What’s wrong with metal cones? Nothing. It’s the crossover (that’s the problem). Metal cones and “soft” cones require different treatment . . . most crossovers that seem adequate with soft cones simply will not sound good with metal.

All cones break up (for a number of reasons, and in a number of ways). All cones beam (some more than others). Soft cone drivers address the issues with internal damping and “controlled” decoupling, to greater or lesser success . . . and sometimes designers incorporate the resulting cone behavior into the design of the crossover itself. With metal cones the *only* approach to breakup that works is avoidance. Complete and total avoidance. For metal cones incorporating the “breakup” behavior into the crossover fails in some pretty fundamental ways. One doesn’t see that mistake often with the SEAS Excel drivers, for example, because their first breakup produces an obvious (and nasty) peak. Just look at it and it screams “keep off me”.

Not so with the Dayton RS series drivers (RS225, RS180, RS150). Their first breakup produces a null. The natural (but false) assumption is “won’t hear it”, and the common mistake that follows is “treat it as part of the filter curve”, following the habit established with soft cones. One common result is a supposedly LR4 (acoustic) filter composed of a combination of a 3rd order electric filter and the “natural roll off” of the driver. And this often tempts the designer to push the crossover frequency *up* to better take advantage of the better-looking-on-paper response curve that results. But there really is no “natural roll off” with metal cones . . . there is just the on-axis (acoustic) null that is but one symptom of a nasty cone breakup. And moving the crossover frequency up only makes things worse . . . there’s more signal getting through to drive the breakup.

There are three simple rules for getting the best from Seas metal cones, and they apply to the Daytons as well: cross no higher than one third the first breakup frequency, cross LR4 electric, *and* put a notch filter on the first breakup (whether it’s a peak or a null) because the first two rules alone are often not enough . . . with a metal cone it is *critical* that there be no signal present to drive cone breakup. Such filters are easy to implement “active”, but become expensive in “passive” designs (although one can sometimes “fudge” a little on the second rule if the notch is carefully done, and one sometimes sees a cauer elliptic filter, or an electric 3rd plus a notch working reasonably well). But it makes the Baby Jesus cry to see a RS150 crossed 3rd order at 2kHz . . . if you’re going to do that use a AC130 instead . . .