Well, I would say that in most cases, you would want the crossover between two drivers to have the same order rolloff on both sides. If the tweeter has a 6dB/octave slope, and the woofer an 18dB/octave slope, before the woofer's rolloff, you will have a peak in the response because the tweeter will still be producing a lot of sound. If, however, the woofer you are using has a natural rolloff around the crossover point you want to use, you may be able to get a flat summed response. However, this is more difficult, and doesn't seem to be very common, which would explain why most people try to "mirror" the slopes at a crossover point. It is also much easier (in general) to cross at a point where both drivers have a flat response, if possible. You should always pick crossover frequencies and slopes based on the drivers you're using.
I've got a 3rd order on the woof and 2nd order on the tweet using seas excel drivers. Crossover was designed by madisound using LEAP. Might have something to do with minimizing the 4k-5k nasty breakups produced by the woofer, although they did include a notch filter for that purpose as well.
As mentioned, picking the crossover should be more dependent on the drivers, i wouldn't want to design a crossover and then try to find drivers that suit it!
As mentioned, picking the crossover should be more dependent on the drivers, i wouldn't want to design a crossover and then try to find drivers that suit it!
Hi:
Is one driver is crossed over at 6dB/oct, and the other at 18dB/oct. the response is given by the follow:
If your crossover is Butterworth
In a driver the phase angle is 45°, and in the other driver is-135°.
the total angle is 180°, then, you should invert the polarity of one driver.
Now, the angle is 0°, and because the butterworth has the cut point at -3dB (0.707, with 1 as reference).
The total responce is 0.707+0.707=1.41!!!(3dB peak!!!!), the add is direct because the phase angle is 0° .
Of couse this is theory, in a real world is more complex.
Is one driver is crossed over at 6dB/oct, and the other at 18dB/oct. the response is given by the follow:
If your crossover is Butterworth
In a driver the phase angle is 45°, and in the other driver is-135°.
the total angle is 180°, then, you should invert the polarity of one driver.
Now, the angle is 0°, and because the butterworth has the cut point at -3dB (0.707, with 1 as reference).
The total responce is 0.707+0.707=1.41!!!(3dB peak!!!!), the add is direct because the phase angle is 0° .
Of couse this is theory, in a real world is more complex.
Yes. The tweeter will have to be crossed very high. The woofer itself is crossed "steeper" than the Butterworth. I'm afraid that there will be too deep of a valley in the crossover region (shallow one is fine with me, and flatness is not a priority).
Why is it so? Are you telling me that it would be better if I crossed the woofer with 4th order?
Tweeter protection is not an issue because the Gainclone is the biggest power amp I have ever used since the last 12 years.
What I'm trying to do is to build a speaker where both drivers work in their optimum state, and no compensation component in the circuit. I've tried full range, 1st order, 2nd order, 3rd order, 4th order, and so now I want exactly a 3rd order on the woofer. I have 2 candidates in the 3rd order, and I'm looking for other candidates in 3rd+1st and 3rd+2nd (3rd+3rd has phase problem anyway). The 3rd+1st has the best transparency, the 3rd+2nd has the best vocal but terrible phase problem. I'm seeking for an advice whether to stop looking into the last 2 candidates and refine the first 2.
Good info thank you. I will look more deeply into the 3rd+2nd. Is the tweeter inverted?
Is it really so? I thought the roll-off has something to do with the voice coil to some extent. The 3rd order will put an inductor in SERIES with the driver coil.
Hmm, unfortunately I don't really understand about phasing and delay in the crossover. I know the physics, it's just I don't know how it relates to the crossover problems. This is the reason why I asked at the first place.
Ok, let me try to interpret what you explained. The single filter will delay the current, or simply said will have a 45 deg phase angle. The 3-filter one will have 135, so the difference is 180, hence the 0 deg of the inverted polarity. May be that applies ONLY in the x-over region (not span the whole freq) I don't know, but my ears tell me that it isn't the case. Third order doesn't have to be butterworth, and if it is not symetrical (as is the case), the 3dB peak is irrelevant I guess.
Jay said:
There is a power handling issue here as well. Most normal signals, such as music, have a spectral tilt of about -6dB/octave. Sometimes more, but almost never less. So there is a lot of energy at low frequencies, but less at high. This is the reason that in a 100W system the tweeter need and can take only about 10W. But if the slope of the tweeter filter is only 6 dB/octave, the amplitude of low frequencies (in typical music signals), after the filter, will be about the same as for high frequencies. This would limit the power handling capacity for normal music signals somewhat, and possibly introduce unnessecary distorsion.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.