As someone who would likely prefer final acoustic slopes no higher than 3rd order, I'm interested in this topic.
I've observed that 1" compression drivers on horns often exhibit a roll off at about 26 dB per octave below their natural cutoff. Therefore the only way to achieve a 2nd or 3rd order roll off would be to cross well above the cutoff (at least 1 or 2 octaves), and live with the steep slope further down.
However, I've noticed a few compression driver/horn combinations, usually of the larger variety, that have a shallower roll off. For example, the Celestion Axi2050 mounted on an SEOS 30 (from Celestion's spec sheet):
As you can see, the roll off is only 12 dB per octave, starting at about 600 Hz. While other horns of similar size may often begin to roll off at or below 400 Hz, those other horns usually have a much steeper cut off slope.
So, is this shallower roll off characteristic achievable from a smaller 1" exit waveguide? If so, how? And are you aware of any available horns/waveguides that accomplish this?
One more thing:
Didn't Geddes use a 6 dB/oct electrical filter on the compression drivers in his speakers? I had always assumed at least a 12 dB/oct filter was basically mandatory on 1" exit compression drivers, even in a home/consumer setting. But if that worked well for him, I don't see why not.
Any thoughts?
Thank you. 🙂
I've observed that 1" compression drivers on horns often exhibit a roll off at about 26 dB per octave below their natural cutoff. Therefore the only way to achieve a 2nd or 3rd order roll off would be to cross well above the cutoff (at least 1 or 2 octaves), and live with the steep slope further down.
However, I've noticed a few compression driver/horn combinations, usually of the larger variety, that have a shallower roll off. For example, the Celestion Axi2050 mounted on an SEOS 30 (from Celestion's spec sheet):
As you can see, the roll off is only 12 dB per octave, starting at about 600 Hz. While other horns of similar size may often begin to roll off at or below 400 Hz, those other horns usually have a much steeper cut off slope.
So, is this shallower roll off characteristic achievable from a smaller 1" exit waveguide? If so, how? And are you aware of any available horns/waveguides that accomplish this?
One more thing:
Didn't Geddes use a 6 dB/oct electrical filter on the compression drivers in his speakers? I had always assumed at least a 12 dB/oct filter was basically mandatory on 1" exit compression drivers, even in a home/consumer setting. But if that worked well for him, I don't see why not.
Any thoughts?
Thank you. 🙂
Last edited:
You're just looking at the frequency response. The thing is, that's at and below the fs of the most drivers. That means, you will have big decay ridges. And the horn doesn't load anymore, which increases the excursion of the diaphragm, which increases the distortion. Plus, at these frequencies your horn doesn't have any dispersion control anymore.
To avoid that, you need to use a bigger and more potent driver with lower fs and/or a bigger horn. That said, the SEOS 30 is a very big horn (with a moderate gain) and the Axi2050 is a very specialized driver which can be used far below the cut off frequency of the horn (aside from the fact it's not decorated with the most modest price tag).
I don't know what he did. With low spl requirements and a big enough driver that can work but that's more likely to fail than to work.
Hi, if I remember correctly, Geddes first order highpass is set very high, somewhere 10kHz or so, which now flattens the down sloping constant directivity response as well as providing enough high pass. If it was 10kHz, then for example 625Hz is four octaves down so -24db attenuation, seems fine for home use. I'd choose rugged driver.
Edit. looking at mabats sims on the ATH thread, it looks like loading makes peak in the response which makes the knee sharper, frequency response graphs look similar as closed box bass vs reflex. Loading lowers excursion down there I think, so less current and excursion and distortion is affected. Side product seems to be bit more complex crossover, which is no biggie if DSP. In the end the driver runs out of xmax somewhere no matter loading.
Edit. looking at mabats sims on the ATH thread, it looks like loading makes peak in the response which makes the knee sharper, frequency response graphs look similar as closed box bass vs reflex. Loading lowers excursion down there I think, so less current and excursion and distortion is affected. Side product seems to be bit more complex crossover, which is no biggie if DSP. In the end the driver runs out of xmax somewhere no matter loading.
Last edited:
Here link to Mabat ath thread, see this page for example for some throat impedance and frequency response graphs. https://www.diyaudio.com/community/threads/acoustic-horn-design-the-easy-way-ath4.338806/page-765
I wasn't putting emphasis on the specific roll off frequencies shown, but rather the shallower roll off characteristic of the horn/driver combination and if that could be achieved with a 1" exit driver, albeit at a higher frequency.
I've also noticed an 18 to 20 dB per octave roll off of typical 1.4" exit drivers with 3" diaphragms in waveguides such as the 18Sound XT1464. Still shallower than the inherent 24+ dB seen with 1" exit horns.
tmuikku: Makes sense, but I still wouldn't have thought it'd be that high. I guess their response sloped down more than I remembered. I guess it's probably been 15 or so years ago. Thanks for the info.
I've also noticed an 18 to 20 dB per octave roll off of typical 1.4" exit drivers with 3" diaphragms in waveguides such as the 18Sound XT1464. Still shallower than the inherent 24+ dB seen with 1" exit horns.
tmuikku: Makes sense, but I still wouldn't have thought it'd be that high. I guess their response sloped down more than I remembered. I guess it's probably been 15 or so years ago. Thanks for the info.
Mabat's waveguides are less constant directivity so have less of the rolloff.
Can't you relieve or restore the lower frequencies with your filters?