What's new

Jeff Secker Realistic SCT-90 picture
Joseph Fabricatore Acoustat topic reply
Marek Mechliński Copland CTA-405 refurbishing
Richard Hines Robertson Audio topic reply
Brett Robertson Audio topic reply


I started my third subwoofer project at summer -99. Element is still the same Focal 13V that was used in earlier version. Box type was selected to be either closed or passive radiator. Some extra effort was put to box construction. Box is damped with bitumen and bend veneer structure makes it stiff compared straight sides. Wavelengths of subbass are typically more than 2 metres. 24dB/oct active filter should take care of standing waves but structure is made so that they are minimized just to be sure. Box was originally designed for passive radiator and there is still a place for one. Suitable element was not found and I heard some discouraging comments about passive radiator so I left it out. Passive radiator might be best in very small enclosures.

DIY Dynamic loudspeakers


Passive Radiator vs. Reflexed vs. Closed

Passive radiator system was originally designed to produce very low frequencies. It had many advantages compared to reflexed system. Reflex port becomes impractically long when tuning frequency is low. At the same time air volume passing through the tube increases. To avoid compression and wind noise, port diameter has to be huge, which in turn increase the length. Pipe resonances may cause colouration. Passive radiator works very similarly than reflex tube but solves some of those problems mentioned above. Structure of a passive radiator is more or less an element without magnet structure and that is tuned to right frequency with by adding weight to the cone. Transient response is poorer than with reflexed enclosures and resonance problems are usual. There is a lot of energy stored in heavy cone. Moreover passive radiator is more complicated and expensive to build.

At the time when sub was build it was cool to extend bass as low as possible. In newer systems I rather minimize distortion and group delay instead. At tuning frequency cone moves less in resonance-loaded boxes than in closed box that is producing same sound pressure. Phase properties are superior in closed box. Benefits of passive radiator are not significant if tuning frequency is increased. In this case box is rather large and element has good power handling so it is possible to equalize closed box low enough and maintain rather low group delay.

Element and Box Tuning

Element model is 12" Focal 13 V 7601 DB. Element has many parameters that makes it good for very low tuning. Closed box design requires rather high Q-value. Cone weight is good compromise between sensitivity and resonance frequency that is 26 Hz. Force factor is relatively high 11.1 so heavy cone should be well controlled. Long excursion ± 12 mm is essential in closed boxes as well as high power rating and low power compression. Cone is extremely stiff as well as chassis. Mechanical structure seems reliable. Element has two 4 ohm voice coils that are in this case driven with separate amplifiers.

Box is rather large (100 litres) thus its rolloff is shallow. Total Q-factor is 0.59. Q-factor of 0.5 would produce best transients, 0.7 has best frequency response. Total Q-value can be easily adjusted with active filtering as well as frequency response.

Box Structure

DIY Dynamic loudspeakers

As always, I liked to do box as well as possible. Box is shelf braced with three internal shelves and four blocks between each of them. shelves contain big unsymmetric holes for free air flowing. Successive shelves are made different to lift possible standing waves as high as possible. Lowest possible (without shelves) standing wave frequency is around 250 Hz and filter used is 4 th order. When cut off is at 70 Hz first standing wave frequency is about 40 dB attenuated. With second order filter that would be only 20 dB. So it might be worth a while to pay attention to standing wave problem even if operational frequency range is lower than lowest standing wave frequency. Sides of the box are curved. Front, top and bottom are straight. Bottom is at different angle.

Box resonances may cause problems especially in large boxes. In my case, box surface is nearly 2 m², while area of 12-inch cone is only about 0.053 m². Box area is 28 times larger that one element. For moving same amount of air box needs to move only 0.035 mm while element move 1 mm. Better air coupling increase this problem especially at lowest frequencies. Total sound pressure depends how different resonances are summed. Measurements show that sound output can be unbelievable high from box sides.

DIY Dynamic loudspeakers

I had earlier decided that I use bitumen in damping. It is very inert and heavy material, where friction converts vibration energy to heat. An outer layer simply keeps bitumen in place. Layer thicknesses are more crucial than it seems at first. If supporting layers are too thick and rigid compared to bitumen, system doesn't work efficiently, because vibrations aren't transferred to bitumen efficiently. On the other hand, if supporting layer is too thin and flexible, bass may lose its accuracy, even if resonances were negligible. So, some kind of compromise had to be made.

I wanted the box to be rigid. That is why I ended up to curved structure. Side layers are bent of single 6 mm birch-plywood pieces so there is also reduced amount of undesirable joints. To make bitumen comparable to supporting layers I made it 15 mm thick. Melted bitumen was poured in from bottom before outer bottom plate was set in place. So whole system has two almost separate boxes with 15 mm bitumen in between.

Result of damping and listening tests

Box resonances are almost negligible at bass frequencies. It can be easily demonstrated by keeping hand on the box when played loud. Normally resonances can be easily felt, but in this case you can hardly feel anything during bass transients. Higher frequency resonances are not significantly damped they can be easily felt by removing filter. Structure has proved to be efficient at bass frequencies but rather difficult, expensive, time consuming and messy compared to normal square MDF-enclosures. It also weights about 70 kg so it is not the easiest piece of furniture to be moved.

With two similar subwoofers aligned to suitable places in room it is possible to straighten room response. With suitable measurement equipment it is very easy to find places, which creates best response. Effects in response are quite remarkable but unfortunately this doesn't make reverberation time shorter so boomy room sounds still boomy.

Amplifiers and filters

Both subs have two 80 W IC-amplifiers (LM3886); one for each voicecoil and so 160 watts in total. Amplifiers have balanced and normal line inputs, voice activation and regulated power. Electronics is attached below the subwoofer box. Mirror Xover is used in filtering. Besides subfilters it has highpass for main speakers.

DIY Dynamic loudspeakers