+3dB at the crossover point.

Well, I've pretty much finished designing my preamp/crossover and am just waiting for a few parts to begin building it. My questions is, I'm making my crossover point 600Hz, but at 600Hz, neither the high or the low is being attenuated, correct? Therefore there would be a +3dB increase at 600Hz because of both the low and the mids reproducing it, correct again? In that case would it be a good idea to attenuate by 3dB at 600Hz? And since I'm using a 12dB per octave crossover, it would be even better if the attenuation rolled of at the same, right?

Or would I be better off attenuating less then 3dB? I don't know if in reality it ever does achieve the +3dB bost, since the mids and woofers and physically quite a ways apart (relatively speaking that is).
 
Actually, for the majority of home listening, you will find that if each driver is -3dB, the nett effect will be +3dB @ the crossover point. So, each driver must be -6dB @ the crossover point to gain a flat response. Off the top of my head, this makes Linkwitz Riley alignment crossovers particularly useful as they are -6dB @ the crossover point. If the drivers are physically quite a distance apart, then each driver will need a little less than 6dB attenuation @ the crossover point but it will still be relatively more than 3dB.
 
kelticwizard said:
I seem to remember reading that if you are using 12 dB/octave crossovers, both drivers should be -6dB at the crossover point. But if you are using 6 dB/octave crossover, the drivers should be 3 dB down at the crossover point.

Of course, these were for a passive crossover system.

Butterworth alignment both drivers are -3dB @ crossover

Linkwitz Riley alignment both drivers are -6dB @ crossover
 

Nelson Pass

The one and only
Paid Member
2001-03-29 12:38 am
In point of fact, after you consider both the phase
and frequency response of both the filters and the
speakers, you can't count on anything. All you can do
is give yourself some wiggle room on the filter
parameters and adjust for the best result.