I have designed an active crossover using two 2nd order linkwitz riley filters Q= 0.5 cutoff freq. 2.5kHz. Now the signals after the crossover go into two separate amplifiers and finally to the speakers where they combine into the air. So if we sum the two output we get a deep notch at the cut off frequency, but if we invert one of the signals and sum them we get a completely flat response. The question is: if I use an inverting circuit for the time delay of the bass then the bass will be 180° out of phase and the treble will be in phase so then they will sum flat. is it right?
an other way will be make the outputs in phase with the input and then connect one of the speakers with the polarity reversed but i would prefer the first even though one of the output is not in phase with the input.
an other way will be make the outputs in phase with the input and then connect one of the speakers with the polarity reversed but i would prefer the first even though one of the output is not in phase with the input.
It would seem that the inverting delay will get your phases back in line.
That would be the simplest and it allows for correct phase matching at the input and correct phase matching at the output.
You don't need to remember to swap phases when you swap gear.
I hope I have got that right, can anyone else confirm?
That would be the simplest and it allows for correct phase matching at the input and correct phase matching at the output.
You don't need to remember to swap phases when you swap gear.
I hope I have got that right, can anyone else confirm?
The crossover you describe, (two cascaded second order Butterworth filters giving a Q of .5) is the normal Linkwitz Riley fourth order crossover. This is called an "in phase" crossover, meaning that no phase inversion is required to bring the HP and LP outputs into phase.
What you seem to be describing are second order crossovers where the HP and LP are 180 degrees apart requiring phase inversion of one output to avoid the cancellation null.
Keith
Edited PS, are you building a 12dB/octave crossover or a 24 dB/octave?
What you seem to be describing are second order crossovers where the HP and LP are 180 degrees apart requiring phase inversion of one output to avoid the cancellation null.
Keith
Edited PS, are you building a 12dB/octave crossover or a 24 dB/octave?
Last edited:
With just an inverting circuit on the woofer side should bring both outputs into a state that adds to get a flat response. However if the delay is a RC type circuit shouldn't it affect the phase vs frequency around the crossover point. Would it now sum flat ? I guess a simple sim of this should give the answer.
An LR2 active circuit doesn't guaranty LR2 acoustic output. Have you measured the driver(s) response?
two 2nd order LR would be, one 2nd order LR for the low pass and one 2nd order LR for the high pass. This would give the inverted phases at the crossover frequency.
A 4th order LR crossover would require four 2nd order Butterworth (Q=1/sqrt(2)).
A 4th order LR crossover would require four 2nd order Butterworth (Q=1/sqrt(2)).
I confirm. The two outputs of the linkwitz-riley filter have opposite phase and you will have to invert one to get a flat sum. Which one and where is not really important.
The problem you will encounter is, that the transfer function of the drivers will add to the electronic filter and the acoustical result will be anything, but not the desired linkwitz-riley filter...
The problem you will encounter is, that the transfer function of the drivers will add to the electronic filter and the acoustical result will be anything, but not the desired linkwitz-riley filter...
I have two 2nd order filter separate one for the bass the oder for the treble 12dB/octave.
so is right to do the first way i described?
so is right to do the first way i described?
are all your amps exactly the same ?
but its like puppet said, varying acoustic output/response plays the major part
and actual speaker design
you may have to trust your ears, or measure
but its like puppet said, varying acoustic output/response plays the major part
and actual speaker design
you may have to trust your ears, or measure
Something to keep in mind.
Reversing the polarity of one driver does effect the sound.
While the phase will be correct at the cross over point, the phase will be 180 degrees off at all the other frequencies.
Some people are more sensitive to this than others.
Reversing the polarity of one driver does effect the sound.
While the phase will be correct at the cross over point, the phase will be 180 degrees off at all the other frequencies.
Some people are more sensitive to this than others.
The whole thing is complicated by acoustic offset.
The exact L-R characteristic assumes that the acoustic centers are aligned. If the driver roll off is at least an octave either side of the crossover then the driver characteristics have a minimal effect upon the result, but in may cases for instance you can get close to a quadrature phase shift due to acoustic offset.
If this happens either in or out of phase connection gives the same amplitude although group delay is different.
rcw
The exact L-R characteristic assumes that the acoustic centers are aligned. If the driver roll off is at least an octave either side of the crossover then the driver characteristics have a minimal effect upon the result, but in may cases for instance you can get close to a quadrature phase shift due to acoustic offset.
If this happens either in or out of phase connection gives the same amplitude although group delay is different.
rcw
In an average 2-way speaker the signal of the tweeter arrives about 0.1 ms before the woofer signal. If you want to insert an electronic delay, that makes only sense in the tweeter path. The delay changes the phase at the crossover frequency and you will not necessarily have to invert one path. A delay of half a wave length has the same effect.
- Status
- Not open for further replies.