This is the original horn, driven at the 1.4" throat:
And this is driven at the 1" end of the conical extension:
Honestly, it doesn't make much sense to me.
And this is driven at the 1" end of the conical extension:
Honestly, it doesn't make much sense to me.
Last edited:
It's described here: http://www.at-horns.eu/athex.html (below the waveguides)
Sorry to hear that but I'm afraid I can't help here.
Faital HF108 (several dB more sensitive than the Peerless).
(This driver has a large exit angle (31°), not a good match for the 6° adapter.)
(This driver has a large exit angle (31°), not a good match for the 6° adapter.)
This was the same driver (or the other one from a pair) in a regular 1" horn (the Sandhorn, roughly 360 mm across):
This is definitely smoother through the midrange, so obviously something is happening that's still not quite right...
This is definitely smoother through the midrange, so obviously something is happening that's still not quite right...
Last edited:
Did you try to simulate with subdomains instead of external -domain-only approach ? Your results look rather suspicious, since the adapter increases the length of the horn quite a lot, while "the cutoff frequency" of the horn is almost not reduced. It looks to me not very correct and contradicts to some of my simulations results.
That was a good remark! Here it is with an internal subdomain added.
Later I'm going to try to reverse the interface orientation, which also has an effect sometimes. (No NUC here.)
Anyway, it seems that this kind of throat impedance actually helps to make the drivers more flat.
The directivity doesn't seem to suffer, which I explain by the conical shape of the extension - it in fact doesn't change much the wavefront at the original throat cross section.
Now think of what can be done with drivers like HF146...
Last edited:
Interesting thing is that Abec3 simulates correctly in axisymmetric mode using both subdomain or external-only domain approaches (there are very minor differences). While in 3D mode only external-only approach gives incorrect acoustic loading. It seems to be some kind of bug in the Abec3 code.
Last edited:
Anyway, this really seems to be a way how to extend the usable range of a driver to lower frequencies, without changing the directivity.
A regular ST260 including a generic driver model (left) and extended with a conical section (right):
Perhaps only the resonances are better damped in the real world (?).
A regular ST260 including a generic driver model (left) and extended with a conical section (right):
Perhaps only the resonances are better damped in the real world (?).
Making the horn itself larger. On the right an overlay -
It seems the larger horn also increases the absolute radiated power below ~800 Hz, peaking 5 dB of equivalent omnidirectional SPL at 500 Hz. Which is not something I would have guessed.
It seems the larger horn also increases the absolute radiated power below ~800 Hz, peaking 5 dB of equivalent omnidirectional SPL at 500 Hz. Which is not something I would have guessed.
Last edited:
The same 1" driver in the extended 1.4" horn (as in the photos), only made axisymmetric. I think we're lacking a proper driver model here, as the reality seems to be a lot more favourable in general.
Not all drivers work out equal -
Note, that if the IR window is to short, you will get smoothed SPL frequency response at the lower frequencies side.
Yes, I know, that was all the same, approx 5 ms.
This is the FANE CD131 with a longer window (~16ms) including some of the reflections, closer to the horn:
This is the FANE CD131 with a longer window (~16ms) including some of the reflections, closer to the horn:
Last edited:
Intuitivelly I would expect measured SPL like in your simulations above (with a sharp peak at the lowest resonance of the pipe). It seems to me, that it is worth to try to increase the IR window.
Hmm, I wouldn't say the difference is significant with longer windows. And it is not clear what conclusion follows from this. The trivial conclusion is that the simulations oversimplify reality a bit, although the picture is qualitatively quite plausible (perhaps some acoustic losses in the compression chamber/phase plug should be included in the simulation?)
RTA, white noise averaged, no smoothing.
My conclusion is that this is unbelievable. 🙂 Considering it's a small 1" driver in a more or less constant-directivity horn.
Now I only wonder what a DFM-2544 could do.
- Definitely a good study material, obviously something must have been done right. I have no clue what it is...
And of course it would be nice to have some means of modeling this for any existing driver, by measuring its parameters first, to be able to optimize the whole thing.
My conclusion is that this is unbelievable. 🙂 Considering it's a small 1" driver in a more or less constant-directivity horn.
Now I only wonder what a DFM-2544 could do.
- Definitely a good study material, obviously something must have been done right. I have no clue what it is...
And of course it would be nice to have some means of modeling this for any existing driver, by measuring its parameters first, to be able to optimize the whole thing.
Last edited:
Here's the impulse response in various formats. Maybe someone else has some better tools for an analysis.