1st Order XO with -6dB XO frequency. Has anyone tried this?

[primalsea]

Has anyone used 1st order HP & LP filters together with different -3dB frequencies so that when the cross over point of the amplitudes meet they are -6dB down instead of 3dB.

If my mental calulations are correct the HP will be set for 600Hz and the LP set for 150Hz. The crossver point will be 300Hz@-6dB.

This will mean that different values of the caps and resisters (active XO) for the HP & LP will be used. This will mean some compromises but hopefully they wont be as bad as those of a higher order filter.

Has anyone tried this? and did it work or was it a disaster.

Kind regards
Primalsea


Dont settle for shampoo, Demand real poo!!

[SY]

What is it that you're trying to accomplish? The problem you're trying to solve?

[phase_accurate]

I don't see any reason for doing so !

Regards

Charles

[primalsea]

Sorry I should have said.

As 1st order filters have a gentle roll off I'm concerned that they may not be suitable for my intended speakers as the attenuation outside the cross over point may not be enough. I dont want the drivers operating outside of their most linear frequency range. I understand this is the time old problem with 1st order filters.

However if the overlap frequency of the xo was -6dB down, like that of a Linkwitz-Riley filter there will be more attenuation of frequencies outside of the xo frequency. This, I hope will give some extra margin.

I foresee that there may be some issues with how the xo and drivers responses sum but not sure at the moment.

[sreten]

Offset crossover frequencies are used all the time in the
real world, to allow response shaping / corrections.

However there seems little point in the alignment you suggest,
the phase response of a 2nd order L/R (-6dB at c/o) is more
suitable for even power summing.

sreten.

[phase_accurate]

Looking at the crossover alone you will end up with a dip in the FR.

Depending on the natural rolloff of the drivers you can make it work. Some Dynaudio speakers for instance use an acoustical 3rd order crossover, using 1st order filters with xover frequencies that are separated by more than one octave. One of the drivers used has to be connected with reverse polarity in this case.

Regards

Charles

[primalsea]

I've read all sorts of reasons why 1st orders filters are best if your drivers can work with them they have become my filter of choice.

People say things like amplitude and phase response when summed are accurate duplicates of the original signal. Also square wave and transient response are best.

I current believe most of what I have read.

I have witnessed a LR 2nd order LP filter turn a distorted sine wave from my sig generator into quite a clean & tidy sine wave. I would hazard a guess that it would have more than a negligable effect on a music signal that a sine wave that constantly changes amplitude and frquency, a bit like the sine wave from my sig generator. As 1st order filters have the best square wave response I thought that if you can use then then you should.

[phase_accurate]

If you are willing to use more than just two simple RC filters you can also do higher-order transient perfect crossovers.

Regards

Charles

[kelticwizard]

What were you planning to do with the dip in the frequency response? Electronically equalize?

[SY]

Originally posted by primalsea
As 1st order filters have a gentle roll off I'm concerned that they may not be suitable for my intended speakers as the attenuation outside the cross over point may not be enough. I dont want the drivers operating outside of their most linear frequency range. I understand this is the time old problem with 1st order filters.

However if the overlap frequency of the xo was -6dB down, like that of a Linkwitz-Riley filter there will be more attenuation of frequencies outside of the xo frequency. This, I hope will give some extra margin.

I foresee that there may be some issues with how the xo and drivers responses sum but not sure at the moment.

Your understanding is correct- this IS one of the time-old problems with first order filters. Lobing is another. But if you do an early rolloff and you have to re-equalize to get back to a flat response, you've gained nothing as far as the driver is concerned.

First order filters actually work well in a very, very, very limited number of cases.