Hi Art,
Excellent - many thanks.
The two things that continue to puzzle me are:
1. Why your test results and now those of Wayne, show only a 6 dB gain at low frequencies rather than a 9 dB gain (my previous post refers) - particularly since you indicated in an earlier post that the combined response of multiple front loaded horns is extended downwards in frequency (if I understand correctly).
2. Why the response of multiple tapped horns is not extended downwards, similar to the front loaded horns you compared them against.
Kind regards,
David
Hi Wayne,
Perhaps at the higher frequencies the two signals being combined are no longer completely coherent or correlated?
10 metres is quite a distance, particularly if there happens to be any movement in the air. Also, the two diaphragms may no longer be acting together as rigid pistons, with independent resonance modes starting to develop depending upon the slightly different physical properties of each driver.
Just a thought...
.Kind regards,
David
Cool, will this have an effect on the simulated LF response of multiple TH
Hi Art,
Not unless I make a special adjustment, just for multiple tapped horns
.Kind regards,
David
Yes, this has always been my understanding too. I mean, on-axis, the two should sum coherently since the path lengths from microphone to each source are the same. But still, the two sources are not acoustically close above 100Hz or so. And then there is also the matter of acoustic centers, which probably shift w/frequency a little differently due to the larger effective mouth size. Lots of little differences between a single horn with one or two square meters mouth size and a couple horns with twice that much area.
Hi Wayne,
Provided that there is not a slight breeze blowing, and that there are no random reflections from the ground
.Kind regards,
David
I am puzzled when simulations don’t agree with measured results, you are puzzled when your simulation does not agree with measured results that can be seen from many different sources
. Your Hornresp program is incredibly useful, it will be even more so when it agrees more closely with measured results.Doubling the radiating area of two equally powered speakers adds 6 (.02) dB.
In the case of "too small mouth" FLH, such as most bass horns, an additional LF increase occurs, Wayne's 12Pi appears to have about a 1.5 dB LF increase.
If Hornresp predicts an additional 9 dB rather than the 7.5 dB LF increase, it is inconsistent with measured results.
Wayne, having both the Hornresp sims and his LMS results could give a more accurate account of the disparity, perhaps he will reply to that .
Wayne's measured increase with a pair of his FLH correlates well with my own measurements of multiple FLH, and others he has tested.
Silas Pradetto has measured 1, 2, and 4 Labhorns, similar increases occurred.
Typical TH have a much smaller radiating exit than similar volume FLH, the TH exits are usually only a couple times SD.
My Keystone TH has nearly the same front dimension as the 12Pi, but does not exhibit any LF increase from one to two cabinets.
My 2x10" small TH (26.5 x 22.5 x 15 inch) also does not exhibit a low corner reduction in multiples.
Tom Danley has mentioned many times that TH do not have an increased low corner in multiples.
Bass reflex cabinets don't have much of a reduced LF corner in multiples either, other than the slightly reduced FS from mutual coupling .
Having used BR cabinets from one cone to 112 cones, and FLH from one cone to 80, I can tell you that the difference in FLH LF gain compared to BR gain is substantial.
I will leave the theory to why FLH, BR and TH enclosures respond differently in multiples to theorists, I am more an empiricist, I measure and listen, and make changes until results improve.
Art
I thought probably most people had seen the Hornresp models and measurements of my 12Pi hornsub. They've been posted online for years and subject of many discussions on various messageboards. Otherwise I would have pointed them out earlier. In many of those discussions, I've mentioned how very good I thought the Hornresp model was, and how close the response curve predicted was to what was measured. I even went to the trouble of modeling and building three different versions of the 12Pi hornsub, each with slightly different front and rear chamber sizes. The models and the measurements correlated very well.
Here is a link with all this data:
Pi Speakers - 12Pi Basshorn Subwoofer with Push/Pull Drive
This page shows the Hornresp model of the prototype 12Pi and the LABhorn:
PiSpeakers Forum - 12Pi observations and comparisons with other designs - Wayne Parham, July 08, 2005 at 03:53:47
I used a much older version of Hornresp at the time I designed this horn, but the input forms are shown so you can see how I described the model on that page.
This page shows measurements of the prototype 12Pi:
ProSpeakers Forum - Re: Hornsub shootout RESULTS - 12Pi - Wayne Parham, October 15, 2005 at 14:20:02
Measurements were done using an LMS system outdoors at high power levels at 10M. Even at this distance, the signal to noise is very good and results perfectly consistent and reliable. You can run a sweep several times and always get the same chart. So this is very repeatable and reliable. I would say this is probably the most universally accepted method for obtaining accurate results - 10 meters / 100 watts outdoors ground plane.
This page describes the first and second production versions, including Hornresp models:
PiSpeakers Forum - 12Pi basshorn subwoofer - the best yet! - Wayne Parham, October 04, 2007 at 15:33:34
I say Hornresp models are shown, but they aren't really - only the output screens are shown. But the differences are described. The only changes are the front and rear chambers. So you can take the input screens from the prototype and change the front and rear chambers to get the 12Pi v1 or v2 models.
Both of these pages show measurements of the 12Pi v1:
ProSpeakers Forum - 12Pi - Wayne Parham, October 18, 2006 at 11:31:36
ProSpeakers Forum - Pi Speakers 12Pi v1 - Wayne Parham, October 21, 2007 at 14:56:15
I hope these models and charts prove to be helpful. There is a lot of data there that can be used to compare Hornresp models to accurate measurements of physical models.
Here is a link with all this data:
Pi Speakers - 12Pi Basshorn Subwoofer with Push/Pull Drive
This page shows the Hornresp model of the prototype 12Pi and the LABhorn:
PiSpeakers Forum - 12Pi observations and comparisons with other designs - Wayne Parham, July 08, 2005 at 03:53:47
I used a much older version of Hornresp at the time I designed this horn, but the input forms are shown so you can see how I described the model on that page.
This page shows measurements of the prototype 12Pi:
ProSpeakers Forum - Re: Hornsub shootout RESULTS - 12Pi - Wayne Parham, October 15, 2005 at 14:20:02
Measurements were done using an LMS system outdoors at high power levels at 10M. Even at this distance, the signal to noise is very good and results perfectly consistent and reliable. You can run a sweep several times and always get the same chart. So this is very repeatable and reliable. I would say this is probably the most universally accepted method for obtaining accurate results - 10 meters / 100 watts outdoors ground plane.
This page describes the first and second production versions, including Hornresp models:
PiSpeakers Forum - 12Pi basshorn subwoofer - the best yet! - Wayne Parham, October 04, 2007 at 15:33:34
I say Hornresp models are shown, but they aren't really - only the output screens are shown. But the differences are described. The only changes are the front and rear chambers. So you can take the input screens from the prototype and change the front and rear chambers to get the 12Pi v1 or v2 models.
Both of these pages show measurements of the 12Pi v1:
ProSpeakers Forum - 12Pi - Wayne Parham, October 18, 2006 at 11:31:36
ProSpeakers Forum - Pi Speakers 12Pi v1 - Wayne Parham, October 21, 2007 at 14:56:15
I hope these models and charts prove to be helpful. There is a lot of data there that can be used to compare Hornresp models to accurate measurements of physical models.
Last edited:
I remember mr.Danley talked about the soundwaves coming out of the horn not as a flat wave but more like a sphere/bubble.
Wouldnt this make the horn longer than its physical dimensions?
And because of that, make 4 coupled TH's even longer?
Or does it only apply to FLH's?
In any case, would this difference show up in a impedance-sweep of one and four?
Dag
Wouldnt this make the horn longer than its physical dimensions?
And because of that, make 4 coupled TH's even longer?
Or does it only apply to FLH's?
In any case, would this difference show up in a impedance-sweep of one and four?
Dag
Hi David and all,
It's great that you found where the difference was. I found a good source with sound power level and sound pressure level calculators (engineeringtoolbox.com):
Adding Decibels
It still confounds me too, that two horn-style cabinets mounted mouth next to mouth (in fact creating a new mouth with twice the area) would not show an additional increase in the SPL owing to the increased radiator size. Maybe the response peak that results from the foreshortening and the undersized mouth is reduced when the mouth and the horn cross-section are practically doubled. But this should also show in multi-driver cabinet v. single driver cabinet comparison.
I would also expect a slight increase in effective length for the two mouth model, and a slight difference in the directivity.
Some of the measurement differences from theory may come from the different boundary conditions.
Thanks to all for working so hard on this problem.
Regards,
It's great that you found where the difference was. I found a good source with sound power level and sound pressure level calculators (engineeringtoolbox.com):
Adding Decibels
It still confounds me too, that two horn-style cabinets mounted mouth next to mouth (in fact creating a new mouth with twice the area) would not show an additional increase in the SPL owing to the increased radiator size. Maybe the response peak that results from the foreshortening and the undersized mouth is reduced when the mouth and the horn cross-section are practically doubled. But this should also show in multi-driver cabinet v. single driver cabinet comparison.
I would also expect a slight increase in effective length for the two mouth model, and a slight difference in the directivity.
Some of the measurement differences from theory may come from the different boundary conditions.
Thanks to all for working so hard on this problem.
Regards,
The TH horn acoustical length is basically the same as the physical path length.
The length does not change in multiples.
FLH do not "grow longer" in length in multiples, but the increased mouth area being shared does increase the LF output slightly compared to the HF.
This effect does not happen with TH or BR cabinets.
Last edited:
DJK
A tapped horn is not a classical horn. It is a resonant system in the lower end of it's response. You see that when you are trying to design one and you end up with a saddle shaped response. The lower bump is a product of the location of the driver in the tapered line and the amount of air available. It can be quite a boost in the efficiency.
But having compared all of the normally available boxes in a very controlled manner I can say that I prefer the cleaner sound of a true front loaded horn. A tapped horn has a rather distorted character reminiscent of a poorly made vented enclosure.
Hi Djk,
I think you are right Djk, TH's don't seem to work down low as a horn (pipe). I think if HornResp had the ability to have variable frequency impulse responses (also known as time-stretched Pulses, although used differently in this context) instead of the fixed one, you could show the difference between the direct signal from the driver and the amplified signal of the horn (pipe). It would not surprise me to see that down low the direct signal becomes more prominent after the horn (pipe) start to loose its efficiency.
I think you are right Djk, TH's don't seem to work down low as a horn (pipe). I think if HornResp had the ability to have variable frequency impulse responses (also known as time-stretched Pulses, although used differently in this context) instead of the fixed one, you could show the difference between the direct signal from the driver and the amplified signal of the horn (pipe). It would not surprise me to see that down low the direct signal becomes more prominent after the horn (pipe) start to loose its efficiency.
Last edited:
Historical perspective
This is an interesting short thread that predates Danley's introduction of the tapped horn. It's fairly clear to me he wasn't thinking about it at the time, since he does not make reference or even talk about the idea. But this thread does actually start down that road. It talks about the nature of basshorns being resonant because they are almost always acoustically small even if physically large. There is a link to an article about modifying a horn by placing expansions at key places along the line in order to influence the reactive ripple. Historical perspective:
This is an interesting short thread that predates Danley's introduction of the tapped horn. It's fairly clear to me he wasn't thinking about it at the time, since he does not make reference or even talk about the idea. But this thread does actually start down that road. It talks about the nature of basshorns being resonant because they are almost always acoustically small even if physically large. There is a link to an article about modifying a horn by placing expansions at key places along the line in order to influence the reactive ripple. Historical perspective:
Many thanks Wayne,
Your link with Tom’s comparison comes exactly on the right moment for me.
back to flutes and clarinets????
Jbell, for the first time I began to see the relation between the local fanfare and my neighbour who hates them; 'flutes and clarinets' and 1/4WL dipole with 'balun' and mirror element. Still following
Hi Art,
Not unless I make a special adjustment, just for multiple tapped horns
Kind regards,
David
Hi Art,
It seems that it will not be necessary to make any special adjustment for multiple tapped horns - the proposed new multiple speakers response calculation method appears to correct the problem automatically
.Kind regards,
David