For a while I've played with the idea of creating a speaker that is not a multiple entry speaker per se, but keeps the woofers outside the horn but fed through slots to ensure proper summing and directivity without the complexity involved with synergy-style speakers.
I recently read Scott Hinson's thorough document outlining his multiple entry horn (MEH). He uses the B&C coaxial compression driver and heavy duty 10 inch woofers to create what seems to be a fantastic speaker.
I thought i could use his general outline (driver selection, dimensions, tuning) but to see if it would work with the woofers outside the horn. The idea being that if the speaker would work well without the taps (quite large) it could improve the sound since the compression driver won't see the taps until the frequency is low enough for it not to matter. Since the horn won't load woofers at the crossover frequency (300hz), it seems it could be possible to build the horn this way:
At the throat is a expansion based Hughes' quadratic throat waveguide. The overall dimensions of the horn is otherwise similar to Hinson's design, around 580mm wide, 90x65 degrees expansion
To check effect of the slot loading on the frequency response, i created a model of the woofer chamber entry and exit and has a go at simulating this. I didn't model the driver (fixed driving) just to see the effect of the geometry:
The results on-axis show a few 3b reduction at 200hz, and the first cancellation dip at 430hz.
I was thinking that if the cancellation here is too low in frequency, this project is dead on arrival. So my question is:
Does this frequency response make sense in your experience, or did I do some serious modelling mistake? Intuitively, if we assume all the input comes from the center of the driver, that there should be a large at the half the wavelength of the center to edge and back (in this case the same as the driver diameter, around 10 inches). So i would expect the first cancellation to be at 340/0.25 which is 1360hz.
This is the normalized simulated responses if we assume the opening of the slots are driven (ideal source) and the compression driver (ideal source). Ideally i would like the slot to not affect the frequency response until much higher in frequency, so that i don't compound the directivity effects of the horizontally separated woofer sources and the cancellation effect of the slot itself (unless the slot itself can be seen as an acoustic filter that matches the crossover frequency).
So i really have two questions: 1) Is the idea sound at all? 2) Is my Akabak modelling or intuition way off? If so, what can i do to improve it?
I recently read Scott Hinson's thorough document outlining his multiple entry horn (MEH). He uses the B&C coaxial compression driver and heavy duty 10 inch woofers to create what seems to be a fantastic speaker.
I thought i could use his general outline (driver selection, dimensions, tuning) but to see if it would work with the woofers outside the horn. The idea being that if the speaker would work well without the taps (quite large) it could improve the sound since the compression driver won't see the taps until the frequency is low enough for it not to matter. Since the horn won't load woofers at the crossover frequency (300hz), it seems it could be possible to build the horn this way:
At the throat is a expansion based Hughes' quadratic throat waveguide. The overall dimensions of the horn is otherwise similar to Hinson's design, around 580mm wide, 90x65 degrees expansion
To check effect of the slot loading on the frequency response, i created a model of the woofer chamber entry and exit and has a go at simulating this. I didn't model the driver (fixed driving) just to see the effect of the geometry:
The results on-axis show a few 3b reduction at 200hz, and the first cancellation dip at 430hz.
I was thinking that if the cancellation here is too low in frequency, this project is dead on arrival. So my question is:
Does this frequency response make sense in your experience, or did I do some serious modelling mistake? Intuitively, if we assume all the input comes from the center of the driver, that there should be a large at the half the wavelength of the center to edge and back (in this case the same as the driver diameter, around 10 inches). So i would expect the first cancellation to be at 340/0.25 which is 1360hz.
This is the normalized simulated responses if we assume the opening of the slots are driven (ideal source) and the compression driver (ideal source). Ideally i would like the slot to not affect the frequency response until much higher in frequency, so that i don't compound the directivity effects of the horizontally separated woofer sources and the cancellation effect of the slot itself (unless the slot itself can be seen as an acoustic filter that matches the crossover frequency).
So i really have two questions: 1) Is the idea sound at all? 2) Is my Akabak modelling or intuition way off? If so, what can i do to improve it?
Last edited:
The usual slot load response is a peak followed by a null. The Hornresp predictions are quite accurate, I don't know enough about AKABAK to tell you why yours lacks the peak, but the null frequency looks correct for those slot dimensions.
Here is a similar slot in a push push subwoofer using 2x 12":
With the conical horn not providing much loading below ~800 Hz for the co-ax, I'd agree the slot cancellation frequency makes this DOA.
The center to center distance between the slots and co-ax entrance would also ruin the horizontal off axis response.
Thanks a lot Art! I will look into why the peak doesn't show up in the sims. Akabak 3 certainly seems easier to use than the previous ABEC, but it still has quite the learning curve.
I guess it would show the infamous peak if he would simulate a simpler shape for the LF waveguide, i.e. one that is of simple rectangular shape that is neither rounded in the back and which doesn't flare towards the mouth either.
Regards
Charles
Regards
Charles
Thanks phase_accourate, i did a second attempt with the externals of the body included and got this response. I suppose it's closer to the Art's example. It does seem like 10 inch woofers are too large to use slot loaded for this application. If we assume for a moment that they weren't (or that i used smaller drivers) that would push the peak/dip higher in frequency, i still don't quite understand why one couldn't have the woofer output outside the horn, as long as they sum constructively so that one is able to get a smooth directivity response: The argument is that since even the lower octave of the CD is not hornloaded, the crossover would happen in a "fully unloaded" region. The cancellation effect from the two slots could even be beneficial if it can be carefully matched with the directivity from the CD?
EAW uses side mounted slot loaded woofers on many of their MEH (multiple entrant horn) line array modules to extend horizontal pattern control down to ~250Hz:
The 4x5" mid drivers on the EAW CSA MEH are capable of a considerably lower crossover point than the B&C DCX464 on a 580mm wide, 90x65 conical horn.
To maintain ~1/4 wavelength spacing between the DCX464 and the side 10" exits, the crossover would need to be above ~250Hz, so possible if you cross as low as Scott Hinson's MEH which seems to be ~310 Hz.
Art
Last edited: