phase correction in crossovers


2001-07-07 7:15 am
I was wondering if anybody had any experience with, or knew of any sites that might have any information regarding using phase correction in crossovers? There doesn't seem to be any sites on the internet with circuit patterns, equations or anything like that. I did find one guy using one particular circuit, but he didn't seem like he had any idea how to actually do it well. I know that you can use software and a cheap mic to find a driver's phase characteristics (to help design the circuit). It seems to me that speakers that use circuits like this to improve phase reponse around the crossover frequency have more transparent, less aggressive tweeters (not to mention smoother transistion from driver to driver). I've heard that the old Tannoy Speaker manuals basically taught phase correction circuitry, but then they realized that other manufacturers were reading them! Anyway, I'd love to hear what you have to say....Thanks,
This is a topic of interest to me as well.
If you're speaking of passive crossovers, phase correction is easy--time align the drivers by moving them back and forth during the design phase of your cabinetry. This leads to some pretty odd looking speakers, but that's why they invented grill cloth. There are lots of commercial examples of this.
To implement phase in an active crossover...ouch. The only way I know of is digitizing the signal, at which point you can delay it as much as you like. Unfortunately, all the DA and AD conversion is going to extract a penalty. This is generally done only for subwoofers.
You might consider all-pass filters, but be aware that they are only 'all pass' over a finite frequency range.
Horowitz & Hill show a variable phase circuit, but again, beware of getting past the 'all pass' region.

Siegfried Linkwitz gave full details for implementing phase correction in active crossovers (for time aligning tweeters/mid-range) in his article 'Loudspeaker system design' published in Wireless World in May/June 1978 with an updated version in Speaker Builder in Feb/March/April 1980. A copy of the article is available for download at his site:

Here is a snippet of some of my acoustics notes on crossovers.

For 3 way system.

"1 determine the crossover frequency and attenuation slope required by the high frequency device and design the necessary active high pass filter. Denote the transfer function Hh2.

2 Determine the crossover point and required slope from low to mid and design an active high pass filter with the required characteristics. Denote it as HH1.

3 Use the synthesized low pass technique to produce a low pass filter whoes transfer function will then be HL=1-HH1

4 Synthesize a bandpass by HBP = HH1-HH2"

The combined acoustic output will be all pass and have zero phase.

Any differentiations in path length, wind currents, driver imperfections, unmatched amplifiers, wires, etc. will of course alter the output.

If you are looking to correct minor phase differentiations, then speaker placement sounds like the easiest way unless you want to convert to digital.