Discussion in 'Archived Threads 2001-2004' started by Travis G, Jan 7, 2002.

1. ### Travis G Stunt Coordinator

Joined:
Jan 6, 2002
Messages:
94
0
Trophy Points:
0
I am trying to learn to build fullrange loudspeakers and I've been reading some books but none so far have really explained the pros and cons of a Linkwitz Riley vs a Butterworth crossover. The last book said that a Linkwitz Riley crossover provides better damping. What exactly does this mean? Will a Linkwitz Riley crossover lower Group Delay? Will a second order L-R roll off at 13 dB per octave instead of 12?

2. ### Patrick Sun Moderator Moderator

Joined:
Jun 30, 1999
Messages:
38,749
480
Trophy Points:
9,110
Have you read Vance Dickason's "Loudspeaker Design Cookbook"? It's on its 6th edition, and available at places like Amazon, maybe Barnes and Noble.

A L-R is more damped because it requires to crossover frequency to be down 6dB (relative to the rest of the driver's overall output through its operating frequency range). It has a Q (describes the quality of the damping) of 0.5, which is considered critically damped. The damping is what makes the driver operate "normally" without introducing peaks (underdamped condition) in the SPL output. When you sum 2 drivers with L-R, they should sum flat at the crossover frequency.

A Butterworth filter is down 3dB at the crossover frequency. Its Q=0.707 (underdamped), and when you sum 2 drivers with Butterworth filters at the same crossover frequency, you'll get a 3dB peak/hump at the crossover frequency (but the acoustic response of the drivers may drag the hump down if you are judicious where you design for separate crossover frequency for each driver filter.

Electrical filters do what they are supposed to do, but you need to design them to work in tandem with the driver to take advantage of the driver's natural roll-off to get the "acoustical" response you are aiming for with your design.

3. ### BruceD Screenwriter

Joined:
Apr 12, 1999
Messages:
1,220
0
Trophy Points:
0
Patrick,

Are there any benefits to the L-R zero degree phase changes vs. the Butterworth's xx? degree phase changes?

Bruce

4. ### Patrick Sun Moderator Moderator

Joined:
Jun 30, 1999
Messages:
38,749
480
Trophy Points:
9,110
It just means getting the polarity of the tweeter hook-up correct w/r/t the polarity of the woofer hook up.

For a 4th order L-R, if you have the drivers wired correctly, there were not be a dip in the response at the crossover region, but reverse the polarity on one of the drivers, and you'll get a null/dip in the region (because one of the drivers will be roughly 180 degrees out of phase at the crossover region).

For Butterworth filters, the phase will be trickier to get right.

Each reactive component (inductor and capacitor) will contribute 45 degrees of phase each (inductors lag by 45 degrees, capacitors lead by 45 degrees). By using the typical parallel filters, the component phase add, thus a 4th order (order means the number of reactive components) is 4 times 45 degrees, or 180 degrees. So if one of the drivers is out of phase, reversing its polarity restore its phase. (I know this is confusing, and would suggest getting the Dickason book if you haven't already).

4th order L-R filters are mainly just 2 Butterworth filter connected in tandem (the trick is to come up with the component values for the filters).

2nd order Butterworth filters will introduce a 90 degree phase shift off the "real" axis of SPL output (if you get the polarity correct, else you'll get a null with out-of-phase polarity). Whether this is good or bad, you can't really say quantitatively, it depend on the overall design.

5. ### Travis G Stunt Coordinator

Joined:
Jan 6, 2002
Messages:
94
0
Trophy Points:
0
Thanks Patrick. That helps alot. It's not too confusing it's just that the only books that I can find at the library are vague. Not only that but they often have misinformation in them which makes it hard to know what to believe and what to disregard.

I did order the LDC last week friday though off of amazon. I tried to order it from Barnes and Nobel about a year ago but they couldn't get it in. I do have Loudspeaker Design Recipies Vol 1 but I don't fully understand it yet.

6. ### Jeff Mills Stunt Coordinator

Joined:
Feb 8, 2002
Messages:
95
0
Trophy Points:
0
I know this is an old post but I thought patrick may be interested in this information

0.707 is critically damped...not 0.5 as you mentioned

below 0.707 is underdamped and above 0.707 is considered overdamped.

You do have the theory somewhat correct though regarding the LR and butterworth....just a little wrong though. Its the underdamped nature of the LR that makes it exhibit a flat response without the hump at Xover freq..