End-Fire Array with All Pass

Frank Koenig

Frank Koenig

Sophomore
Mar 7, 2011
187
0
16
Palo Alto, CA USA
www.dunmovin.com
You can build a two element end-fire sub array using a first-order all-pass instead of pure delay. The on-axis response looks pretty good on paper with about 18 dB rear rejection over a useful passband. It almost seems like a compromise between an end-fire and a cardioid array.

The example graphed below has 4ms of spacing between the speakers and the center frequency of the all-pass, at which it has 90 deg of phase shift, is 62.5 Hz, corresponding to 1/4 wave spacing. Neglecting attenuation due to distance, the front summation at the center frequency is total (6 dB) since the 90 deg from the filter cancels the 90 deg from the spacing. (I found that the first zero in the front response can be pushed up by adding a small amount of electrical delay without messing up the rest of it too badly.)

So does anyone use this? And how does it work in practice? When I get a chance I'll take my TH-Minis out to an open area and try it.

AllPassSubArray.jpg

F is front (audience side)
B is back (stage side)
A is all-pass filter response

Reinventing the wheel instead of cleaning the house.

--Frank
 
Re: End-Fire Array with All Pass

I have thouht about trying it at some time but never had suficient equipment to try up to 2 weeks ago .
I purchased 2 APEX intelli-x processors for some personal projects and for giving a short coarse in doing measurements .
Within each output channel i have the possibility of 12 filters wich can be ap1 or 2 besides al normal eq filters .
As soon as i have time and 4 subs i can try it and post some screen shots if you like .

It might be a bit difficult because a ap2 filter is symetric and delay causes phase wraps that get narrower as frequency rises . With just 2 subs i do not think it will be a problem but with 3 or 4 ..... do not know (yet)
 
Re: End-Fire Array with All Pass

Timo,

Thanks for publishing all those measurements. I think we all appreciate how much work that represents. I confess that it's a little hard for me to follow exactly all the experiments you did from looking at the graphs, but my take-home is that, in general, these phased arrays do what we expect and that the various implementations (all-pass vs pure delay, for example) make relatively small differences in practice. Is this a correct assessment?

As we say in the States, there are many ways to skin a cat. (Which is not to say that the pursuit of the various ways is not interesting.)

Best,

--Frank
 
Re: End-Fire Array with All Pass

Frank Koenig said:
Timo,

Thanks for publishing all those measurements. I think we all appreciate how much work that represents. I confess that it's a little hard for me to follow exactly all the experiments you did from looking at the graphs, but my take-home is that, in general, these phased arrays do what we expect and that the various implementations (all-pass vs pure delay, for example) make relatively small differences in practice. Is this a correct assessment?

As we say in the States, there are many ways to skin a cat. (Which is not to say that the pursuit of the various ways is not interesting.)

Best,

--Frank
Click to expand...

As Timo says in the last frame, pick your flavor. It appears that the 2nd order all-pass implementation yields 3-5dB more rear attenuation from about 55Hz to 90Hz, but the forward response is nearly identical.

Great work, Timo, and thanks for putting up the screen caps. You rock.
 
Re: End-Fire Array with All Pass

So first : the mic at the back of the array was about less then a meter away from the back sub so if i had had more room the mic at the back would/should have been roughly at the same distance from the back sub as the front mic from the front subs was . This means that the cancels at the back were deeper but i'm just guessing on that .

With the all-pass thing you have more control over the cancels in the back by playing around with the 2nd and 3rd row of subs in the picture (S2 and S3) .
So if you're on a small time window for the alignment use delay (maybe along with allpass filters so a mix of both have not tried that).

If given enough time and the noise pollution police eager to give you a fine for to much level breathing down your neck (10.000 euro's in the Amsterdam city centre for example) it can give you just enough control to keep them in there offices (that'll be the day;-)
@frank thanx and yes it was a lot of work but a very interesting experiment .
@tim i'm just curious (and rocking all the time if i get the chance)
 
Top Bottom