Hornresp

[just a guy]

SPL response is now called Acoustical Power. It changed a few revisions ago.

I haven't done the latest update yet, is Acoustical Power not working or were you just looking specifically for SPL response?

BTW, all Hornresp users should frequent this thread even if diyaudio isn't their forum of choice. The updates have been fast and furious, if you don't check in here you are missing out on a lot.

You have 30 minutes to edit a post, after that edits are only possible by a request to the moderation team. They will do it but you have to request it.

 

[publius]

thank you!

 

[David McBean]

Maybe Norton have given up monitoring because you update the software too often for them to keep up !

McBean crushes Norton!

Woohooo!

Almost sounds like a title match.

.

 

[David McBean]

Hi Kurt,

1. Would it be possible to have the impulse response calculation take the damping material into account, now that it is stored in the record. I guess that would mean taking damping into account when calculating the combined frequency response and other things (pressing "Calculate" in the main input window).

Damping is restricted to the Loudspeaker Wizard so that the frequency range can be limited to a maximum of 2000 hertz (for the reasons given in earlier posts). The way that the calculations are done, it is not practical to include impulse response in the wizard - if it was feasible then the feature would have already been there, similar to the Filter Wizard .

2. I often miss one more segment, mostly because of special aesthetic considerations of folding etc., unfortunately it is sometimes it is before the driver and sometimes the need is after the driver in an offset driver configuration.

Strange as it may seem, to add another segment would require that Hornresp be almost completely re-written - something that I am definitely not in a position to do. Being able to specify another segment would impact on most parts of the program.

I think there is a small bug, if you edit a record and only change the damping, Hornresp doesn't discover that the record was changed, and hence if you are sol silly as to move away from the record, your changes are discarded silently, where Hornresp normally is friendly enough to ask you if you want to save.

The functionality seems to work fine for me . Could you please post the export file for a record that does not recognise that the damping value has been changed, so that I can try to replicate the problem. Many thanks.

Kind regards,

David

 

[David McBean]

I think there is a small bug, if you edit a record and only change the damping, Hornresp doesn't discover that the record was changed, and hence if you are sol silly as to move away from the record, your changes are discarded silently, where Hornresp normally is friendly enough to ask you if you want to save.

Hi Kurt,

Is this problem present in the latest release, Product Number 3480-140728?

Kind regards,

David

 

[nielsenkb]

Hi David

Yes it is in the latest version. Did a little bit more testing. The problem as far as I can see only occurs if you you modify only the percentage of filling in a segment (or more) and nothing else. If you touch the Fr1 for a segment, Hornresp correctly asks if you want to save the record.

Record tested right now is included, but to my knowledge the "feature" (Microsoft term for bug) exists independent of record content.

happy hunting, Kurt

 

[nielsenkb]

Hi Kurt,

Damping is restricted to the Loudspeaker Wizard so that the frequency range can be limited to a maximum of 2000 hertz (for the reasons given in earlier posts). The way that the calculations are done, it is not practical to include impulse response in the wizard - if it was feasible then the feature would have already been there, similar to the Filter Wizard .

OK, I see. I was just dreaming that perhaps the calculation done from the main input window could be allowed more time and take the filling into account. I would just be using the frequency response and impulse response as final documentation of the simulation.

Strange as it may seem, to add another segment would require that Hornresp be almost completely re-written - something that I am definitely not in a position to do. Being able to specify another segment would impact on most parts of the program.

OK

 

[nielsenkb]

Aspect ratio

Hi Dave, another thing that I wondered about:

Often the designs we simulate end up having and output slot with a far from square aspect ratio (perhaps 1:10), like short transmission lines, please see attached. As I understand it a high aspect ration will introduce an acoustic resistive element in the formulas. I sometimes try to add a little filling in the port tee what its effects are, although this is of course pure guesswork. Is there any correct way of doing this? Or is it something that Hornresp potentially could deal with?

Best regards, Kurt

 

[David McBean]

Record tested right now is included, but to my knowledge the "feature" (Microsoft term for bug) exists independent of record content.

Hi Kurt,

Many thanks for the record file.

It took me quite a while, but I finally managed to replicated the problem .

It seems that it only occurs when a setting is directly entered into the slider control by typing the value and then pressing the Enter key. If the slider control setting is changed using the mouse, then the "Save changes to current record?" message appears as it should.

Is this your experience also?

Kind regards,

David

 

[David McBean]

Often the designs we simulate end up having and output slot with a far from square aspect ratio (perhaps 1:10), like short transmission lines, please see attached. As I understand it a high aspect ration will introduce an acoustic resistive element in the formulas. I sometimes try to add a little filling in the port tee what its effects are, although this is of course pure guesswork. Is there any correct way of doing this? Or is it something that Hornresp potentially could deal with?

Hi Kurt,

Hornresp assumes that the acoustical impedance loading at the output of a transmission line, or at the mouth of a bass horn, is the same as that of a circular vibrating piston. For practical purposes this is accurate enough, even when the actual opening happens to be rectangular. Taking rectangular cross-sections into account in the simulations would complicate Hornresp greatly, for little noticeable change in the overall results.

Kind regards,

David

 

[nielsenkb]

Hi Kurt,

Many thanks for the record file.

It took me quite a while, but I finally managed to replicated the problem .

It seems that it only occurs when a setting is directly entered into the slider control by typing the value and then pressing the Enter key. If the slider control setting is changed using the mouse, then the "Save changes to current record?" message appears as it should.

Is this your experience also?

Kind regards,

David

Hi David, that is consistent with what I see

kbn

 

[David McBean]

that is consistent with what I see

Hi Kurt,

Excellent, many thanks for the confirmation.

Hopefully the bug has now been fixed. Product Number 3480-140801 refers.

Note that if the parameters on the main input screen are left unchanged but a change is made in the Loudspeaker Wizard and the Cancel button rather than the Save button is then pressed, the "Save changes to current record?" message will not appear when moving to a different record. The "save changes" message is not really applicable in this case because any changes made in the wizard are disregarded when the Cancel button (rather than the Save button) is used to close the wizard.

Kind regards,

David

PS - I should have said "to replicate" rather than "to replicated" in my previous message .

 

[David McBean]

Hornresp Update 3480-140801

Hi Everyone,

CHANGE

The bug identified in Post #4738 has now been fixed. My thanks to Kurt for reporting this problem.

Kind regards,

David

 

[nielsenkb]

Hi Dave, solved in latest version

I am still puzzled by the aspect ratio of ports etc. If I do a standard BR box in WinISD of about 10 l it recommends a duct of 4 cm diameter 8 cm length, if I change that to rectangular slot of 1*12,6 cm (same area), it recommend only 6 cm length to reach the same tuning frequency.

I found and old article with a design sheet (see attached) at reflex, which (unless I have my head totally upside down) seems to move in the other direction .

So could you tell me what you think should be done in Hornresp to compensate for a 1:10 aspect ratio in a mass-loaded TQWT. As I said I have tried to ad filling to the port under the assumption that hi aspect ratio introduces a resistive element in the port but of course it is pure guesswork right now.

Still hoping, Kurt

 

[AndrewT]

................ it recommends a duct of 4 cm diameter 8 cm length, if I change that to rectangular slot of 1*12,6 cm (same area), it recommend only 6 cm length to reach the same tuning frequency..................

the 4cm port has an area of 12.6cm², a 1x12.6cm slot has an area of 12.6cm²
But the volume taken up inside the box will be different. Sometimes very different. That internal volume of the port including it's walls affects the calculation of the port length.

 

[David McBean]

Hi Kurt,

solved in latest version

Excellent .

So could you tell me what you think should be done in Hornresp to compensate for a 1:10 aspect ratio in a mass-loaded TQWT. As I said I have tried to ad filling to the port under the assumption that hi aspect ratio introduces a resistive element in the port but of course it is pure guesswork right now.

Hornresp models the port as a cylindrical tube, as shown in the axisymmetric schematic diagram.

As far as I am aware, there is really nothing that can be done to more accurately take into account the effects of a high aspect ratio rectangular port in Hornresp, without actually resorting to using a proper rectangular port tube model, which is not an option in this case.

I would be surprised if adding absorbent material to a simulation in an attempt to "correct" for a rectangular cross-section port tube, increased the accuracy of the overall results.

Kind regards,

David

 

[Kolbrek]

As far as I am aware, there is really nothing that can be done to more accurately take into account the effects of a high aspect ratio rectangular port in Hornresp, without actually resorting to using a proper rectangular port tube model, which is not an option in this case.

What will be the biggest difference between a circular and a rectangular tube (when used as a port), is the end correction. This is an added virtual length due to the radiation mass at the end(s) of the port. This added mass makes the port appear to be longer than its physical length. The end correction is different for circular and rectangular openings.

I would be surprised if adding absorbent material to a simulation in an attempt to "correct" for a rectangular cross-section port tube, increased the accuracy of the overall results.

It will not help, since the most significant difference at these low frequencies is in the radiation mass. Resistance is futile.

I would suggest the best way would be to figure out the difference in end correction, and change the length of the port accordingly. I.e. if the rectangular port has an end correction of 3cm, and the equivalent circular port has an end correction of 4cm, shorten the port in the simulation by 1cm. I have some equations by Mechel (Formulas of Acoustics), I can scan them and post them here.

-Bjørn

 

[Kolbrek]

My scanner refuses to work today, but here are the formulas:

Circular duct in baffle: dL = 0.849*a
Circular unflanged duct: dL = 0.613*a
(a = radius)
Rectangular duct in baffle, dimensions 2a x 2b, beta = a/b, a is the smallest dimension:

dL/a = 2/3pi * [beta + (1-(1+beta^2)^(2/3))/beta^2] + 2/pi * [1/beta*ln(beta+sqrt(1+beta^2))+ln(1/beta*(1+sqrt(1+beta^2))]

Since Mechel gives a more complicated formula for the circular duct in baffle (with two extra terms), I expect the last formula could be simplified to

dL/a = 2/3pi * [beta + (1-(1+beta^2)^(2/3))/beta^2]

without much error.

Hope this helps.

-Bjørn

 

[Kolbrek]

Now it works. Here are the formulas from Mechel.

 

[nielsenkb]

Thanks a lot for the answers. Kolbrek, thanks for the formulas. I seem to get a surprising result for a 1:10 I get a correction of 8.7, with -0.3 from the first part, so if I have a slot of 2 cm versus 20 cm, I should add more than 17 cm to the physical length, is that it?

kbn