what you mean with "linearly in total radiated power" ?
"Total power" is not exactly what you are after when feeding a horn
Remember a horn is a "diffraction alignment device" - meaning it is capable to bend the usual "one direction vector" of wavefront at origin in a controlled way so that you possibly get smooth sound field over 4 PI in the end.
If you start out with some "chaotic wavefront" at the throat you need better software to predict horn behavior I guess
Michael
"Total power" is not exactly what you are after when feeding a horn
Remember a horn is a "diffraction alignment device" - meaning it is capable to bend the usual "one direction vector" of wavefront at origin in a controlled way so that you possibly get smooth sound field over 4 PI in the end.
If you start out with some "chaotic wavefront" at the throat you need better software to predict horn behavior I guess
Michael
It was my mistake. By accidence I fed a right mode with 25mm throat and get nice smooth response. With a proper feeding we we could manage modal traducer I'm pretty sure.
From this time my measurements are calibrated. Measured the same horn with STX D800-Ti. It's nice that JMLC is linear on axis and does not require any eq.
From this time my measurements are calibrated. Measured the same horn with STX D800-Ti. It's nice that JMLC is linear on axis and does not require any eq.
Attachments
Hello,
The funny thing is that we have lot and lot of axial frequency response curves been measured for many different Le Cléac'h horns (mounted on compression drivers).
All of them or nearly are the flatest axial unequalized responses we can obtain with horns.
And ... in the end no simulation (Hornresp, Akabak, BEM,...) ever succeed in predicting a flat axial response...
Best regards from Paris, France
Jean-Michel Le Cléac'h
The funny thing is that we have lot and lot of axial frequency response curves been measured for many different Le Cléac'h horns (mounted on compression drivers).
All of them or nearly are the flatest axial unequalized responses we can obtain with horns.
And ... in the end no simulation (Hornresp, Akabak, BEM,...) ever succeed in predicting a flat axial response...
Best regards from Paris, France
Jean-Michel Le Cléac'h
Hi Jean-Michel, have you tried to adjust motor parameters in simu to match measured FR ?
Michael
Hello Michael,
Even with a theorical constant velocity source, simulations done using such software like Hornresp give for the Le Cléac'h horn a very flat power response (called "constant directivity response" in Hornresp).
In such case it is obvious that due to the increasing directivity of the horn, the axis response is rising with frequency and this is clearly seen on the O° response curve given by Hornresp. (Give a look to the attached graph for the compared power response and on axis response curves predicted by Hornresp for a Le Cléac'h horn having a Fc of 160Hz and a T of 0.8.)
BEM simulations lead to the same.
I used to talk about that with David Mc Bean and eventually I asked him to try a constant acceleration source (but the correction was too large).
As you can see the on axis response possess a +3dB/octave ( = 10dB / decade) slope and I was asking myself if a linear dependance on f (frequency) was forgotten...
Best regards from Paris, France
Jean-Michel Le Cléac'h
Even with a theorical constant velocity source, simulations done using such software like Hornresp give for the Le Cléac'h horn a very flat power response (called "constant directivity response" in Hornresp).
In such case it is obvious that due to the increasing directivity of the horn, the axis response is rising with frequency and this is clearly seen on the O° response curve given by Hornresp. (Give a look to the attached graph for the compared power response and on axis response curves predicted by Hornresp for a Le Cléac'h horn having a Fc of 160Hz and a T of 0.8.)
BEM simulations lead to the same.
I used to talk about that with David Mc Bean and eventually I asked him to try a constant acceleration source (but the correction was too large).
As you can see the on axis response possess a +3dB/octave ( = 10dB / decade) slope and I was asking myself if a linear dependance on f (frequency) was forgotten...
Best regards from Paris, France
Jean-Michel Le Cléac'h
Attachments
Depends – I played around with AxiDriver BEM software and gave it a try with your TAD 2001 data :
http://www.audioasylum.com/forums/hug/messages/12/128658.html
on the LeCleach 0.8 / 160Hz horn
Horn contour and sound field:
On-axis response at 1cm to compare to measurements done at the same distance:
Would you consider that to be too far from those measurements of the Azuara-Horn seen here :
http://www.azurahorn.com/6681_on_160.pdf
( same horn - but on Yamaha 6681B compression driver )
Michael
http://www.audioasylum.com/forums/hug/messages/12/128658.html
on the LeCleach 0.8 / 160Hz horn
Horn contour and sound field:
On-axis response at 1cm to compare to measurements done at the same distance:
Would you consider that to be too far from those measurements of the Azuara-Horn seen here :
http://www.azurahorn.com/6681_on_160.pdf
( same horn - but on Yamaha 6681B compression driver )
Michael
Last edited:
Take a look
Older guys having fun with building serious horns
6moons audio reviews: Hornographic Pursuits I
Older guys having fun with building serious horns
6moons audio reviews: Hornographic Pursuits I
I am quite curous. What is the main purpose of sim and measurment at the mouth? How does it compare with response at 1meter from the mouth?Depends – I played around with AxiDriver BEM software and gave it a try with your TAD 2001 data :
High Efficiency Speaker Asylum - Hornresp and TD2001 simulation - Jmmlc - November 06, 2007 at 03:34:29
on the LeCleach 0.8 / 160Hz horn
Horn contour and sound field:
On-axis response at 1cm to compare to measurements done at the same distance:
Would you consider that to be too far from those measurements of the Azuara-Horn seen here :
http://www.azurahorn.com/6681_on_160.pdf
( same horn - but on Yamaha 6681B compression driver )
Michael
For this particular case, it does make sense since the mouth is so large. I think the distance to mouth diameter ratio best be at least 5:1, 10:1 might also be considered.
... that may be a good question - but for me the question to answer was merely about does or doesn't BEM provide (more or less) comparable results.
IMO correlation between measurements (though with different driver) and the simulation I knocked up is "not that bad at all" - what you think ?
Should you provide measurement, contour and motor spec of a design of yours - at whatever distance - I give it a try !
Michael
Last edited:
In the geddes thread, I have already posted measurements and BEM (in baffle) and BEM (free standing). My model is smaller with a mouth diameter about 1/3 of yours. So there is not way we can compare apples to apples. I think we should just know why we select a specific sim and measurement distance.... that may be a good question - but for me the question to answer was merely about does or doesn't BEM provide (more or less) comparable results.
IMO correlation between measurements (though with different driver) and simulation is "not that bad at all"
Should you provide measurement, contour and motor spec of a design of yours - at whatever distance - I give it a try !
Michael
If I recall correctly, 1cm in AxiDriver is distance in front of the sound source, which is inside the horn. I have no idea what Hornresp reference point is.
Last edited:
Helo Jzagaja,
Why your BEM simulation of the Le Cléac'h horn is so rough?
Bjorn Kolbrek used to perform a lot of long duration BEM simulations and the maps he obtained are so smooth when yours are so bad!
Mige0 seem to obtain smooth simulations too...
Best regards from Paris
Jean-Michel Le Cléac'h
Why your BEM simulation of the Le Cléac'h horn is so rough?
Bjorn Kolbrek used to perform a lot of long duration BEM simulations and the maps he obtained are so smooth when yours are so bad!
Mige0 seem to obtain smooth simulations too...
Best regards from Paris
Jean-Michel Le Cléac'h
Last edited:
Do you mean middle map? I should explain it was for 20kHz frequency where BEM has some difficulties. For any lower it is smooth. With a proper motor parameters axial response becomes more linear.
Forgive my ignorance but why power response on Hornsrep simulation is flat? This because "constant velocity source" that isn't achievable in practice?
Forgive my ignorance but why power response on Hornsrep simulation is flat? This because "constant velocity source" that isn't achievable in practice?
Hello Michael,
You on axis simulation of a 160Hz Le Cléac'h horn is the nearest I could see until now from the many measurements performed on Le Cléac'h horns.
I wonder why we are unable to obtain the same flat on axis response when using Hornresp...
Best regards from Paris
Jean-Michel Le Cléac'h
You on axis simulation of a 160Hz Le Cléac'h horn is the nearest I could see until now from the many measurements performed on Le Cléac'h horns.
I wonder why we are unable to obtain the same flat on axis response when using Hornresp...
Best regards from Paris
Jean-Michel Le Cléac'h
... that may be a good question - but for me the question to answer was merely about does or doesn't BEM provide (more or less) comparable results.
IMO correlation between measurements (though with different driver) and the simulation I knocked up is "not that bad at all" - what you think ?
Should you provide measurement, contour and motor spec of a design of yours - at whatever distance - I give it a try !
Michael
Hello Jzagaja,
OK this simualtion is done at 20kHz, I understand now why the map looks bad...
I used to ask David McBean about the questionable on axis response prediction for a Le Cléac'h horn obtained under Hornresp using both my model of the TD2001 or the constant velocity model .
See my message:
http://www.diyaudio.com/forums/subwoofers/119854-hornresp-65.html#post1862698
His reply is here:
http://www.diyaudio.com/forums/subwoofers/119854-hornresp-66.html#post1863488
Here is the comparison between a simulated (by another method) and a measured on axis response for a tractrix horn (from Michael Makarski's thesis, 20 april 2006) . It is pretty acurate.
Best regards from Paris, France
Jean-Michel Le Cléac'h
OK this simualtion is done at 20kHz, I understand now why the map looks bad...
I used to ask David McBean about the questionable on axis response prediction for a Le Cléac'h horn obtained under Hornresp using both my model of the TD2001 or the constant velocity model .
See my message:
http://www.diyaudio.com/forums/subwoofers/119854-hornresp-65.html#post1862698
His reply is here:
http://www.diyaudio.com/forums/subwoofers/119854-hornresp-66.html#post1863488
Here is the comparison between a simulated (by another method) and a measured on axis response for a tractrix horn (from Michael Makarski's thesis, 20 april 2006) . It is pretty acurate.
Best regards from Paris, France
Jean-Michel Le Cléac'h
Attachments
Last edited:
As shown in the geddess thread, AxiDriver on axis response and off axis response of an in-baffle sim was close enough to measurements in free air. So I found valuable for a first try for prototyping. For most simulation using a compression driver, usually there is not enough data to model the portion from throat to diaphragm correctly. This makes a significant difference.