Trying to: 2-way Linear Phase Crossover

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

I guess I'll start by explaining what I'm doing.

I had been using a wideband driver and it's phase was mostly centered around zero degrees from the treble down to a few hundred Hertz. The stereo image and 3-D depth was fantastic.

The thing was, the wideband needed baffle step compensation and a tweeter. Adding even a first order crossover introduced phase shift. Along with increases phase shift I noticed diminished imaging and depth.

2nd order low pass
2 notch filters

1st order high pass
1st order allpass to to align the phase 5kHz-20kHz
voltage devider

There is also impedance correction to reduce impedance phase. None of these had any practical effect on the acoustic phase, other than the allpass aligning the phase where it was needed.


The point I'm getting to is that I noticed one design in XSim3D called the "Cosyne3 with AllPass practical" as shown in the next image below. It has a mostly linear phase down to 500Hz. How exactly is this linear-phase response achieved?

Last edited:

The first plot shows the latest combined response of both drivers. This is as good as I can get right now considering the response of the wideband driver.

The Green line in the center of the graph is the minimum phase response and the Grey in the center is the real phase. It's linear phase (phase coherent) and time aligned at ear height. Always a work in progress.

Last edited:
When I took the intial measurements there was little stuffing inside the cabinet. I ended up with some peaks and dips below 3kHz. I used a different stuffing technique and removed most of the ripple. I have one dip at 150Hz that I need to address with careful stuffing placement. I was able to change some of the notch filters to optimize the power response.

The phase below 200Hz begins to rotate positive with the bass tuning. While I can correct this with adding more drivers, it's not what this design was about. I'm satisfied with what I have here and am starting to wind the inductors to see how it compares.

Last edited:
Hello Art,

just found your thread by chance and am greatly impressed by your work! Achieving this with passive implementation is high art! Congratulations to your phantasic results! It would be interesting to see responses under angles as well :)

But more importantly: How would you describe the resulting sound in your system and room subjectively?

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