Active filter

Successful use of a passive filter in a subwoofer is highly unrealistic. The seemingly simple task of just creating a crossover function at 100-150 Hz with a 4th order  slope might be within reach, but you can just forget about a lower crossover frequency or equalizing with a passive approach. The reasons for this is the difficult impedance of the driver in the cross-over range, and the extremely large component values needed.  A successful result requires extensive impedance correction, which will lower the system impedance dramatically. Passive EQ will also greatly reduces the efficiency. In the end, the system becomes virtually impossible to drive. Active filters is the only solution!



The filter is EQ'ing the system as I have described in a previous section, and crossing at 100 Hz with 4th order Linkwitz-Riley slope (In my opinion the only “correct” filter type). The filter block schematics is shown below. The filters are best made up with op-amps, I used the LM837 (quadruple). LM837 is designed for audio use, has very good data and is fairly inexpensive.

The graphic above describes the block schematic of the filter and equalizer. The first two sections are the equalizer and the low-cut filter. The two last 2nd order filters are in combination a 4th order Linkwitz-Riley filter. The equalizer's roll-off above 100 Hz creates a steeper total roll-off than a 4th order function, but this does not cause any phase problems in the crossover region, as the 180 degrees phase shift is located exactly at 100 Hz.

This graph shows a simulation of the equalized frequency response including the crossover (solid line) compared to the 4th order Linkwith-Riley function (dotted line).

The main system used with the subwoofer must be crossed with the same filter slope as the sub. The main system should also have approximately flat response at least one octave below the crossover frequency in order to get a flat summed response using a by-the-book 4th order filter for the main system. Using the sub with "satellites" (small speakers) with its cut-off above 50 Hz will require a different filter. The best way to design this is by simulation of the whole system in a proper crossover simulation software.

My own solution is using a 2nd order filter at 100 Hz with quite high Q for the main systems. Their acoustic response falls off 12 dB/octave below 100 Hz (closed box) with low Q at resonance. The sum is a 4th order roll-off with -6 dB at 100 Hz, and perfect phase correlation at 100 Hz (180 degrees shift for both the sub and main system).

And here’s a more detailed schematic of the filter and equalizer:

The OP's I used was the quadruple LM837, which makes one piece of silicon enough for a complete filter channel (two needed for stereo). A better choice today would be OPA2134 (dual) or similar high-end op-amp.

The OP’s may need offset adjustment to avoid DC potential at the output. This is somewhat dependent on the OP type you choose. The LM837 will need it, but the OPA2134 will not. The offset adjustment is best carried out by feeding the negative input of the last OP the required DC voltage for zero output offset through a high value resistor. Alternatively, you can put in a large high quality capacitor at the output. The value of this must be large enough to avoid low frequency cut-off, 5 Hz is a good target to aim for. If your power amp has an input impedance of 47 kOhm this cap should be at least 0.68 µF.

Resistors should be 1% metal film types, and the caps should be some good polyprolyplene types. It's difficult to find 1% or 2% caps, and they are usually quite expensive. Instead I buy a large amount (hundreds) of caps, and measure them to pick out the ones that fall within 2% tolerance.

A power supply is of course necessary. This must be a regulated type, and its voltage must not exceed +/- 18 V. If it does, your OP-AMPs will release its storage of smoke*. The simplest solution is to put in a pair of 78xx / 79xx regulators, and feed them from the main power-supply of your power amplifier. What I have done is to use two 12V lead-acid batteries, and trickle charge them from a dedicated charger. This is possibly the best sounding power supply solution.

My active filter concept used here is quite flexible, and will work with just about any sub-woofer. The component values must of course be altered when different cut-off  frequencies and filter Q's are needed.

* All electronics depends on smoke to operate. Its obvious - when the smoke is released, it dont work anymore!!