Designing in-wall speakers is a systematic balance of compromises-even more so than
conventional speakers. High "Q" drivers are often selected for in-wall use because each
speaker is effectively placed into a three cubic foot cabinet once installed. I didn't want to
go that route and sacrifice sound quality; high "Q" drivers are said to bounce around the
audio signal rather than directly play it.
I was also shooting for a decent power response since most of the listening will be done
off-axis once these speakers are installed. This led to experimentation with the midrange
tuning. I settled for a sound that was a touch below hot on-axis for a smoother, more
neutral off-axis response.
Initially, I tried crossing the ND20FB-4 with a 2nd-order electrical high-pass filter (one
capacitor in series, one inductor in parallel) that approximated a 4th-order Linkwitz-Riley
acoustic slope with a corner frequency of 3,000 Hz. Listening to the tweeter alone, at
about 2.83 V, revealed too much "hash" for my taste and I started to worry about overexcursion problems. A third-order electrical high-pass filter (two capacitors in series with one inductor in parallel connected between the capacitors) cleaned up nearly all of the lower-frequency junk. The acoustic slope's target was that of a 6th-order Linkwitz-Riley alignment but the actual slope is asymmetric and steeper. Polarity of the tweeter was reversed in relation to the woofers.
The DC130BS woofers do a lot of work in this design and are called upon to run out
to 2,500 Hz before rolling off. A second-order electrical low-pass filter (one inductor in
series with a capacitor in parallel) helps to create a fifth-order acoustic slope with a 3,000 Hz corner frequency. In order to make the woofers play nice with the tweeter, a series RLC filter (a resistor, inductor, and capacitor wired in series and placed in parallel with the woofers) was implemented to contour their frequency response.
A slightly exaggerated bass emphasis has been built-in for two-channel stereo listening.
The electronic high-pass filter built into most home theater receivers will minimize this-
most of those are set to about 100 Hz. I would recommend doing this as it also limits the
woofers' excursion, something absolutely needed during movie playback.
The crossover components were hot glued to the back of the speaker baffle for
installation simplification. Each capacitor used is an electrolytic. I'll leave the capacitor
sound quality debates for someone else to discuss!
All off-axis measurements shown were taken with the microphone referenced to the
"inside" or tweeter-offset side of the baffle. The measured speaker impedance never went below 6 ohms. No receiver on the market will be overly stressed driving these-even super low cost home-theater-in-a-box receivers.