Re: cross firing trap main?
http://www.falstad.com/ripple/
This is a Java Wave simulator. It's built for liquid waves, but it works just as well for sound, since air behaves the same way.
I tried to simulate 2 90 degree boxes at a right angle since this is the easiest thing to freehand. The first image is 2 cabinets in a traditional array. There's a bit of comb filtering here. This obviously changes with frequency. I just went with a default setting. The second image shows the same 2 cabinets toed in towards each other at 90 degrees. Notice there is a complete polarity flip in the middle, and the areas where the comb filtering was in the top image there is now almost complete cancellation. The middle polarity flip means you've screwed up your sub to mains integration and will have a hole in the response.
Those patrons on the sides will be plenty pissed if they paid a bunch of money to hear next to nothing.
Tim,
The cross fire messes up the phase response, not polarity, polarity is invariant with frequency. If you change the frequency in your wave model, the “out of phase” wave positions will change.
The cross fire idea works so poorly because it puts the maximum output of both cabinets in the area of greatest overlap, intensifying the phase cancellation resulting from different path lengths to the listening position.
Reflections off the adjacent cabinet face make for more cancellations, none of which can be EQed since they all change with angle of incidence and frequency.
At the low crossover point, say 100 Hz, splaying a pair of small front loaded speaker in or out will make little difference, as either will result in a spacing difference of less than 1/4 wavelength.
Cross firing of larger cabinets can actually be beneficial in the 100 Hz crossover region as it can reduce the center to center distance between cones to within 1/4 wavelength.
A “normal” out firing splay, with the HF horn 6 dB down point in the area of overlap will have far less comb filtering, and less deep nulls in the critical speech frequency range.
The screen shots below shows the on axis response of a single cabinet (10T EQ on). The off axis response for this cabinet is pretty uniform up to 8K out to about 40 degrees off.
The lower two traces show a pair of the same cabinets cross fired with the mic in the center of the cross fire region (10T V on), and 10 degrees off (10T V 10 off).
There is more than 20 dB variation in level, (the single cabinet was +/- 3 dB) and the null spacing changes with angle.
There is little change in the low frequency response, but above 400 Hz things go very badly, outdoors it was easy to hear how grossly the sound differed as I walked from left to right.
Unfortunately, I discarded the further off axis traces, but suffice it to say they looked even worse, and quite different at each 10 degree interval.
Indoors, the reflections from boundaries make things more complicated. If a single cabinet dispersion is wide enough, splaying a pair outward will result in wall reflections, in that case it is better to vertically stack the cabinets, usually with the HF together.
Art
