The club desktop speakers provide surprisingly good bass given the relatively small four-inch midrange drivers and the small box. However, basic physics still imposes fundamental limits — the little speakers roll off below 200 Hz. If you like listening to any music that has lots of bass, the desktop speakers will be lacking in low-end punch.

To fill the missing low frequencies, we can add a sub-woofer to the system. The "sub" is made with a single driver, but it is much bigger. The bigger driver, matched with an appropriate amplifier, will generate of lots of low-frequency sound. A few aspects of building a subwoofer to use with the desktop speakers:

• Low frequencies only. Subwoofers are generally limited to frequencies below 200 Hz. The upper limit may be a little higher or little lower than 200 Hz, but there is usually a low-pass filter before the sub — probably before or in the amplifier driving the sub — to limit the frequencies.
• Typical sizes: 6, 8, or 10-inch diameter for a medium-sized subwoofer and 12, 15, or 18-inch diameter for a big one. Subs for a concert hall or a rave may be 24 inches across or more. Obviously, the enclosure for the sub will be correspondingly bigger.
• Lack of directionality. The low-frequency sound waves have very long wavelengths — ranging from about 20 m at 20 Hz to 2 m at 200 Hz. Humans have a hard time localizing these long-wavelength sounds, leading to a couple of consequences for using subwoofers. First, since stereo imaging requires the listener to be able to localize the sounds, there will be no stereo effect for subwoofer frequencies. This means that only one subwoofer is needed in a sound system. Secondly, the subwoofer can, in principle, be located anywhere within listening environment. And the driver can be oriented so that it "points" in any direction: to the front, the back, to one side, up, or down. It is a common trick to hide a subwoofer inside a piece of furniture, like a coffee table, with the driver pointed down in a "floor-firing" configuration.
• Closed or open. The enclosure can be entirely sealed up so the no air from the interior can move to the exterior — known as an acoustic suspension design — or the enclosure can have a port or vent to allow air to move back and forth from inside to outside — known as a bass reflex design. A sealed speaker generally gives a more accurate response with less distortion. The enclosure is usually smaller. However, sealed enclosures have difficulty reaching the lowest frequencies, so getting good bass in the 20 Hz - 50 Hz range might be difficult. Also, sealed speakers are less efficient, meaning that a bigger amplifier is required to get to a specific sound pressure level (SPL). Vented subwoofers used a tuned port to extend the frequency range to lower frequencies. Vented sub are also more efficient and so can get by with less powerful amps. (Or produce louder sounds for the same amp power.) However, vented subs exhibit more distortion than sealed subs. Also, a vented enclosure usually needs to be larger than that of a sealed sub. These are important tradeoffs to consider when choosing a design.
• More power. The bigger driver will require a higher-power amplifier. The small GTDT amp (< 5 W) or the 3122 Class D amp (10 W) will not cut it. For a medium-sized sub, we might want 50W to 100W. Even more for a big sub. Concert-hall sized subs might use 1000s of watts.
• Bigger power supplies. Higher power amps need higher-voltage power supplies. The components become (much) more costly and more care is needed to deal with the higher voltages.
• More wood. Obviously, bigger speakers require means more wood for the bigger cabinet. In addition, the wood panels must be more heavy duty to support the heavier driver and remain stiff with the high SPL sound waves are bouncing around inside. If using MDF, that probably means using 3/4-inch thick panels, instead of 1/2 inch. For really large subs, the 3/4-inch panels might doubled up. Internal bracing may become necessary for large subs.
• Bigger losses. If using classic analog Class-B output stages, the power loss becomes an issue. Class-B amplifiers are only about about 75% efficient at best, meaning that 25% of the power is being dissipated in the output-stage transistors. If an amp is delivering a full 100 W to a subwoofer, then 33 W is being dissipated in the transistors — they will get HOT. So heavier-duty transistors are needed along with heat sinking to dissipate the heat, both of which add to the cost and require more careful design. Alternatively, a more efficient Class-D amps can be used, but designing high-power Class-D amps can be tricky, too.
• Active or passive. An active sub integrates an amplifier in the enclosure together with the driver. A common active arrangement is to us a "plate amplifier", which is an amp in a flat-ish package that can be attached to the back of the sub enclosure. Having all the components in a single package makes for a convenient purchase and simplifies siting the sub in the listening environment. A passive sub separates the driver and the amp in separate packages. A passive sub is simply the driver and its enclosure, an arrangement that is probably easier for DIY construction.

Links of interest:

• Subwoofers on Wikipedia.
• Acoustic-suspension speakers on Wikipedia.
• Bass-reflex speakers on Wikipedia.
• A list of plate amplifiers at Parts-Express.