Hornresp

HORNRESP VERSION 19.10

David McBean said:
Please note that it is quite okay to run multiple copies, provided that they all have the same version number. The important thing is not to edit, add or delete any records while the multiple copies are running.

Hi Oliver,

I have done some further testing and found, contrary to my earlier comment above, that it is really too risky to allow multiple copies of Hornresp to run, under any circumstances. Adding or deleting records in the first copy before the second copy is opened, can lead to all sorts of unpredictable outcomes. If the two copies happen to have different version numbers, then the potential problems are even worse.

Hornresp Version 19.10 safeguards against any inadvertent corruption of the data file in this fashion by allowing only one copy of Hornresp to run at a time.

Also, from Version 19.10 onwards it will no longer be possible to open the Hornresp.dat data file if it has previously been opened by a more recent release of the program. This should also assist in preventing inadvertent corruption of the data file, going forward.

Thanks for raising the issue - you have helped to make Hornresp a better product :).

Kind regards,

David
 
Hello Lars,

As an example of compression driver simulation under HornResp you'll find in that link, my simulation using the TAD TD2001:

http://www.audioasylum.com/forums/hug/messages/12/128658.html

Best regards from Paris, France

Jean-Michel Le Cléac'h

revintage said:
Hi David,
How do I simulate a horn with a compression driver, in my case JBL 2441 ? Have checked your help and googled but haven´t found anything.
 
Hello Jean Michel,

Thanks for the link, very interesting. Funny, I had your or the kugelwellen horn in mind. Wanted to do a 180Hz horn as this seems to be the flare of the short horn in the driver. My intention was to try it from 300Hz.

I have used your spreadsheet and another KW-spreadsheet for calculations. Was interested in a simulation just to verify.

As I mentioned in another post I have found data for the 375 in an AES-paper by Geddes but the area and volume figures do not make sense.


Brgds
 
These are the of the JBL 375. Anyone who cares to translate to metric?
 

Attachments

  • 375-data.jpg
    375-data.jpg
    79 KB · Views: 1,586
Re: Re: HORNRESP VERSION 19.10

Jmmlc said:
Hello David,

Do you know this new software

http://www.djgroundbass.com/HyperHorn.html


Best regards from Paris, France.

Jean-Michel Le Cléac'h

Hi Jean-Michel,

No, I was not aware of this new software - many thanks for bringing it to my attention.

It looks very interesting, although it does not seem to be available for download just yet - at least it wasn't when I tried.

Kind regards,

David
 
HORNRESP UPDATE

Hi Everyone,

Just letting you know that I have released yet another minor update to Hornresp. The change does not justify a new version number, but the product number is now 1910-080515.

Change Details: Typically when developing a loudspeaker system, Hornresp users tend to start with a base design and then iterate and refine from there. In the past, when a record was edited and a new record then added, any changes made to the old record could either be saved to both the old and new records, or not saved to either record.

With the new release, when a record is edited and a new record then added, any changes made to the old record can either be saved to both the old and new records, or just to the new record only.

This means that it is now much easier to generate and preserve a series of iterations leading from the base starting design to the final system specification. In other words, it comes in handy if you wish to maintain an “audit trail”, documenting the evolution of a particular design project over a number of records.

On another issue, a number of users have asked me to explain the significance of the Hornresp product number. It simply represents the version number with the point separator removed, followed by the software release date in year / month / day order. For example, the current release of Version 19.10 was uploaded on 15 April 2008. This means that the product number for this release is 1910-080515 (format is vvvv-yymmdd).

At the moment the product number can change quite frequently, sometimes almost on a daily basis. I am continually tinkering with the code looking for ways to further refine and improve the program. Sometimes the change may not be apparent to the user, but in other cases it will be quite obvious. Please note that most of the changes are now related to the operation of the context-sensitive user interface, and the elimination of very subtle bugs and logic errors that I still manage to find from time to time. Very few changes are being made to the actual simulation models used in the program.

Incidentally, if anyone does come across a bug or “unexpected outcome” while using Hornresp, please do not hesitate to let me know, as I greatly value such feedback.

Also, I would be very interested to learn if Hornresp works with the new Vista operating systems (32-bit and/or 64-bit). Unfortunately I only have access to Windows XP myself.

Thanks in anticipation for any information.

Kind regards,

David
 
Dear David,

After I send with the message

http://www.diyaudio.com/forums/showthread.php?postid=1520648#post1520648

a graph with simulations performed under Hornresp that allow to compare the response of different type of 160Hz horns (exponential, hypex, tractrix, Le Cléac'h, ...) to the response of the waveguide having the same length and mouth diameter than the exponential, Earl Geddes wrote in that message:

http://www.diyaudio.com/forums/showthread.php?postid=1520813#post1520813

that the simulations given by Hornresp for the oblate spheroidal waveguide "are wrong".

Well, I would just let pass the information...

Best regards from Paris, France


Jean-Michel Le Cléac'h
 
Jmmlc said:
Dear David,

After I send with the message

http://www.diyaudio.com/forums/showthread.php?postid=1520648#post1520648

a graph with simulations performed under Hornresp that allow to compare the response of different type of 160Hz horns (exponential, hypex, tractrix, Le Cléac'h, ...) to the response of the waveguide having the same length and mouth diameter than the exponential, Earl Geddes wrote in that message:

http://www.diyaudio.com/forums/showthread.php?postid=1520813#post1520813

that the simulations given by Hornresp for the oblate spheroidal waveguide "are wrong".

Well, I would just let pass the information...

Best regards from Paris, France


Jean-Michel Le Cléac'h

Hi Jean-Michel,

Many thanks for the above information.

Hornresp uses the same isophase wavefront model for all horn flare types (including the OS waveguide).

Not having any measured results to compare the OS W/G predictions against, there is probably little that I can do at this stage to improve the generalised model.

It is interesting to note that it appears to be only with OS waveguides that the predicted versus actual results are significantly different (reportedly). I see that you have assumed corner loading in your calculations. Perhaps the results might have been closer if 2 Pi had been used?

Kind regards,

David
 
Hello David,


Thanks for your reply which confirms my thoughts.

You'll find as attached file 2 simulations performed under Hornresp using the 2 Pi solid angle as you recommanded.

Please note that I choose the same dimensions for the horn (Fc = 160Hz as in the Azura 160 horn and the waveguide.

Here is a picture of the Azura 160 horn
http://www.azurahorn.com/images/Yamaha_6681B_on_AH160.jpg

and its measured response is here:
http://www.azurahorn.com/6681_on_160.pdf

The simulations performed with Hornresp confirm what I know since my reading of H. Olson (e.g. fig 5.5 p.105 in Acoustical Engineering, Ed. Prof. Audio. Journ. 1991):

throat impedance for a conical horn (and an OS waveguide behaves very much the same regarding throat impedance) is very low in the low frequency domain compared to the throat impedance of an exponential horn or an hypex horn. From this lesser loading at LF of the diaphragm there is with the conical horn or the waveguide a lost in efficiency in the conversion by the driver from electrical energy to acoustical energy (at frequencies at which exponential or hypex horns still have a high efficiency).

I was puzzled the first time when Earl Geddes wrote on AudioAsylum that the LF part of the reponse of different horns and waveguide (... of comparable dimensions) is not related to the profile between the throat horns and waveguide :

see: http://www.diyaudio.com/forums/showthread.php?postid=1521956#post1521956

"Any device placed on a compression driver loads the same above some LF point. They will all approach the characteristic impedance of air divided by the throat area. So everything "loads" the same. Thats why I ignore "loading" as irrelavent. Loading and cutoff are both concepts that are not relavent factors in a device whose primary intent is to control directivity."


As he possess a recognized authority even I thought that I was minsinterpreting his thought and pass over.

What do you think of Earl Geddes formulation? (I'll undesrtand if you don't reply...)

Best regards from Paris, France


Jean-Michel Le Cléac'h


David McBean said:


Hi Jean-Michel,

Many thanks for the above information.

Hornresp uses the same isophase wavefront model for all horn flare types (including the OS waveguide).

Not having any measured results to compare the OS W/G predictions against, there is probably little that I can do at this stage to improve the generalised model.

It is interesting to note that it appears to be only with OS waveguides that the predicted versus actual results are significantly different (reportedly). I see that you have assumed corner loading in your calculations. Perhaps the results might have been closer if 2 Pi had been used?

Kind regards,

David


Earl Geddes: Any device placed on a compression driver loads the same above some LF point. They will all approach the characteristic impedance of air divided by the throat area. So everything "loads" the same. Thats why I ignore "loading" as irrelavent. Loading and cutoff are both concepts that are not relavent factors in a device whose primary intent is to control directivity.
 

Attachments

  • lec_os_compar.gif
    lec_os_compar.gif
    8.6 KB · Views: 1,354
Hi Jean-Michel,

Many thanks for the additional information. The comparison of predicted vs measured results for the Azura 160 horn is very interesting. The low frequency roll-off point seems to have been reasonably accurately predicted. Are your simulation results showing the constant directivity power response, or the on-axis pressure response?

"Any device placed on a compression driver loads the same above some LF point."

The above statement can be readily verified using Hornresp, by checking the acoustical impedance charts for different horn types having identical throat areas. At 20000 hertz the throat acoustical impedances are effectively the same for all types (normalised resistance ~ 1, normalised reactance ~0).

With regard to the OS waveguide itself, strictly speaking the device is actually a horn and not a waveguide (by definition, a waveguide is a non-radiating transmission line with the propagated energy being "guided" from one point to another). Also, from a physics and geometry viewpoint there is nothing really special or unique about the 'OS waveguide'. It is simply yet another horn flare, asymptotic to a conical horn, but with a transition at the throat which facilitates the wavefront shape changing from plane to curved as the sound propagates down the horn.

Kind regards,

David
 
Hello David,

In both examples, the curves are related to constant directivity response.

Thanks for your remark about Earl's sentence.
As in the discussion it came when we disussed comparison of the efficiency at low LF between horns and waveguides, I was thinking (probably wrongly) that this not well defined "some LF point" was situated by example in the frequency domain for which the resistive part of the acoustical impedance was a fraction ( e.g. "half") of the resistive part of the acoustical impedance at highest frequency.

Also I did a parallel with what Earl Geddes wrote in his reply to the AudioXpress editor"

http://www.gedlee.com/downloads/Horn Theory reply.pdf

B. Kolbrek wrote: "Well at least the horn equation is good for predicting the shapes that one needs in order to achieve good impedance loading – at least at low frequencies!”

E. Geddes: "Sure, except that it turns out that virtually any shape connecting a given throat and mouth area will yield approximately the same impedance – within a dB or so. In the long wavelength region where the Horn Equation is valid, shape is simply not a significant factor, the waves don’t see shape. Basically they only see the inlet and the outlet areas and the distance between them and everything else is of secondary importance."

I interpreted this, and I am not the only one (it was even translated by someone on a French audio forum), as if Earl Gedddes was saying that whatever is the flare shape the throat impedance was the same at low frequency for horns and waveguides having comparable dimensions...

But this is probably due to my poor knowledge of the arcanes of the English language.

Best regards from Paris, France

Jean-Michel Le CLéac'h


David McBean said:
Hi Jean-Michel,

Many thanks for the additional information. The comparison of predicted vs measured results for the Azura 160 horn is very interesting. The low frequency roll-off point seems to have been reasonably accurately predicted. Are your simulation results showing the constant directivity power response, or the on-axis pressure response?

"Any device placed on a compression driver loads the same above some LF point."

The above statement can be readily verified using Hornresp, by checking the acoustical impedance charts for different horn types having identical throat areas. At 20000 hertz the throat acoustical impedances are effectively the same for all types (normalised resistance ~ 1, normalised reactance ~0).

With regard to the OS waveguide itself, strictly speaking the device is actually a horn and not a waveguide (by definition, a waveguide is a non-radiating transmission line with the propagated energy being "guided" from one point to another). Also, from a physics and geometry viewpoint there is nothing really special or unique about the 'OS waveguide'. It is simply yet another horn flare, asymptotic to a conical horn, but with a transition at the throat which facilitates the wavefront shape changing from plane to curved as the sound propagates down the horn.

Kind regards,

David
 
Jmmlc said:
"Sure, except that it turns out that virtually any shape connecting a given throat and mouth area will yield approximately the same impedance – within a dB or so. In the long wavelength region where the Horn Equation is valid, shape is simply not a significant factor, the waves don’t see shape. Basically they only see the inlet and the outlet areas and the distance between them and everything else is of secondary importance."

Hi Jean-Michel,

Thanks for the extra background information - I can now see why there could be some confusion out there :).

The above statement is certainly true for very short horns and for long slow-flaring horns where the mouth area is not much larger than the throat area. Following are the measured throat impedances for an exponential horn and a tractrix horn having similar throat areas, mouth areas and axial lengths. Based on this data, I leave it to you to decide whether the comment can be applied generally to all horns :). It would be really interesting to be able to compare the measured exponential and tractrix throat impedance results against those for either a parabolic or conical horn having identical dimensions.

To complete the picture, I have also included Hornresp-predicted results for the above exponential and tractrix horns (Ang = 2.0 x Pi, S1 = 20.27, S2 = 2307.22 and L12 (Exp and Tra) = 55.90). The Hornresp data was exported to Excel and converted to show values of ka (2 x Pi x f / 34400 x 27.1) rather than frequency along the x-axis, to make it easier to directly compare the predicted and measured results.

You can see that the predicted resistance and reactance curves for the exponential horn have the same overall characteristic shapes as the measured results for the exponential horn. Also, the predictions for the tractrix horn compare reasonably well to their respective measured results. The comparisons are certainly close enough for the purposes of calculating theoretical power responses (to an acceptable degree of accuracy).

If the Hornresp predictions for an equivalently-dimensioned conical horn are anything to go by, then the above statement could probably be justifiably questioned :).

Kind regards,

David
 
Hello David,

Thanks for the reply. I can see that we agree on a possible questionning about Earl's sentence ;-)

Thanks also for the impedance measurements (is it possible for you to indicate where did the measured curves of the acoustical impedance were published).

Too bad I gave my conference about "axisymetrical horns" last saturday, my introduction to Hornresp would have been improved with those data.

Best regards from Paris, France

Jean-Michel Le Cléac'h




David McBean said:


Hi Jean-Michel,

Thanks for the extra background information - I can now see why there could be some confusion out there :).

The above statement is certainly true for very short horns and for long slow-flaring horns where the mouth area is not much larger than the throat area. Following are the measured throat impedances for an exponential horn and a tractrix horn having similar throat areas, mouth areas and axial lengths. Based on this data, I leave it to you to decide whether the comment can be applied generally to all horns :). It would be really interesting to be able to compare the measured exponential and tractrix throat impedance results against those for either a parabolic or conical horn having identical dimensions.

To complete the picture, I have also included Hornresp-predicted results for the above exponential and tractrix horns (Ang = 2.0 x Pi, S1 = 20.27, S2 = 2307.22 and L12 (Exp and Tra) = 55.90). The Hornresp data was exported to Excel and converted to show values of ka (2 x Pi x f / 34400 x 27.1) rather than frequency along the x-axis, to make it easier to directly compare the predicted and measured results.

You can see that the predicted resistance and reactance curves for the exponential horn have the same overall characteristic shapes as the measured results for the exponential horn. Also, the predictions for the tractrix horn compare reasonably well to their respective measured results. The comparisons are certainly close enough for the purposes of calculating theoretical power responses (to an acceptable degree of accuracy).

If the Hornresp predictions for an equivalently-dimensioned conical horn are anything to go by, then the above statement could probably be justifiably questioned :).

Kind regards,

David