Slot Loaded, Dual Opposed 8" Subwoofer (suitable for on wall/high shelf mounting)

[Speakerofthehouse]

Why not copy this design?

Dynaudio 18s, dual opposed sub. Seems that this is way easier to make.

View attachment 455361

hi. Do we know what this design is? Asking as I am considering a diy sub

 

[kiwifi]

Construction of the cabinet is complete. The following is a photo history...

It all starts with the baffles


T-nuts are used to mount the drivers and the plenum


Easier to paint the baffle faces before assembling the slot


Sanity check!
The cone and magnet look really close together, so I also ran a 10-200Hz sweep through the assembled drivers to ensure that the cone would not contact the magnet with the drivers at Xmax.


Completing the "inner" box and adding the top


Adding an access port to the top (secured with T-nuts)


Internal wiring and sound absorption is added prior to sealing the ends.
I am using woolen felt for the lower chambers and hollow fill stuffing for the top chamber.


All sides of the cabinet are wrapped with a second layer of 12mm MDF, for a total thickness of 24mm throughout.


The removable plenum has been redesigned with holes added.
The back of the slot will be covered with a layer of felt to dampen any slot resonances. The plan is to measure the effectiveness of this, before and after adding the felt



The cabinet design won't win any beauty contests but rounding all the edges helps


Undercoat, prep and paint. Spayed with a black, quick drying, oil based enamel.


Installing the drivers and connectors. The feet are $2 rubber door stops.


Side by side with the M&K MPS-2810 (MX-700) that uses identical drivers to my design, which I am calling the PP-DOS-8


Measurements to follow.

 

[A Surfer]

Construction of the cabinet is complete. The following is a photo history...

It all starts with the baffles
View attachment 462559

T-nuts are used to mount the drivers and the plenum
View attachment 462560View attachment 462561

Easier to paint the baffle faces before assembling the slot
View attachment 462562

Sanity check!
The cone and magnet look really close together, so I also ran a 10-200Hz sweep through the assembled drivers to ensure that the cone would not contact the magnet with the drivers at Xmax.
View attachment 462563View attachment 462564View attachment 462565

Completing the "inner" box and adding the top
View attachment 462566View attachment 462567View attachment 462568

Adding an access port to the top (secured with T-nuts)
View attachment 462569View attachment 462570

Internal wiring and sound absorption is added prior to sealing the ends.
I am using woolen felt for the lower chambers and hollow fill stuffing for the top chamber.
View attachment 462571View attachment 462572

All sides of the cabinet are wrapped with a second layer of 12mm MDF, for a total thickness of 24mm throughout.
View attachment 462573View attachment 462574

The removable plenum has been redesigned with holes added.
The back of the slot will be covered with a layer of felt to dampen any slot resonances. The plan is to measure the effectiveness of this, before and after adding the felt

View attachment 462575View attachment 462576View attachment 462577

The cabinet design won't win any beauty contests but rounding all the edges helps
View attachment 462578View attachment 462579View attachment 462580View attachment 462581

Undercoat, prep and paint. Spayed with a black, quick drying, oil based enamel.
View attachment 462582View attachment 462583View attachment 462584View attachment 462585View attachment 462586

Installing the drivers and connectors. The feet are $2 rubber door stops.
View attachment 462587View attachment 462588

Side by side with the M&K MPS-2810 (MX-700) that uses identical drivers to my design, which I am calling the PP-DOS-8
View attachment 462558

Measurements to follow.

Fantastic work.

 

[AcousticTheory]

I am curious to see if measurements of the output of the slot will support the simulation results.

 

[kiwifi]

I am curious to see if measurements of the output of the slot will support the simulation results.

The measurements broadly follow those predicted by HornResp. The differences are because the complex profile of the plenum could not be accurately modeled.

Here is the impedance plot of one of the drivers


Here is the impedance plot of the system (both drivers and cabinet) which shows additional resonances at 254Hz and 333Hz


Measuring very near field to eliminate the room as much as possible


The response is measured first with no felt (NF) attached to the plenum and then with the felt (F) attached


Here is the frequency response showing the effect of the felt. It smooths out the high frequency peaks and dips but as expected, has little effect lower down


Here is the measured response vs what was predicted by Hornresp.
Instead of one peak at 250Hz, there are now two slightly smaller peaks at 254Hz and 333Hz, which I think that this is due to the revised plenum profile.


Slot resonance increases with volume. Here are the THD vs frequency plots before and after felting


Looking at the slot induced distortion in more detail, there are two resonant peaks. One at 254Hz and one at 333Hz.


In both cases the primary components of the distortion are 3rd and 5th harmonics. The even harmonics have been minimized by the push-pull driver configuration
For the largest peak at 333Hz, the 3rd harmonic at 8.4% would result from a fundamental frequency of 111Hz and the 5th harmonic at 2.8% would result from a fundamental at 66.6Hz.
For the second peak at 254Hz, the third harmonic at 2% would result from a fundamental frequency of 84.7Hz and the 5th harmonic at 1.5% would result from a fundamental at 50.8Hz.
If the sub cross over is at 80Hz then the fundamental frequencies responsible for the 3rd harmonics will be attenuated by that filter, together with the fundamentals at 254Hz and 333Hz.

In conclusion, the slot appears to provide good coupling of the push-pull drivers resulting in a very low even order harmonic distortion. Because of the slot dimensions in this particular case, the odd order harmonic distortion is also low, because it mostly originates outside of the operating frequency range of the subwoofer.

Here is the frequency response over the sub-woofer's operating range (no-DSP applied yet)

 

[kiwifi]

Here is the response with DSP enabled. This flattens the response to 40Hz and is 3dB down at 30Hz, which is plenty low enough for an 8 inch sub.


I have found that the easiest way to find the correct values for the low frequency boost is to use the EQ window in REW


Using the unfiltered very near field response as the input, I manually add a Low Shelf with Q (LSQ) with a gain of 12dB. Then adjust F and Q until the desired response is achieved.

In my case the filters are implemented using QSC DSP-4 which allows multiple LSQ filters to be cascaded but for some reason, the first one has a maximum Q of 0.747. This is why the first LSQ is at 160Hz with a Q of 0.707. At 80Hz and below it just becomes a small gain stage, which I use in conjunction with LSQ #2 to achieve the response that I want.

Over excursion in the driver is prevented by using a limiter function just before the final output in the QSC DSP-4

 

[kiwifi]

I wanted to be able to locate a subwoofer in each of the four quadrants of my room. This picture of the back left corner illustrates why I designed this sub.


The PP-DOS-8 is located in the top right of the picture. It is sitting on a small shelf above the windows which run the length of the room. As you can see there was no option to put a sub on the floor because of the bench seating under the windows and the entrance to the dining room. The M&K MPS-2810 (MX-700) is located tangentially opposite the PP-DOS-8, on a small table in the back right corner. Each of the two front quadrants have a large, dual 12", floor standing sub, so the main purpose of the pair of dual eights in the rear, is to even out seat to seat variations. That is also why all four subs are located 1/4 of the room length away from the front/back corners. The rear eights are closer to the MLP than the front twelves, so they don't need to be as large and powerful as the front pair of subs, to be effective.

The finished PP-DOS-8 weighs in at 46lbs (21Kg), so the shelf needs to be sturdy. That is 5lb more than the M&K despite having no embedded amplifier. A knock test reveals that cabinet reverberates less than the M&K as well. With a hand on the M&K, you can feel a tiny amount of vibration in the cabinet when it is playing, but nothing at all from the PP-DOS-8 due to the dual opposed drivers. This means that the PP-DOS-8 could probably be attached directly to the wall studs if you wanted.

Of all the subs that I have built, this is the one that I would build again. It just works, really well.

 

[scorlee]

I have also tried this slotted installation method, the effect is a little better than expected, the sound artifacts are inevitable, but very slight, so don't worry too much

 

[kiwifi]

I have also tried this slotted installation method, the effect is a little better than expected, the sound artifacts are inevitable, but very slight, so don't worry too much

I can see that for larger drivers and slot dimensions, artifacts might be inevitable. In the case of my design, using 8 inch drivers, any artifacts are inaudible and this is born out by the measurements. I would go further and say that the PP-DOS-8 out performs the equivalent M&K. The problem is that the cabinet is more complex to build than a simple box, so unlikely to be mass produced any time soon!