Hornresp

Which one would be more accurate? Lec1 or 2?

I ran a simulation in BEM (which solves the wave equation by numerical integration), I attach the results. They show the throat impedance and power response of the system. Two things are evident:

- Lec2 seems to be closest to the more accurate BEM simulation
- Hornresp results are amazingly close to the BEM results.

Unfortunately I can't simulate a 180 deg. LeCleac'h horn in an infinite baffle (2Pi), so I can't comment on how well Lec4 holds up against BEM.

The problem with Lec3 vs. Lec4 may be the 2Pi termination, but maybe David has more thoughs on this.

Bjørn
 

Attachments

  • Lec1_2_BEM.gif
    Lec1_2_BEM.gif
    6.7 KB · Views: 269
  • Lec1_2_BEM_pr.gif
    Lec1_2_BEM_pr.gif
    6.5 KB · Views: 264
For your interest, here are the plots for throat impedance and SPL power response for Lec3. Again Hornresp is very close.

The dip in the response, visible in both the SPL power response plots, comes from rear and front chamber resonances. They also appear in Hornresp, if masking is turned off. That they are so similar, is because I use the same rear chamber model in the BEM program as Hornresp uses.

Bjørn
 

Attachments

  • Lec3_Z.gif
    Lec3_Z.gif
    6.9 KB · Views: 251
  • Lec3_pr.gif
    Lec3_pr.gif
    6.5 KB · Views: 249
Hi David_Web,

Which one would be more accurate? Lec1 or 2?

As Bjørn has so kindly shown, in this particular case the Lec2 predictions are the most accurate. This is because the Hornresp Le Cléac'h horn isophase wavefront model is optimised for horns having a mouth flare tangent angle of not less than 90 degrees. The isophase wavefront model is therefore not really suited to the two examples given.

Both Lec1 and Lec2 are severely truncated Le Cléac'h horns - the mouth flare tangent angles are only 26 degrees. The two examples are effectively identical to catenoidal hyperbolic-exponential horns having the same dimensions - this can readily checked by changing L12 from 'Lec' to 'Hyp'. The Hornresp plane wavefront model is optimised for use with slow flaring horns, which is why those results are the more accurate in this case.

Also decreasing a Le Cléach horns mouth doesn't seem to have too bad influence in a mid horn. Would this be an ok compromise in regards to SQ?

I have no practical experience with Le Cléac'h horns but it seems to me that it would defeat the purpose of using the profile if the mouth flare tangent angle was not at least 90 degrees.

If Jean-Michel happens to read this, then he might care to comment - after all, he is the real expert on Le Cléac'h horns :).

Attached are 90* and 180* Le Cléach horn.

Thanks. Once again, Jean-Michel would be the best person to comment on the level of "bass boost" that could be expected by extending the horn to a fully-formed 180 degree mouth. It seems from Bjørn's BEM simulations that the Hornresp Lec3 predictions are reasonable at least.

Kind regards,

David
 
Hi Andrew,

Do we need to know how to identify that software changeover?

Not really. Prior to Version 5.50 the wavefront model was chosen by the user, but now Hornresp makes this decision.

How do we use Hornresp so that this anomaly is avoided?

In most cases the horn system you wish to simulate will fall clearly into one wavefront model category or the other, and either model will still allow you to optimise a particular design - unless you happen to be sitting right on the cross-over point. It should be rare that a problem will actually arise in practice. In a worst-case scenario, the maximum difference between the two models is still only about 3dB anyway :).

Kind regards,

David
 
I ran a simulation in BEM (which solves the wave equation by numerical integration), I attach the results.

Hi Bjørn,

Many thanks for these simulations, they are greatly appreciated!

I am very pleasantly surprised at how well the Hornresp predictions hold up against your BEM results - almost too good to be true, considering the profound differences in the calculation methods used :).

Kind regards,

David
 
Thanks.
I did a multi segment conical approximation and got close to what you showed above.

The "bass boost" seem to happen in any pi.

BTW Thank you for the Horn wizard!!!
I love it. Spent the last hour "butchering" the art of horns.

I don't mean to pick on my wonderful new toy but,
Could you change the sliders to behave like the TH version, so they don't "loop back" to 0 every thousand cm2. It's a little hard to keep track on size based on slider position.
Could you also introduce this behavior on chambers as going from a very large to smaller would take a very long time unless you exited the wizard and typed it in.


The BEM simulations really shows how good hornresp really is if you treat it well.
Thanks.
 
Hi David_Web,

The two examples are effectively identical to catenoidal hyperbolic-exponential horns having the same dimensions - this can readily checked by changing L12 from 'Lec' to 'Hyp'.

One thing that I forgot to mention in my earlier message:

If hyperbolic-exponential flares are used for Lec1 and Lec2 (which would most likely be the case in practice) then the results predicted by Hornresp for the two horns are effectively identical. There is no longer a cross-over point issue.

Also when bending the horn back 180* you seem to pick up a tremendous bass boost. Is this really true?

The attached screenprints show:

* The throat acoustical impedance of a Le Cléac'h horn having a mouth flare tangent angle of 120 degrees, as predicted and measured by Bjørn.

* The throat acoustical impedance for the same horn, as predicted by Hornresp.

A comparison of the results would seem to suggest that the Hornresp model is valid for Le Cléac'h horns having mouth angles up to 120 degrees, at least.

Kind regards,

David
 

Attachments

  • AH425_BEM_vs_Measured.gif
    AH425_BEM_vs_Measured.gif
    14.1 KB · Views: 181
  • AH425_HR.png
    AH425_HR.png
    25.9 KB · Views: 73
Hi David_Web,

Could you change the sliders to behave like the TH version, so they don't "loop back" to 0 every thousand cm2.

I am reluctant to make the change, as the maximum allowable slider value would then have to be reduced to 64000 cm2, as is the case with the TH Wizard (due to a slider range limitation in Visual Basic). The area limit is not really an issue with tapped horns, but it may be with other very large bass horn systems.

It's a little hard to keep track on size based on slider position.

If you are viewing the Response chart, for example, then you can quickly "sanity check" your slider settings by pressing the S key to instantly see the schematic diagram. Releasing the S key will immediately return you to the Response chart.

Could you also introduce this behavior on chambers as going from a very large to smaller would take a very long time unless you exited the wizard and typed it in.

It is not necessary to exit the Wizard. To directly set a slider to a specified value, key in the value and then press Enter while the slider has the focus.

The Loudspeaker Wizard description on page 11 of the Help file refers.

Kind regards,

David
 
I am reluctant to make the change, as the maximum allowable slider value would then have to be reduced to 64000 cm2, as is the case with the TH Wizard (due to a slider range limitation in Visual Basic). The area limit is not really an issue with tapped horns, but it may be with other very large bass horn systems.

It would not make sense to use the slider at that size anyway as it takes "ages" to scroll to anywhere. (hence the request in the first place) You could always add a multiplier when going over a certain value to get around the limitation.
The front chamber now takes one sweep per liter to get anywhere so some sort of value depending multiplier to make adjustment easier would be nice.

*tries to remember to read the manual every release* :D
 
Hi David_Web,

It would not make sense to use the slider at that size anyway as it takes "ages" to scroll to anywhere.

It should not be necessary to scroll for "ages" - just directly enter a starting baseline value as explained in my previous message, to get into the general range you wish to operate in, and then move the slider bar to make any required fine adjustments. A completely different value can be quickly entered if required.

The front chamber now takes one sweep per liter to get anywhere so some sort of value depending multiplier to make adjustment easier would be nice.

Each slider sweep range has been chosen to ensure that the sensitivity of the control is optimised for that particular parameter - making chart changes not too fast or not too slow. I suspect that multipliers would complicate, and possibly confuse, things.

Kind regards,

David
 
I know I found a program somewhere a while back that allows you to combine and edit dat files from hornresp but I can't find it now. Does anybody know what I'm talking about?

I wrote such a program some time ago, called HornrespMerge. It may be what you are looking for. It can be downloaded from my homepage at folk.ntnu.no/bjorkol under SW and Tools.

Bjørn
 
@ Bjørn
Thanx, this looks very interesting, hope I get the time to investigate this. Theoretically, one could enter any horn-contour and look, how it behaves - right?

Yes. One just have to make sure the surface elements are small enough (typically less than 1/6th wavelength), and set the right boundary conditions. I have used the axisymmetric code, which makes the surface definition very simple. For rectangular horns you need to divide the surface into triangles, which can be a bit harder.

You may also want to take a look at AxiDriver, another BEM package with GUI.

Bjørn
 
Hi David_Web,

It would not make sense to use the slider at that size anyway as it takes "ages" to scroll to anywhere. (hence the request in the first place) You could always add a multiplier when going over a certain value to get around the limitation.
The front chamber now takes one sweep per liter to get anywhere so some sort of value depending multiplier to make adjustment easier would be nice.

Just letting you know that in the next release of Hornresp the Loudspeaker Wizard S1 to S6 area sliders for other than tapped horns, and the Vtc slider for all horns, will have value-dependent multipliers.

The operation of the non tapped horn S1 to S6 area sliders will be as follows:

If value < 10000 then:

Sweep range = 1000
Small change = 1
Large change = 10

If value >= 10000 then:

Sweep range = 10000
Small change = 10
Large change = 100

Moving from the minimum to maximum allowable value using the area sliders will then require only 19 full sweeps, compared to the 100 currently needed.

The Vtc slider will also have a similar "times 10" multiplier for values greater than or equal to 10000.

Thanks for raising the slider operation issue, and for suggesting the multiplier solution.

The next release of Hornresp may still be a few weeks away - I am very busy working on other things at the moment.

Kind regards,

David
 
OK, I am close to admitting defeat. I can build most anything from Hotrod to racing airplanes to hi end audio gear that looks like art. But I am just not that good at understanding electronics and in particular speakers. So, I need help.

1. what would be a good 15" driver for a horn with a small back chamber to start with. I don't have drivers now so I am open to anything.

2. Given your recommendation for drivers, can you post a screen shot of all the number that go in the boxes on the bottom half of HornResp?

I think I have the top half of the program down fairly well. But I just can seem to wrap my head around the driver data portion. Thanks.