ones do performe best in your opinion with your horns :

- in the mid-bass region ( 200hz - 1000hz )
- in the midrange region ( 1000hz - 10khz )

if i remember right, you mentioned you use Yamaha drivers. What is your opinion abouth them ? i can get a pair with 1" exit, and 3" aluminium diaphragm. Suspention seems to be the same as from the JA 6681B. Do you think these are worth a try, and a contender with S2 ?

Angelo


---------------------------------------------------------
http://www.audiovoice-acoustics.com/forum/

[Jmmlc]

Hello Soongsc,


In the version you have of the spreadsheet the tangential angle of the wall of the horn is imposed by the evolution of the area with the distance to throat (resulting from your choice of throat diameter, T coefficient and Fc).

For an arbitrary set of those 3 parameters the angle at throat is fixed. You have two solutions to obtain the angle you want:

1) try different values of the diameter, T coefficient and Fc until you reach the desired angle

2) calculate a spline curve that will progressively go from the tangential angle you desire to a given Le Cléac'h horn profile.

(the angle value you can enter in the actual spreadsheet will have some effect on the profile only if the angle of the throat of the driver is quite near of the throat angle of the horn)

Best regards from Paris, France

Jean-Michel Le Cléac'h

Originally posted by soongsc
Jean-Michel
If I want the initial angle to be a certain angle, how can I do that?

[Jmmlc]

Hello Tinitus,

I know thar some difference exists between my recommendations and what people do with those.

You'll see on the pictures on the page:
http://www.melaudia.net/rueil0903.php
what I consider as the maximum width for an axial Le Cléac'h horn.
We'll have a public demo of that horn during the weekend of 28 & 29 (march 2009).

For cut-off frequency under 150Hz I don't recommand the complete axial horn calculated by my method.
I recommand a quarter of axial horn placed on a corner of the listening room as on the picture:
http://ndaviden.club.fr/pavillon/axial_murs.gif


For very low cut-off frequency I recommand 2 halves of quasi-cylindrical waves Le Cléac'h horns
http://ndaviden.club.fr/pavillon/cyl_mur.gif

The spreadsheet to calculate that horn can be dowloaded for free at: http://ndaviden.club.fr/outils/cylind.zip

You can see a picture of such a realisation at:
http://img383.imageshack.us/my.php?i...280x768ac0.jpg

Best regards from Paris, France

Jean-Michel Le Cléac'h

Originally posted by tinitus
Hi Jean Michel

I understand what you say about mouth curves, that they should prolong all the way, and not stop at 180gr

But surely you cant demand that at low frequencies, where we are bound to make compromises

Is there a frequency where it gets important or where its less important

Im mostly thinking of 150-1khz

edit,
I have been trying to find your previously posted layout of an idialised elliptical horn, but cant find it

[soongsc]

Hello Jean-Michel,

Is there no relationship between T and the initial angle? Is there a geometrical deffinition in relation with T?

[Jmmlc]

Hello Soongsc,

T is a coefficient in the general frouma linking the area and the distance to throat for horns of the hyperbolic family

http://www.bd-design.nl/contents/media/hypo.gif

T = 0 : catenoidal horn
T between 0.5 a,d 0.707 : hypex horn
T = 1 : exponential horn
T infinite for conical horn.

Best regards from Paris, France

Jean-Michel Le Cléac'h

Originally posted by soongsc
Hello Jean-Michel,

Is there no relationship between T and the initial angle? Is there a geometrical deffinition in relation with T?

[panomaniac]

Originally posted by Jmmlc
In fact my contribution should be looked more as a method to calculate horns than rather a new expansion.

All anterior methods rely on an hypothesis for the shape of the wavefronts.

Ah! D'accord! Now I understand. Sometimes it just needs to be spelled out in plain langauge. =)
Thanks so much for that - it really helps understand what you are doing. Very interesting information in your post.

Thanks to Angelo for starting the thread.

[tinitus]

Jean-Michel
Fantastic "feedback"
Very informative

Thank you

[Petter Persson]

Jean-Michel, great answers you give!

In wich space do you design this horn:

http://ndaviden.club.fr/pavillon/cyl_mur.gif

floor placement? or floor/wall?


Regards / Petter

[tinitus]

[i]Originally posted by Petter Persson

In wich space do you design this horn:

http://ndaviden.club.fr/pavillon/cyl_mur.gif

floor placement? or floor/wall?


Regards / Petter [/B]

Could it be something like this one
Looks cool btw

Kind of like half a JBL Paragon

[pk]

Hi Jmmlc,

Thanks for sharing all the knowledge regading your horns!

I have a question regarding the system that according to your post above will be demonstrated ultimo March:

http://www.melaudia.net/rueil0903.php

Could you please specify the crossover points of this system?

Thanks!

Best regards
Peter

[tinitus]

[i]Originally posted by Petter Persson

floor placement? or floor/wall?

[/B]

Or maybe more like this one

[tinitus]

I know this is approved

[soongsc]

Originally posted by tinitus
I know this is approved

I hope they don't have animals running around Nice place to hide.

[soongsc]

Originally posted by Jmmlc
Hello Soongsc,

T is a coefficient in the general frouma linking the area and the distance to throat for horns of the hyperbolic family

http://www.bd-design.nl/contents/media/hypo.gif

T = 0 : catenoidal horn
T between 0.5 a,d 0.707 : hypex horn
T = 1 : exponential horn
T infinite for conical horn.

Best regards from Paris, France

Jean-Michel Le Cléac'h
Thanks Jean-Michel,
I'll try to figure things out.

[panomaniac]

Originally posted by soongsc
I hope they don't have animals running around Nice place to hide.

Just play a little Wagner or Jean Michel Jarre. Animals all gone!

[angeloitacare]

hello Jean Michel

thanks for your answer. Interesting, your mention of Onken drivers. In the past, there where a few offers at ebay and some places elsewhere, and i was curious to try them out, since these seem to have been built only for home / high-end use - but hesitated , because of lack of information on the web in regard of their performance, specially comparing to other drivers.

There is a ongoing debate for a long time already on various forums about different horn / wave guides, each one claiming and advocating that this or that is best. You have a lot of advanced theoretical answers , why in your opinion the LeCleac'h profile performs better, and avoid setbacks generally attributed to horns. My most important concern was always to get the most fidel/natural sound, without horn coloration and honk, but also without have to miss dinamics. Only horn systems can deliver that live like sound, which amazes me. In the past, and more intensively in the last 3 Years, i have tried out a number of drivers and different horns and setups at home. The best i have come up until now is my actual configuration, using in the midrange a LeCleac'h horn, and Vitavox S2 driver. While it does reproduce most music incredibly well, beeing it a real pleasure and relaxing, to spend time, listening to music, there are other things, that i wish, it would do better. Namely voices, specially female voices, have the tendency, sometimes to sound irritating. And there is also a high degree of directivity , and little move of listening position, changes the presentation dramatically ( thats less of a concern for me ). After i compared with the Audax Medomex , ( even not well integrated in the system ) , but in a closed, apropriate box, crossed same as S2, it became clear to me : the Medomex sounds far more natural/neutral - and as conclusion : there is still a far way to go, to come close to the performance of a direct radiator, when the main point of observance is neutrality/timbre/tone. After your post at AA, i played around with different crossover points, trying to improve things, but the existing configuration was best. So: whats the reason the Medomex is more natural ? is it the compression driver ? the simple first order crossover ? no linear power response ? or in the end the horn shape ? What other factors could it be, than mouth diffraction, HOM's , unsmooth transmission from driver to throat etc ? I once asked you, it it might be, that the shorter the horn, the less honk ? u said no. But this is getting to be something, i start more to think about. I suspect that main cause of horn honk and coloration is coming from the first exponential part at the throat with small opening. A horn with fast opening looses its function, turning just into a waveguide with directivity control. Than dinamics goes away as well.... In the final, the conclusion : whatever we do , it is always a compromise....

Angelo

[Lynn Olson]

Now that we have a new thread, here's the latest BEM eye candy from Bjorn Kolbrek in Norway. The first example you've seen before in the Onken thread, it's an AH425 with a 288 driver - cutoff frequency 425 Hz, T = 0.707, 1.4" exit diameter and 8 degree flare from the compression drive. Phase plug HF cancellations and diaphragm breakup are not simulated.

The AH425 is a 16.5" wide horn with a nominal 425 Hz cutoff, and usable (resistively loaded) frequency response down to 550~600 Hz. In practice, a 800~900 Hz crossover is recommended to avoid the region with sharp group-delay variations.

BEM at 1 kHz: