horn 3 cai

[arcgotic]

Salut!

Incep faza de proiectare a urmatorului horn de tip Multiple Entry Horn, MEH, genul celui de acu 2ani, Synergi. Ma gandeam sa postez aici ce informatii utile am adunat in anii precedenti, informatii luate de pe alte site-uri, informatii ce sunt utile pentru cine vrea sa construiasca un asfel de tip de incinta. O sa postez aici un text in engleza pentru inceput. Apoi o sa urmeze simularile in Hornresp, pentru difuzoarele de care dispun. Si apoi constructia. 

Enjoy! Cu placere!

 

=====

 

If you can't get a 3" mid to go higher than 900hz on a Synergy horn it's time to find a new midrange IMO.
This is not the case. In every case I've found the tap in point to be the reason why the mids drop off too early. 
Even piece of crap Curtis Mathes mids can work well into the 2KHz to 3KHz range.

http://www.goodsoundclub.com/PDF/Synergy_Patent.pdf
__________________


 Most people are using flare rate and cutoff frequency interchangeably. A driver tapped into a Synergy style speaker will not play LOWER in 
frequency than the local flare rate/cutoff frequency. A driver tapped into a Synergy style speaker will normally play up to a frequency where 
the circumference of the cross sectional area of the tap is equal to one wavelength.

Example: You have a mid tap into the horn where the area is 60cm^2. To find the local flare rate/cutoff frequency you use Horn Response to 
calculate the length it takes for it to reach 120cm^2 (a doubling of area). Lets say it takes 4cm to go from 60cm^2 to 120cm^2 in our conical 
horn. Highlight the horn segment in Horn Response and press the E key on your keyboard to change the horn profile from conical to exponential. 
The flare rate will now be displayed. In this example it should tell you the flare rate is 474.37Hz. This is how low you can expect the mids to play 
provided there are not issues with your horn build. Now to figure out how high it will play. We can calculate what the circumference of a circle 
with an area of 60cm^2. It is 27.45cm long. a 1248Hz sound wave is 27.48cm long. Therefore we can expect our mid to play reasonably well 
between about 470Hz - 1250Hz. One last detail is the cancellation notch. If you space the mids too far down the horn, the notch will over ride 
everything. If the notch occurs at 1000Hz, than that is highest you will get out of the mid.
__________________

I'll go over this one last time. To reach your desired high frequency output of the mids the following must be true:

1.) The circumference of the cross sectional area must be equal to or less than 1 wave length of the highest frequency.
2.) The distance between the acoustical center of the compression driver and the tap point must be one half a wave length or less.

In most cases people are getting the mids close (satisfying number 2), but fail to ensure the area is correct.
Everyone needs to re-read the Synergy horn patent again.

*********************************************

The ideal balance is when the circumference of the area the mids tap into is equal to one wave length of the highest frequency, and the 
distance between the acoustical center of compression driver and mids is half a wave length of the highest frequency. In your case of 1.5KHz 
this would mean that you want to tap the mids into the horn with an area of 41.65cm^2 that is 11.42cm from the acoustical center of the 
compression driver.

***********************************************

Let’s use an example where we are crossing over a compression driver at 1200Hz. At this frequency the full wave length is about 28.56cm long. 
If we take this compression driver and load it into a conical horn, the horn will stop providing a load to the driver around where it expands to 
64.91cm^2. Once you expand to beyond an area that is greater than one full wave length, the driver can no longer exert any pressure against 
the horn walls. (The juvenile analogy I use is it becomes like throwing a hotdog down a hallway.) Now take this a step further and use a 
square/rectangular horn instead of an axis symmetrical horn. The pressure in the corners drops even sooner than the flat side walls. Tuck your 
entry ports in the corners, and there is almost zero chance of them interfering with the compression driver’s output. To fully ensure you won’t 
get interference, you want to tap the mids in where the area is slightly larger than 64.91cm^2. We design to have a small gap between when 
the compression driver unloads, and when the mids turn on. There will be enough summing at the crossover point to provide a flat power response. 
There should not be a dip in response.

*******************************************

What I aim for is having the cancellation notch slightly lower in frequency than my crossover point for the mid. (e.g. Crossover is 1st order 
lowpass at 1300Hz, notch at around 1200Hz to 1250Hz.) You want the acoustical cancellation notch doing most of the filtering work. If you 
place your crossover point too close to the notch, you end up with a peak in response once it sums with the compression driver and its crossover. 
The 1st order lowpass on the mid is there for phase shifting, and to filter out the mids high frequencies once it rebounds from the notch an 
octave above.

*******************************************

But the flare rate itself is expressed in a reciprocal length (doubling of cross-sectional area). The cut off frequency and the flare rate are proportional:
100 Hz : 3,66 x 1/m = 0,0931 x 1/inch   = 1.12" per foot
200 Hz : 7,33 x 1/m = 0,1861 x 1/inch   = 2.32" per foot
300 Hz : 10,99 x 1/m = 0,2792 x 1/inch = 3.36" per foot
400 Hz : 4.48" per foot 


A. To determine throat area: At = (2Pi)(Fs)(Qts)(Vas)/c
where At = throat area in sq. ft.; Fs = driver's resonance frequency in Hz;
Qts = driver's total Q factor; Vas = compliance as equivalent volume of air in cu. ft.

B. To determine mouth area: Am = [1/(SF)(4Pi)](c/Fc)(c/Fc)
where Am = mouth size in sq. ft.; SF = Size Factor (SF1 is free space--like hanging above a cornfield; SF2 is half space--on a floor; SF4 is quarter space--on a floor and next to a wall; and SF8 is 1/8th space--in a corner); c = speed of sound in feet per sec. (1130); Fc = desired cutoff frequency (-3 dB).

The horn length should be at least 1/4 wavelength--in practice, the distance from the throat to the calculated mouth size using the contour formula below.

C: To determine contour (flare rate or flare frequency): Ax is equal to At times e raised to the power of 2x divided by xo.
where Ax is the area of the expansion at distance x from the throat; xo is equal to 2/k where k = (4Pi)(Fc)/c, where Fc is the desired cutoff and c is the speed of sound; e = a constant (like Pi, or 3.1416....); e = 2.71828....

So, for an example, using a Pioneer 10-inch instrument loudspeaker, A25GC40-51-F-Q, with an Fs of 30 Hz, a Qts of 0.15, and a Vas of 5.08 cubic feet, and figuring an exponential horn with a cutoff of 50 Hz with a size factor of 8 and a cabinet width (internal) of 16 inches, we find a throat area of 24 square inches, a mouth area of 730.4 square inches, a length of 73.6 inches; the distance in which the cross-sectional area doubles is about 14.94 inches. 


Calculate Flare FC
4 x Pi(3.1416) x Fc... divide the result by 13,200... take that answer and divide .7 by the answer to obtain the Fc.


To get 'in the ballpark', find one that either is a closed back unit that resonates at ~SQRT(200*1400) = ~529 Hz, or if open back, then ~529 = 2*Fs/Qes. How you get there doesn't matter if you're not trying to get max efficiency out of it with 'X' diameter x 'Y' long vents. IIRC, TD's prototype driver's published specs were ~2*250 Hz/0.99 = 505 Hz, so I wouldn't get too hung up on the apparent need for a low Qt just because it's a horn app.

DANLEY INPUT

	
Hi All
Bear, that is correct, it is the mutual radiation which the drivers share when closely coupled which raises the electroacoustic efficiency of the direct radiator and it is the increase in radiation resistance which causes a horn to have a higher efficiency. For multiple drivers to coherently drive a horn as one source, that spacing must also be recognized.

Dumptruck poses questions;

#1, yes the physical / acoustic spacing MUST be held (that ¼ wl spacing or less) in order for the sources to radiate spatially as one acoustic source. With a larger spacing than about 1/3 wl, then the sources radiate independently, the mutual coupling is gone and what you get, depends on the vector sum of the two sources.
As you move around the speaker, whenever the difference in path length is N X 1/2wl, one has a cancellation notch and null in the polar pattern. In commercial sound while unavoidable using multiple sources, an interference pattern is very bad because the lobes that point out in the wrong directions (not the listening area) excite room sound and the directivity (energy in the right direction compared to energy going everywhere else) directly relates to being able to understand random words etc (and preserving information related to the stereo image in the recording).

#2, in all “normal crossovers” like those with names, once past first order, the summed output has an “all pass” phase response. That is a flat amplitude but phase rotation equal to the number of orders times 90 degrees. This places the upper and lower crossover outputs at two different times (hf first, lf later) and this can be seen looking at the Group delay of any simple crossover. Most crossovers do this and is normal behavior.
For the Synergy horns, part of the design uses the inverse spacing of the low, mid and high sources, conceptually like an FIR filter where passively, the hf is delayed the most, the mids less and lf the least to offset the electrical part.
This will not work with any “named” crossovers even without the magnitude and phase of the individual frequency bands added in. In this case, with all the drivers directly coupled acoustically, like signals through a resistor network, one cannot get away with an error or on gets a whopping big cancellation notch everywhere not just one spot..

This then requires a crossover who’s magnitude and phase is what is required to mate the magnitude and phases of each range and this is never a normal test book shape but one that is adapted to the conditions. How i do that has evolved but is certainly based on careful measurements of the real thing.

The SH-50 referenced that I designed in 2005 will reproduce a square wave from near perfect to fair on an O-scope, from about 250Hz to about 2900 Hz, spanning both crossovers and like most all of them, even the largest 10 feet tall with over 100 drivers, sounds like, acts like, measures like a single crossover-less driver. 

The hf driver (BMS 4550) is already one of the top couple most efficient drivers available at any price. It is the wide dispersion, maintained up high, which forces it down. On a horn which narrowed up high, that driver can deliver over 110dB on axis sensitivity.
Keep in mind, on a CD horn, it is usually the hf drivers sensitivity in the top octave that limits the overall sensitivity.
The holes are another part of the design. You may have noticed that as you drive loudspeakers harder and harder, they get “bright” and eventually harsh sounding. Harmonic distortion starts an octave above the real signal and extends upwards by 2,3,4,5 etc times the input frequency.
While that brightness may not be objectionable, the object here is to be a faithful reproducer and one thing that means the spectral balance should not change with level if possible.
The trapped air volume under the cone and the small holes, form an acoustical 2nd order low pass filter, like an electrical filter but in air. The object being that the harmonic distortion the drivers invariably produce, will be attenuated and not enter the horn. The effect does limit the bandwidth and that is the object of it. The idea is you DO NOT want sound the driver produces on it’s own (not part of the input signal) which is always above the electrical crossover and the short obstruction does not affect the horns radiation resistance.

[arcgotic]

Am simulat cu ce am acasa.

O sa folosesc un SBAcoustic full range de la 800-1k in sus

https://www.audioxpress.com/article/Test-Bench-SB-Acoustics-SB65WBAC25-4-2-5-full-range-driver

 

Pe 120-1k, un B&C 10CL51. In incinta inchisa de 22L, cu intrare in horn la 5cm pe adancimea hornului, pe o gaura de 60cm2.

https://www.poweraudio.ro/difuzor_BCspeakers_10CL51

 

Horn-ul de test e un https://www.poweraudio.ro/horn_p-audio_PH-642?search=642 ce il modific, ii mai adaug o continuare pana la 3000cm2 aria gurii.

 

Deocamdata integrez astea 2 difuzoare doar. Sa vad cum suna in combinatie cu 18" in incinta inchisa. Taiate la 120Hz, center-to-center intre horn si mijlocul 18" ar trebui sa nu depaseasca 1/4 din lungimea de unda (283/4 cm adica).

[arcgotic]

Am decupat ce era de luat din horn. Urmeaza continuarea laturilor cu placaj sau mdf.

[arcgotic]

Am facut aseara teste cu SB65WBAC25 in hornul asa cum e acum in poza si nu sunt multumit de ce am auzit. In concluzie o sa regandesc sistemul, in jurul altui difuzor de inalte-medii. Probabil o sa folosesc tot CD B&C DE250 pe care il am, taiat la 950Hz.

Si modific acest horn ca sa fac un synergy.

https://h-audio.de/Lautsprecher-PA/Sica/Hoechtoener-Hoerner-/driver-horns/Sica-1-Hochtonhorn-Q07015A-Abstrahlung-80-GRAD-x60::2429.html

 

I'll be back! :D

[arcgotic]

Am decis sa fac o clona de Synergy de la 0, fara sa ma folosesc de hornul Sica sau altele.

 

Lemnul a fost comandat, placaj fag 8mm, costa 170ron pt ambele hornuri. O sa mai fie nevoie de niste bucati pentru suport horn si alte piese mici. In cateva zile asamblez prima bucata.

Difuzoarele sunt B&C DE250 1" pentru 1600Hz si mai sus, 2buc Visaton FRS8M in paralel (600-1600) si 1buc B&C 10CL51 cu deschidere bass reflex in interiorul hornului (100-600hz).

https://www.poweraudio.ro/driver_BCspeakers_DE250

http://www.visaton.de/en/products/fullrange-systems/frs-8-m-8-ohm    <-- difuzorul asta este excelent.. il recomand.. magnet puternic, masa membrana mica, pret foarte bun!

 https://www.poweraudio.ro/difuzor_BCspeakers_10CL51?search=10cl51

 

Hornul e proiectat pentru o dispersie de 80/50 grade orizontal vertical.

O sa fie maricel (94/60/43 cm) dar din experienta cu primul facut, stiu ca o sa sune cum trebuie.

 

More to follow.

[marius75]

N-am experiență cu hornuri, dar ma pasioneaza subiectul. Ar trebui luată în calcul și varianta sa lăsați diful b&c de 10" să  zburde pana la 1600hz, (are spl bun), iar de acolo in sus, mers pe driverul b&c 1".

Renunțat la visatoane.

Mers pe varianta: doua căi, fiecare cu hornul ei. 

 

[arcgotic]

Idea e sa le inghesui pe toate in acelasi horn astfel incat sunetul sa fie point-source,  distanta intre porturile unde intra difuzoarele sa nu fie mai mare de 1/4 din lungimea de unda a punctului de crossover.

1/4 din lungimea de unda la 1600hz e 5.3cm

1/4 din lungimea de unda la 600hz e 14.1cm

(http://www.procato.com/calculator-wavelength-frequency/)

Deci gaurile pt intrarea medilor e imediat, la 5cm de intrarea compression driverului

Si mai jos pe lungimea hornului, la 14cm, o sa fie portul pentru Midbas.

In poza de mai jos, ce au facut altii, cu 4 medii si 4midbasi, dar pentru camera 2medii si 1midbas e suficient.

 

 

Am avut si solutia cu 2 hornuri separate dar distanta intre centrele acustice era mare, nu era la fel de coerent ca un pointsource synergy horn.

[arcgotic]

Un link pt cine vrea mai multe detalii 

http://www.hificircuit.com/community/threads/index-of-diy-synergy-and-unity-horns.23/

[arcgotic]

Update mic, putin offtopic dar sa pun o masuratoare, poate e curios cineva. 

 

Am cumparat o pereche de Radian 760Neo si le-am pus in hornul  https://www.poweraudio.ro/horn_p-audio_PH-642?search=642 

Le ascult de test taiat la 550hz LR4, 2cai, jos difuzorul de 18" in incinta inchisa .

Mai jos masuratoarea Radian-ului fara eq, de aproape de horn (30cm) cu un HighPass 200hz LR4. 

E destul de liniar, la 10k7 incepe break-ul.

 

[arcgotic]

meanwhile, teste 2cai. Suna foarte bine! 

Ingrediente: panou 60/70, 18", 2" driver. JBL2380 clona horn. XO 500hz LR4. 

 

[arcgotic]

Proiectul inainteaza destul de incet.
Am renuntat la a fabrica hornul de la 0. (nu am sculele si atelierul necesar pentru taieturile precise necesare..)
In schimb am cumparat hornuri B52, intrare de 1", dimensiuni 10" pe 14" 
https://www.parts-express.com/b-52-phrn-1014-1-horn-10-x-14-bolt-on--299-2303

pe care am pus driverul Sica cd 60.38/n92
https://en.toutlehautparleur.com/media/catalog/product/datasheet/sica/Z009484.pdf

si mai departe montez 4 medii Visaton FRS8M
https://www.visaton.de/en/products/fullrange-systems/frs-8-m-8-ohm

si midbass o sa vad, cand ajung acolo. Ma gandesc la FaitalPro 5Fe100. Am 2 si o sa testez pe un horn, incinta bass reflex.

Deocamdata ascult hornul + driver, taiat la 1400Hz cu un 12" mai jos si cu un 18" mai jos.
 

[arcgotic]

Se printeaza flansele pt prindere Visaton pe horn. 

 

 

Intre timp 12" cu driverul taiat la 1400, masurat la canapea. Bun bun!

[arcgotic]

Au venit 2 flanse, arata bine lucrarea, sa vedem cum le prind de horn.. i'll be back.

 

[Sharky]

frumos proiect si frumoasa evolutie, felicitari!

 

Crezi ca ai putea sa incerci o diferenta intre point source si non point source?

Evident fazand difuzoarele in zona de cross in unghiul de interes.

 

Oare e asa deranjanta diferenta? Subiectiv vorbind.

[arcgotic]

Inca nu e gata prototipul sa pot compara dar tin minte de cand am avut synergi-ul pe 2cai, ala mare, 90/60cm hornul, ca efectul era impresionant, si din cauza point source dar mai ales din cauza controlului pe directivitate - scotea camera din ecuatie destul de mult. 

 

Din ce am citit la Geddes, cel cu boxele pe 2cai (12 sau 15" + horn OS), el a studiat si zicea ca ar fi bine ca de la 800-1000hz in sus sa fie point source, ceva legatura cu creierul si cum decodeaza / localizeaza / separa instrumentele. Deaia el taia Driverul de inalte pe la 875Hz.

http://www.gedlee.com/Loudspeakers/Abbey.aspx