What Xover order is better?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I may be wrong here but I think that in the case of 1st order filters a Butterworth is a Bessel isn't it?.

In some filter tables you will indeed find 1st order Bessel, 1st order Butterworth ..... for the same 1st order filter.


If you look at the cutoff slope you can establish the rollof rate and the frequency of the filter. Butterworth's are down 3dB at this frequency, Bessels are down a little under 5dB I think?

You are talking about what happens at the pole frequency of a 2nd order filter.
But you will also find filter tables giving the pole frequencies and Qs for the sections of other types than Butterworth that are referenced to -3 dB.

Regards

Charles
 
Phase_accurate,

I'm not certain about the first order labels so I'll pass on that for now. With a second order filter, the Q is the linear magnitude of the response, relative to the passband, at the point where the ideal slope reaches 0dB.

Could I be wrong but I thought the label for a filter slope was directly indicated by this Q?
 
Sometime in the last 10 years Neville Thiele (of TS fame) did an artilcle on Cauer/elliptical filters for cross overs in the JAES. There are passive high level formulas and active low level circuits with differing 'm' or 'k' values (can't remember which). They looked interesting and one point that I remember was that he said that from a phase point of view they could be used with the LR driver offset etc. I'll post the exact reference if I find the p/copy. Also if you put cauer/elliptical cross overs in Google I think you'll get another similar paper at some guy's site in The States. This also has practical high and low level examples.

BTW it might help if we agreed on -3db for the place were we specify the bandwith of a filter. I think that this half-power frequency is pretty much universally agreed on by the professional world. (having said that I am aware that in the case of LR circuits they are each 6db down at transition but that is for the cross over and not a specification of the individual filters sections)
 
Yes Jonathan,

The 3dB point is considered a universal cutoff point as it is half power down. The bessel does roll off earlier than a Butterworth for the same filter frequency, and a chebychev rolls off later.

The filter frequency is the point that the phase is in the centre of its transition (90 deg for 2nd order, 45 deg for first). This is the text book crossover point for any filter type. You choose your type according to what response you want at that point, as well as phase roll off rate desired.
 
Jonathan Bright

Here's the reference :

Loudspeaker Crossovers with Notched Responses
Volume 48 Number 9 pp. 786-799; September 2000
---A class of crossover systems is described which produce null responses in their high-pass and low-pass outputs at frequencies close to the transition crossover frequency. As a result, both outputs have a high initial rate of attentuation in their stopbands, while the sum of their output has a flat all-pass response. When the nulls are moved to very high and low frequencies, the transfer functions degenerate into Butterworth functions for odd order, and into Linkwitz-Riley functions, for even order. Active and passive realization are presented. The work is the subject of a patent application.
Author: Thiele, Neville---

Crossovers using a notch are not new at all. The first Goldmund speakers used it, the designer being Christian Yvon, I think. There was an article by Bill Hardman in Electronics World in 1999 for an active implementation (there is a DiyAudio thread on it) and an other article in AudioXpress for a passive implementation about two years ago.


Sdclc126
---Crossover topology is dictated by speaker drivers, enclosure, design goals etc---

Philosophy towards crossovers can be different. One can ask : considering the room acoustics and the human physiology, are they crossover frequencies which would have less insidious effects than others ? Second question : having selected the frequencies according to such criteria, what is the better slope for them ?
Then, it is the crossover topology that dictates the drivers. This way to assemble the lot can even be easier than the traditional one.
 
Because I have built and listened to a crossover with a notch (Hardman's circuit), there is no doubt in my mind it 's a circuit much more tolerant than many others towards the selection of the drivers. Nevertheless I am not sure it's the best of all, and would like to discuss with anybody who has tried it and liked it or did not. No need of a forum to avoid arguing.
 
The filter frequency is the point that the phase is in the centre of its transition (90 deg for 2nd order, 45 deg for first). This is the text book crossover point for any filter type. You choose your type according to what response you want at that point, as well as phase roll off rate desired.

What you are talking of is the pole frequency. This has to be different for a given -3dB cutoff frequency for different filter types. And textbooks usually refer to the -3dB point as cutoff frequency and not the pole-frequency.

The best crossover will only work with flawless drivers

Since these don't exist we have to use the second best crossovers.

Regards

Charles
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.