[BrunoB]

I want deep bass but I am living in an apartment with neighbors above me. What to do?

The idea is to focus the bass energy on the listener.

I built a W frame dipole using two 15'' drivers (views with the top removed):


.

This is the first speaker I built from scratch. This sub was very easy to build, this is a one afternoon project.

The sub is installed very close to the listening position, just behind the listening sofa , at ear height:

The red part is a blanket that covers the sofa. I temporally placed two pieces of rockwool between the sub and the back of the sofa in order to decrease reflections from the hard surfaces.

The slot is facing the listener.  The slot acts as an acoustical lens that concentrates bass energy on one small spot. Outside this spot, the bass level decreases very quickly, which gives a neighbor friendly dipole sub.

Another advantage of this set up is that there is zero room interaction for the lowest frequencies. Pure bass!

This is the frequency response measured at the listening position:


Boosting of low frequency and low pass filtering are applied.

My next post will provide more measurements and technical details.

Bruno

[mcgsxr]

Very interesting, I had not thought of that type of install, for that type of implementation.

I look forward both to the objective and subjective details!

What are the drivers?

[iON]

Nice curve! I haven't seen that before, no walls on the side?

/Jon

[BrunoB]

Drivers are a pair a PA Audio WN-15R:
http://www.intertechnik.de/index.html/JTI2bmF2aWQlM0QxNzY4JTI2bGFuZyUzRGRlJTI2c2lkJTNEbjQ4MWRmMjAxOTdjYzMlMjZiJTNE.html?detail=45955

I choose this driver because it has a relatively low moving mass (66 g) and  a good BL/Mms ratio (about 0.24). This ratio is my main criterion for choosing a driver for this project. It is like a car power to weight ratio, which determines the acceleration capabilities of a car. The BL represents the force that moves the moving parts whereas Mms is the moving mass.  With a ratio of 0.24, the PA Audio WN-15R (15'' woofer) has a BL/Mms ratio twice as higher than most 12'' woofers. The magnet looks small because it is a neodymium magnet.

The amplifier is a Panasonic SAXR58 digital amplifier that provides 2 x 100 W. These digital amplifiers have very good control on the woofer. I choose the speaker cables as thick and as short as possible (R ~= 0.01 Ohms). The Panasonic works nicely for this application because it is very quiet (no hum or hiss, which is important for a speaker that is placed very close to the listening position).

The bass module is a Reckhorn B1. This module contains a subsonic filter that, unfortunately, cannot be bypassed. The subsonic filter is set to 20 Hz (minimum). Boost level is set to +12 db and boost frequency is somewhere above 20 Hz (knob set at 9). The low pass filter of the B1 is very bad (< 6db /octave). I am not using it. Presently I am using the sub output of a Sharp digital receiver, which gives a 80 Hz low pass filter of 24 db per octave. A good low pass filter is very important because of the 200 Hz resonance of the sub (more on this in one of my next posts).


Thanks to the W frame, the two drivers move in opposite directions, consequently the sub frame does not vibrate at all. This type of dipole is well described (Linkwitz and/or Ripole). The difference between my sub and Linkwitz/ripole is that I don't use side walls.

Bruno

[TRADERXFAN]

very clever Bruno. Great thinking.

How well does it integrate with the mains though? Do you have phase control?

[BrunoB]

very clever Bruno. Great thinking.

How well does it integrate with the mains though? Do you have phase control?

Thanks for the nice comments.

I don't have phase control but I can set up a delay for the subwoofer. I choose the sub delay by measuring the low frequency response of the subwoofer and main speakers playing together. I guess that the delay was correct when I had the least cancellation between main speakers and the sub. When the setup is right, the deep bass seems to come from the main speakers only. It is always tempting to push the sub level higher. That sounds very nice at first, but after a while, I found that an excess of very tight deep bass becomes tiring, especially with music. Furthermore, when the sub level is set too high, the integration with the main speakers is not as good - bass seems to come from my back. I am still playing with the level, boost frequency, delay to get the bass I like the most (which is not necessarily the most accurate).


Bruno

[BrunoB]

Measurements
The mic is an uncalibrated ECM-8000. THe audio sound card is a M-Audio Duo. Software is SoundEasy v 10.0.
Impedance of the sub only  (no filter/boosting/eq):

FS is 26-27 Hz. FS of driver alone:  about 34 Hz. Strong resonance around 200 Hz.


FR from 20 to 20K at the listening position:

The black line shows the FR at the listening position when the slot of the W frame is directed towards the microphone, as shown in the pictures above. The blue line shows the FR when the sub is rotate 180 deg, the larger openings are facing the microphone, the slot is facing the back wall. No LF boost , no EQ and no filtering performed.
One can see that the 200 Hz resonance is relatively much stronger  when the sub is rotated 180 deg.





To decrease the 200 Hz resonance, I added pieces of memory  foam inside the sides:



Waterfall plot with memory foam measured at the listening position:

The foam makes a visible difference in the water fall plot around 200 Hz.

FR from 10 to 500 plus phase (green line) at the listening position (memory foam present):

No filter boosting or eq.



Bruno

[Graham Maynard]

Hi Bruno,

This is a nicely presented project - far better than anything I could ever do.

Were some measurements taken with the sides removed, or were the photos taken at a different time ?

Have you run it without the sides, and if yes how does it compare/sound ?

I notice the 50Hz blip in the impedance plot which is not noticeable on other responses.  It might equate to a room dimension, but is very sharp and there are not any others similar.  Any known reason for it ?

Cheers ............ Graham.

[BrunoB]

Hi Bruno,

This is a nicely presented project - far better than anything I could ever do.

Were some measurements taken with the sides removed, or were the photos taken at a different time ?


Have you run it without the sides, and if yes how does it compare/sound ?

Hi Graham,

You have probably noticed slight differences between the FR's. I used two different amplifiers and probably slightly different mic positions . I also used foam for the second set of measurements.  I liked the sound better without the sides. I will redo FR measurements with and without the sides.

I notice the 50Hz blip in the impedance plot which is not noticeable on other responses.  It might equate to a room dimension, but is very sharp and there are not any others similar.  Any known reason for it ?

Cheers ............ Graham.

The blib is due to the 50Hz AC from the power supply of my external sound card.


Bruno

[Graham Maynard]

Hi Bruno,

So the drivers facing each other still reduces the resonant frequency, but it sounds better without the rearward facing resonant chambers.

Interesting.

Cheers ....... Graham.

[tubamark]

Bruno,

Good stuff - 'appreciate the measurements!  I've pondered a similar notion, but haven't tried anything serious yet, as I'm moving my listening room soon.
A couple questions, If you can share:
1)  I understand the neighbor/SPL objective drove the placement to be right by yer head.
Have you tried with the frame on the floor just for kicks? like a foot or so behind the seat.  Basically I'm wondering about sofa coverage rather than single-seat as you've described.   Also because SAF goes up when the box goes down (outta sight)  . . .
2)  How far is the frame and/or your listening position from the rear wall?  'Apologies if this was mentioned and I missed it.  Maybe not a big factor at the super-nearfield listening distance you are using, but I'm still curious.


This could prove to be an awesome alternative to IB bass!  Please keep us posted on any new discoveries or variations you try!

Envious and curious, grateful for your sharing,
-- Mark

[BrunoB]

For those who can read German,
I 'd like to recommend this issue of Hobby-Hifi:

http://www.hobby-hifi.de/Archiv/05/02_05/02_05.html


There is one article about bass dipole with different frames (H,U,I,A, ripole). The effect of room positioning, freq of resonance, impedance plots plus also comparison with bass reflex , ... are presented with measurements.



Here is the issue that inspires me for doing this project:
http://www.hobby-hifi.de/Archiv/07/04_07/04_07.html

(look at the bottom left picture)

They compared four different drivers (12") for the dipole sub. All gave very similar FR's but sounded slightly different.

Bruno

[Graham Maynard]

Hi Bruno,

I know you filter out the Ripole peak around 200Hz, and this must be related to the cabinet rather than a driver dimension.

Does taking the sides away help in that regard ?

It also occurs to me that if a Ripole was shrunk in size - as for say 8" drivers - then that 200Hz resonance might be shifted away from bass frequencies - to around 400Hz - and thus further away from the wanted LF working range.  Of course more 8" drivers would then become necessary to match the 2x 15" cone area, though with a potential to be more efficient too.

Cheers ....... Graham.

[BrunoB]

Here is an impedance plot measured on a single woofer:


I first thought  that my dipole sub was transformed into a bass-reflex because of the two peaks .

It seems that the two drivers really see each other although they are not placed on a air tight enclosure.

Bruno

[BrunoB]

Bruno,

Good stuff - 'appreciate the measurements!  I've pondered a similar notion, but haven't tried anything serious yet, as I'm moving my listening room soon.
A couple questions, If you can share:
1)  I understand the neighbor/SPL objective drove the placement to be right by yer head.
Have you tried with the frame on the floor just for kicks? like a foot or so behind the seat.  Basically I'm wondering about sofa coverage rather than single-seat as you've described.   Also because SAF goes up when the box goes down (outta sight)  . . .

I havent' tried with the sub on the floor. On way to increase sofa coverage is to rotate the sub 180 deg, then the slot is facing the back wall. In this case, the amplifier volume must be increased and the low pass filter must be steeper because of the relatively stronger 200 Hz resonance.

2)  How far is the frame and/or your listening position from the rear wall?  'Apologies if this was mentioned and I missed it.  Maybe not a big factor at the super-nearfield listening distance you are using, but I'm still curious.

the listening position is 1.5 m far from the back wall and the on the sub is 0.5 m deep. I have a small listening room. I would not placed the sub on the floor between the sofa and the back wall because it would increase cancellation of the dipole. A better placement would be the middle of the room, far away from walls and corners. I don't know when I will try another position for the sub.


Bruno

[Michael V]

Bruno-

Would you mind sharing the dimensions?  I have a pair of Goldwood 15 inchers lying around that I could try in your design.  Definitely an afternoon project!

Any thoughts on the suitability of these?

http://www.partsexpress.com/pe/showdetl.cfm?&Partnumber=290-384

Regards,
Mike

[mcgsxr]

I have considered the Goldwoods you mention, as their QTS, FS are attractive - what remains to be seen is how they handle excursion, and if they make nasty noises as they bottom etc - or whisle or make other noises.

Lots of swept area, that is for sure!

[BrunoB]

Bruno-

Would you mind sharing the dimensions? 

Unit is millimeter. I used a MDF  thickness of 18 mm.

Bruno

[BrunoB]

Hi Bruno,

I know you filter out the Ripole peak around 200Hz, and this must be related to the cabinet rather than a driver dimension.

Does taking the sides away help in that regard ?

It also occurs to me that if a Ripole was shrunk in size - as for say 8" drivers - then that 200Hz resonance might be shifted away from bass frequencies - to around 400Hz - and thus further away from the wanted LF working range.  Of course more 8" drivers would then become necessary to match the 2x 15" cone area, though with a potential to be more efficient too.

Cheers ....... Graham.

Hi Graham,

I remeasured the impedance plots with and without sides. The two drivers are connected in parallel.

First, without side wall.


Second with the side walls. Note that my side walls cover only 50% of the surface of the sides. Unfortunately, I don't have full size walls to try.


We can see now two resonance peaks, one at about 180 and another stronger at about 240 Hz. Also, the main FS decreases slightly in intensity (41.5 Ohms) and frequency (24 Hz)


Finally, without the sides walls but with plenty of memory foam attached to the walls (I posted a pict earlier in this thread)

.
The intensity of the 200 Hz resonance peak is lower. The FS decreases in intensity (40.5 Ohms) and frequency (24 Hz) .

And now yet another strange plot. The setup is the same as the previous plot: memory foam and no sides. But this time I connected the two drivers out of phase!


There is no more 200 Hz resonance and I don't understand why.

The FS is about 34 Hz and 45 Ohms. This  looks like a free air measurement.
There is a new resonance at 850 Hz.

Bruno

[Davey]

"There is no more 200 Hz resonance and I don't understand why."

Because the cavity is not pressurized now.  More of a free-air result measurement is what you'd expect, and that's what you have.

Dave.