Duelund synchronous crossover

[angel]

Has anyone here had a closer look at Steen Duelund's synchronous phase crossovers? They were mentioned in an earlier thread, but no-one posted anything regarding experiences with these filters, or for that matter comments on the theory.

They seem rather interesting, as the resulting phase graph for a 3-way speaker looks like three overlapping curves that represent a nearly constant group delay, with correct amplitude summation, as with Linkwitz-Riley filters. Should be particularly interesting for active crossovers, as the passive ones tend to grow immensely expensive.

For more information, try
1) http://www.meta-gizmo.com/Tri/speak/STEEN.html
or
2) http://tkhifi.homepage.dk/due3vejs/d...simu3vejs.html

The former is in English, and explains the theory (scroll down a bit for the math), while the latter is in Danish, and shows a theoretical example by another dane. You will have to bear with him on the rants; he can be a bit eccentric at times, but who isn't?

[phase_accurate]

At least the simulated step response doesn't look that special. The step response of really interesting crossover would sum to the same step that has been used as stimulus.

Regards

Charles

[AuroraB]

The first link came up looking rather funny on my screen, but-
this looks suspiciously like what was referred to as "phase linear" or a "filler driver" crossover,- a.o. adopted by B&O in the late 80-s, albeit Duelunds version is quasi 4th order LP/HP and the filler 2.order.
The original idea was quasi 2 order LP/HP and 1st order filler................This filter was described in an article in "High Fidelity" sometimes late 80-ies........

I built a speaker in -89, with a x-over like this ( and SC4 ported bass), which came out very good,- ( still plays very good in fact....)

As a funny point,- the sim's im ref.2 are madeusingMicrocaps v.6,---- I used Microcaps,v.3 in -89.......

[phase_accurate]

One has to watch out: Please state clearly what you mean with a first order filler driver! First order bandpass filters do not exist since they have to be made of at least a first order lowpass AND a first order high pass, resulting in at least a 2nd order bandpass.

The simulation shows definitely NOT a transient perfect behaviour.

OTOH it is well known that it is possible to add a filler driver path to a 4th order 2-way crossover. But you have to be aware that the filler driver now must be capable of delivering 6dB more SPL than the woofer and tweeter (in terms of efficiency and max SPL).

Regards

Charles

[AuroraB]

I am aware of that,- the filler driver is quite naturally driven from a 1st order BP,
but by definition, a 1st order BP has 6dB slopes, and a 1 order LP and HP section. A 2nd order BP will have 12 dB slopes.............
As the original article describes, the filler principle relies on the 2nd order LP/HP being realized with "skewed" component values , in order to shape the stop band contour to match the fillers response. Mathematicaly, the 2nd / 1st order combo sums to linear phase, as I recall....

[phase_accurate]

but by definition, a 1st order BP has 6dB slopes, and a 1 order LP and HP section.

Regarding the terminology I disagree with you: A BP with 6 dB slopes is definitely called a 2nd order bandpass.

Regards

Charles

[Volvo_Victim]

Almost all dynamic drivers naturally exhibit 2nd order mechanical HP and LP unfiltered roll-off characteristics towards each end of their frequency range.

Adding a first order electrical filtering effectively creates a third-order band-pass function as the mechanical 2nd order roll-off sums with the 6db/octave electrical filter.

Having said that, the filter remains effectively first-order (6db/octave) throughout the portion of the frequency response which is flat, then sums and rolls-off 3rd order with the drivers natural response.

This is particularly important to remember when listening off-axis, as all dynamic drivers roll-off significantly earlier at high frequencies off-axis.

[john k...]

An easy way to think of the Duelund 3-way crossover it that it basically a deviation of and LR4 2-way. That is, in the limit it reduces to an LR4 2-way. The phase response is always a derivative of the LR4 2-way with 180 degree rotation at the center frequency of the band pass filters. Note that the denominator of the 3-way transfer function in factored form is

(S^2 + aS + 1) x (S^2 + aS + 1) which is just the detonator of a 2nd order filter squared with a = 1/Q. So what Duelund has, in simpler terms, is HP and LP sections which are formed by cascading two similar 2nd order HP or LP sections. For example the HP section is

HP = s^2 / (s^2 + s/Q + 1) x s^2 / (s^2 + s/Q + 1)

the LP = 1 / (s^2 + s/Q + 1) x 1 / (s^2 + s/Q + 1)

and the BP transfer function is

BP = -1 / (s^2 + s/Q + 1) x s^2 / (s^2 + s/Q + 1) x G

where G is the BP gain = (1/Q)^2 - 2

When Q = 0.707 the HP and LP sections are LR4 and the BP section has G = 0.0.

In any event, the on axis phase and transient characteristics are similar to the LR4 2-way but off axis, for real systems with noncoincident drivers, the response degrades differently than the LR4 2-way.