This is what I will build up now. Components I found in my stock marked in green, the missing just ordered...
Thank you Kwesi for this amazing design, and for your willingness to share this.
I was intrigued by the Mesanovic CDM65, which has a cardioid profile and also uses the SB15nbac woofer.
So I tried to replicate that speaker starting from your design, adding a woofer on each side to try and get the cardioid pattern.
It's just an idea for now, but I will try to build it in the coming weeks.
The new design requires putting the SB15 woofer inside a 2.7 L enclosure, and deepening the box till 420 mm.
The two subwoofers are on each side at the bottom, centered.
I picked the Visaton w 170s 8 ohm for the subwoofers in a sealed construction.
In vituixcad, I performed the following:
I set the subwoofer on the sidewall of the box in the diffraction tool to get the horizontal responses.
In the model, each gets rotated 90°/-90°, 90mm/-90mm on the x axis (= half the width of the baffle) and z=210mm (= halfway the depth of the box).
Not sure if that assumption is right, maybe the acoustic center should be taken instead for the x axis?...
Anyway, this is the vituixcad result:
I couldn't properly get rid of the bass bump, but I think it will sound fine - flat might bass might be boring to some.
The directivity plot looks weird when the tweeter is in there, probably because it only goes until 60° (taken from your measurements).
If I remove the tweeter, the graph better shows what I hoped to get:
The new crossover (simplified somewhat, to keep the total parts count down):
Kind regards,
I was intrigued by the Mesanovic CDM65, which has a cardioid profile and also uses the SB15nbac woofer.
So I tried to replicate that speaker starting from your design, adding a woofer on each side to try and get the cardioid pattern.
It's just an idea for now, but I will try to build it in the coming weeks.
The new design requires putting the SB15 woofer inside a 2.7 L enclosure, and deepening the box till 420 mm.
The two subwoofers are on each side at the bottom, centered.
I picked the Visaton w 170s 8 ohm for the subwoofers in a sealed construction.
In vituixcad, I performed the following:
I set the subwoofer on the sidewall of the box in the diffraction tool to get the horizontal responses.
In the model, each gets rotated 90°/-90°, 90mm/-90mm on the x axis (= half the width of the baffle) and z=210mm (= halfway the depth of the box).
Not sure if that assumption is right, maybe the acoustic center should be taken instead for the x axis?...
Anyway, this is the vituixcad result:
I couldn't properly get rid of the bass bump, but I think it will sound fine - flat might bass might be boring to some.
The directivity plot looks weird when the tweeter is in there, probably because it only goes until 60° (taken from your measurements).
If I remove the tweeter, the graph better shows what I hoped to get:
The new crossover (simplified somewhat, to keep the total parts count down):
Kind regards,
Hi, thanks for your feedback!
The CDM65 is a very nice design and worth as reference. Another design to get inspiration from might be the ASR Directiva r2. Uses also SB15NBAC and Seas DXT to achieve cardioid pattern, still unfinished but the geometrical design is ready: https://www.audiosciencereview.com/...-project-market-requirements-gathering.27212/
How do you plan to realize cardioid pattern with a passive xover? Just by shaping the phase transition between woofers and mid?
About the bump in the woofer response around 80Hz: It is there, and it will sound as boomy as it looks in the simulation;
It appears due to interaction of the woofer impedance bumps around the resonance frequency / cabinet tuning and the typically sized xover components of the woofer lowpass. Usually you implement series RCL notch filters in parallel to the woofer to linearize the impedance that the lowpass can work properly, this will avoid the hump and give a flat response. This measure is also often used for the midrange channel, RCL on the resonance and the following passive highpass behaves more textbook like.
Best regards
Peter
P.S.: In your simulation, there might be also some work open on the phase transition between mid and tweeter - blue and green phase should overlay with same gradient at least +/- 1 octave around the xover frequency, wider is better.
I may also propose an additional capacitor in parallel to the 150uH coil of the mid lowpass, tuned to suckout the cone resonance at 10kHz effectively.
The CDM65 is a very nice design and worth as reference. Another design to get inspiration from might be the ASR Directiva r2. Uses also SB15NBAC and Seas DXT to achieve cardioid pattern, still unfinished but the geometrical design is ready: https://www.audiosciencereview.com/...-project-market-requirements-gathering.27212/
How do you plan to realize cardioid pattern with a passive xover? Just by shaping the phase transition between woofers and mid?
About the bump in the woofer response around 80Hz: It is there, and it will sound as boomy as it looks in the simulation;
It appears due to interaction of the woofer impedance bumps around the resonance frequency / cabinet tuning and the typically sized xover components of the woofer lowpass. Usually you implement series RCL notch filters in parallel to the woofer to linearize the impedance that the lowpass can work properly, this will avoid the hump and give a flat response. This measure is also often used for the midrange channel, RCL on the resonance and the following passive highpass behaves more textbook like.
Best regards
Peter
P.S.: In your simulation, there might be also some work open on the phase transition between mid and tweeter - blue and green phase should overlay with same gradient at least +/- 1 octave around the xover frequency, wider is better.
I may also propose an additional capacitor in parallel to the 150uH coil of the mid lowpass, tuned to suckout the cone resonance at 10kHz effectively.
Last edited:
Thanks a lot for the input!
Indeed, many parts of the crossover should be improved. Still learning!
I know I can't go for a real cardioid, but more like the somewhat 2.5 setup as described in the last point of this post:
https://www.diyaudio.com/community/threads/cardioids-for-dummies.412157/#post-7671899
I noticed that if you cross the subs too high, the directivity gets very wide around the mid frequencies. Would probably sound all bad.
This is what I now came up with.
Cost of all the crossover parts got a bit out of hand unfortunately...
But still worth a try I I think.
Indeed, many parts of the crossover should be improved. Still learning!
I know I can't go for a real cardioid, but more like the somewhat 2.5 setup as described in the last point of this post:
https://www.diyaudio.com/community/threads/cardioids-for-dummies.412157/#post-7671899
I noticed that if you cross the subs too high, the directivity gets very wide around the mid frequencies. Would probably sound all bad.
This is what I now came up with.
Cost of all the crossover parts got a bit out of hand unfortunately...
But still worth a try I I think.