DIY cabs: Simple question about horns and crossovers.

[parapentep70]

Hi all (and specially DukeLeJeune )

I am building a pair of 1x15 cabs, one of them with a horn, the other without. I work on this every now and then, so it is taking so many weeks.

When I started to simulate crossovers (with the goal to re-use or modify a good PA cross-over for sub-optimum performance) I discovered 2 or 3 things:

1) This is anything but trivial, crossovers can degrade a bit a good cab with a good design, or ruin it with a bad design.

2) I've got a couple of subtle additional problems due to the particular impedance curve of my compression driver, as DukeLeJeune anticipated weeks ago. Fortunately this Z curve seems to come from the UNLOADED compression driver, I measured impedance in the lab with the horn and things did change for the good.

3) In traditional 2-way designs with 2nd order filters, tweeter and woofer are connected with opposite polarities. This way, at the cross-over frequency phase lead / lag are +90 and -90 degree, so at this frequency sound from both drivers is in phase.

However, when using my horn it will generate a wavefront ("sound") some 15 cm deeper than the woofer. This is some 440us of delay. And this is almost exactly just half a cycle at my crossover freq. (1200 Hz) !!

So, at least in simulation, I get a much better blend if I DON'T reverse polarity in the tweeter.

Above crossover I get some "comb filtering", but I can limit it to less than +/- 1.5 dB. And this will be "on axis": My response "off axis" will be much better in this range (1.5 to 3 KHz). Simulated "combing" is negligible at higher freqs.

Is this a well-known effect and I am re-discovering the wheel? In case the subject is not in "cabs for dumbs manual", does it make sense? Anybody out there considering the effect of delay from the depth of horns?

Thanks a lot!

[will33]

Signing on for a bit of education here.

[craig.p]

It is well-known among those who have studied up on their theory. With the passives we had "back in the day," it was very common to stagger, physically, drivers running in different ranges so they didn't cancel in the (hopefully narrow) common range, but I believe today it's more common to introduce electronic delay.

The question is whether anyone, including you, will notice any "wrongness" at 1.2 KHz. Especially once the rest of the players show up and raise the noise floor and thus reduce the threshold likelihood of audibility of that wrongness. In other words, little concerns like these often become in-context N/As. I proved it in the field decades ago with a little experiment on a big PA using a two-way crossover with a variable-delay knob. I could spin that thing to any setting I wanted and no one could tell the difference -- and that was down at either 500 or 800 (can't remember which), and it was for voice, too, which should've been far more adversely affected by a delay anomaly than BG.

[billfitzmaurice]

Originally Posted by parapentep70

Is this a well-known effect and I am re-discovering the wheel?

By Harry Olsen, Paul Klipsch, Robert Novak and many others in the 1930s and 1940s. Time align of the woofer and HF driver was the impetus behind the Altec A7 and the like.