I’ve begun with the cross-over simulations based on the frequency measurements I did last weekend. So far everything is going well and it was quite easy to simulate the first prototype cross-over design, which I will build, measure and listen to.
There are a couple of challenges to deal with in this design. First of all the mid-woofer cone break-up needs to be taken care of properly. I will use steep cross-over slopes (24dB/octave acoustical) and a notch filter to reduce the 10kHz break-up mode.
I’m not fully satisfied with the conventional LCR notch approach, instead I will try an alternative method to deal with the break-up and to reduce the complexity of the filter. Distortion measurements and listening tests will decide which method I will use.
Second, I have to deal with the tweeters rising top-end response and how much attenuation it needs in order to get a well-balanced sound that doesn’t sound to bright, but still manage to present an open and airy sound.
Left: On-axis Right: 15deg off-axis
Left: 22.5deg off-axis Right: Individual driver phase integration
The 1m measurement indicates that the mid-woofer relative acoustic off-set is approximately 25mm (1”) compared to the tweeter.
The cross-over simulation is optimized for the 15deg off-axis measurement and at a 2.5m listening distance.
As can been seen I’ve used almost a full baffle-step compensation and the sensitivity is about 81dB 2.83v/1m. Not exactly a high sensitivity design, but as expected when using such a small mid-woofer for woofer duties and not only for mid-range duties.
The mid-woofers cone break-up at 10kHz is well suppressed and the 13.5kHz and higher break-ups is suppressed by 45-50dB and should be off the audible radar.
The cross-over frequency occurs at 2100Hz and the phase integration between the mid-woofer and the tweeter is superb.
The tweeter has a rising top-end at the 0deg, on-axis frequency response and it’s still visible from 15kHz at the 15deg off-axis frequency response. Listening tests will decide if the tweeter needs more attenuation.
Personally I like when the on-axis response has a somewhat rising top-end, since it often contributes to a greater 3-D perspective and soundstage. However, to avoid an overly bright sound characteristic this calls for a non toe-in loudspeaker setup.