Thank you for taking your time. I can't possibly know what causes the difference in the phase response but the delay still makes sense to me since the drivers still have some offset in spite of the waveguide.
I will try and figure out what is going on but my hopes aren't that high as a complete beginner, this is a little overwhelming for me to be honest. Looks like I am gonna stick to that delay and hope for the best.
Thanks again for your input witwald, I'll let you know when this project comes to an end.
I will try and figure out what is going on but my hopes aren't that high as a complete beginner, this is a little overwhelming for me to be honest. Looks like I am gonna stick to that delay and hope for the best.
Thanks again for your input witwald, I'll let you know when this project comes to an end.
I am actually using the factory data for the .zma files, and I am assuming you are not interested in them but I am uploading the files just in case...
You think the problem might be the fact that I don't take my own impedance measurements?
The factory files should be okay in the frequency range above 250 Hz. The tweeter is a sealed unit, and the woofer impedance shouldn't be affected by the enclosure loading until below 100 Hz or so. I just wanted to have all the available data at hand as part of the overall simulation strategy.
The offset distance needed to get the best simulated combined response is 0.295 inches (0.00749 meters). This corresponds to a time offset of only 0.00749/343*1000 = 0.0218 milliseconds. At 15 kHz, the distance corresponds to 32.8% of the wavelength of sound. That time offset looks to be close to the delta time between individual samples when a sampling rate of 48 kHz is used (1/48000*1000 = 0.0208 milliseconds). That's not a big offset at all.
The offset distance needed to get the best simulated combined response is 0.295 inches (0.00749 meters). This corresponds to a time offset of only 0.00749/343*1000 = 0.0218 milliseconds. At 15 kHz, the distance corresponds to 32.8% of the wavelength of sound. That time offset looks to be close to the delta time between individual samples when a sampling rate of 48 kHz is used (1/48000*1000 = 0.0208 milliseconds). That's not a big offset at all.
That's good to know...
I got the same result when I loaded .frd files as is (without the Hilbert Bode transform). I also realised when I am implementing the h.b. function, I have to apply a completely different value to match the inverted summation and the inverted measurement, which doesn't seem right.
So I guess I should just ditch using the H.B. transform function and load the .frd's as measured, and apply the 0.295 inches mod-delay.
Just wanted to let you know that I've made a prototype xo and It measures pretty good with a satisfying reverse null
Once again, thank you very much for helping me figure out the correct value for mod delay. 🙏 
@hautparlurker Thanks for sharing your results. It appears that you've achieved a good blending of the midwoofer and the tweeter filtered responses. You appear to have gone for a crossover design where the two drivers are in phase through the crossover region. Hence, your measured deep null indicates that you are achieving that behaviour to a large extent. Of course, if you'd chosen another type of acoustic filter response function, say a 3rd-order acoustic Butterworth, then changing the polarity of one of the drivers would not have any effect on the summed response.
Have you performed any listening tests? The measured response is tilted slightly upwards towards high frequencies. For example, the 6–9kHz region seems to be about 2 dB higher than the response in the 1–3kHz region. The loudspeaker system may sound a bit bright when listened to on axis.
Have you performed any listening tests? The measured response is tilted slightly upwards towards high frequencies. For example, the 6–9kHz region seems to be about 2 dB higher than the response in the 1–3kHz region. The loudspeaker system may sound a bit bright when listened to on axis.
At first I thought it would be cool to have it on the bright side a little, but someone else also pointed that out in another post and I ordered 6.2 ohm resistors for the R4 instead of 6.8, which will bring that region down a dB or so. I haven't listened to the speakers yet because I am still working on the finish of the cabinets, so hopefully by the end of next week...