Sub crossover point

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
It's not so much a question of the diameter, really, as it is how high or low the driver goes, and how fast the driver is. I prefer to leave at least one octave leeway between the crossover point and the rolloff of the speaker. More would be nice, but life is rarely so kind as to give you the option.
As an example, the Tympani IV woofer panels I'm using are good down to about 35Hz or so, and very quick compared to cone drivers (though not as good as electrostats). I cross over to the subs at 70Hz so that the slope (as determined by the active crossover) has a chance to get established before the woofer panels start pooping out. The result is a compound slope--roughly speaking 18 dB per octave between 70 and 35Hz, and 24-36 dB/oct (or whatever--I'm not sure what the T-IV rolloff slope is--lotsa variables there) below that, once the drivers start rolling off. I'm going to do some more intricate things with my crossover when I get time.
Another option is to let the woofer rolloff naturally and cross the sub in at that frequency. The only commercial example of this that I can think of was the old Wilson WAMM, back when David was still using KEF B-139s for the woofers. They're good down to about 40Hz, at which point he just let them fade into the sunset (sealed enclosure--nice, tame 12 dB/oct), but the sub (an 18", I think) was brought in at exactly the right frequency to mate with the KEFs. Sadly, the KEF B-139 was discontinued, so he was forced to redesign using other drivers; I don't know what strategy they use now. As usual, there are tradeoffs. On the one hand, you get rid of a crossover point, on the other, you're not limiting the signal going into the amp that drives the woofer, so you're wasting power. But it's a cool idea when it works.
15" and 18" inch drivers are rarely going to be fast enough to do you much good over about 50-60Hz. 12" (what I use) are good up to...maybe...80-100Hz. 10" and particularly 8" drivers start getting pretty peppy. You might consider an 8" as high as 200-250Hz, but it'll take a <i>lot</i> of them to move the same amount of air that larger drivers can manage. Yes, yes, yes...I know the specs show that they're good to 500 or 1000Hz. I'm not talking flat response (although dispersion problems can creep in), I'm talking transient response. How quick are they when someone hits a drum head? No one specs this sort of thing. What you'd need would be an impulse into the driver, and a measurement comparing the rise time of the wavefront coming out of the driver. Not a waterfall plot, which is what it sounds like I'm describing, just the rise time. But I've never seen such a thing (the numbers would be abysmal if viewed as distortion percentages, just the nature of the beast) and lacking that information, we just kinda have to get by as best we can.

I have no big experience with active crossovers (never liked it very much either), but in case of a system with built-in passive sub here is what I can say gives the best results:

- use a bandpass box to mechanically filter high frequencies
- help the bandpass with a 1st order crossover to filter out the resonances (normally you have a peak around 400Hz, that you can eliminate with a 1st order tuned at, say, 150 Hz)
- Be careful to linearize the impedance peak of of the bandpass (or any other alignment...) with a RLC net, otherwise your filter will likely interact with the system resonance and you'll end up with a very high Q factor of the system
- Let the bass-mid driver 'die free' like mr Wilson did: you waste some power but you save a lot of other troubles, and at least you don't have a big capacitor in series to your precious signal.

With a bandpass enclosure you can 'play' with the tuning in order to match the slope of the bass-mid, and also you can simplify your life with the crossover filter.

A drawback of this design is that at lower frequencies you have the 2 speakers almost in parallel, and the system impedance will be very low. There is some trick to bring impedance module to acceptable values, but they heavily depend on the system project, I don't know a general rule for that.

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.