diyAudio logo

Subwoofer FAQ


[Paul Spencer]

Subwoofer FAQ

Questions included:

Can a DIY subwoofer compete with commercial subs?

In theory a commercial subwoofer has the potential to be better, since there is the benefit of expert design, a R&D team as well as tools and facilities beyond the financial means of the diy enthusiast. However, commercial subwoofers are generic solutions which have limitations imposed by the need to be able to market and ship the product and make it profitable. The diy enthusiast can spend all of the budget on materials, hence for a given budget a diy sub can have a better quality driver. Compromises can also be removed. Therefore a properly implemented diy subwoofer has the potential to offer considerably better value and performance. A clear advantage of a diy subwoofer is that it can be a custom design intended for a specific room. This means a result which is far superior, especially when one starts to implement large infinite baffle or architectural horn subwoofers, which for obvious reasons, aren't readily available in commercial subwoofers.

There are a number of other advantages:


Which subwoofer is best: sealed or vented?

There are many opinions on this issue, but the prevailing belief is that sealed subwoofers are best when musical accuracy is considered. If this does prove to be true in a blind AB test, then we must ask why this is so. I suspect the reason is not related to sealed subwoofers being inherently superior, but rather a matter of implementation. Often a sealed sub will integrate with room gain where a vented sub will have too much bass when combined with room gain. Both can be resolved, and in fact when driven within their limits both may in fact sound very similar when their response is the same.

Vented subwoofers will have 6db more output and lower distortion, but they must be larger to achieve this and often the extra output is achieved in the bottom octave, and this may not be worthwhile for "music only" subwoofers.

You should become familiar with the characteristics of each, and make an informed choice, balancing compromises to suit your goals.

Is there really such a thing as "fast bass" or is it just another hifi myth?

Fast bass is a subjective term which might be better replaced with the word "accurate." It's unfortunate that when someone uses the term, discussion often digresses and it becomes a bit of a red herring, leading the discussion into heated debate. Part of the problem is that it is a misleading term. Objectively speaking, all subwoofer drivers are slow. Consider a high excursion subwoofer driver moving a inch to reproduce a 20 Hz signal. Its actual speed will be 1 m/s or 3.6 km/h! If you think that is slow then consider a woofer reproducing 40 Hz at the same amplitude. It will be 0.5 m/s or 1.8 km/h. Now this is less than walking speed! It should be clear that subjective "speed" has nothing to do with actual speed. In fact, achieving subjective speed may not in fact follow along similar lines to your intuition, so ignore statements like "small drivers are faster" and "lighter cones are faster."

So how is "fast bass" or accurate bass achieved?

Here is the ideal situation. We start with an ideal driver. It can handle far more power than you will ever require without any power compression. The voice coil inductance is very low, and is flat over the bandwidth used. The BL motor strength is constant over all excursions that you will ever use and the suspension restoring force is also constant, without becoming stiffer with greater excursions. The suspension and voice coil systems do not exhibit any "memory effect." The driver excursion is perfectly linear over the range that you use it. All forms of distortion are no higher than your amplifier over the range in which you use it.

We put this ideal driver into an ideal box in an ideal room. The ideal room won't have any room modes. In such a room, the ideal box may be a horn which is loaded down to below your intended cut-off with a perfectly flat response. Or it might be an IB where the construction is such that structural/enclosure vibrations don't audibly colour the sound. Or if your room isn't ideal, then your box might be no box - an open baffle dipole where your driver is capable of delivering the desired low frequency cut-off with very low distortion.

Given all of these conditions, the group delay should be below the audible threshold. There should be no degradation of any time domain information, the frequency response should be perfectly flat and in-room transient response should precisely match the source without audible deviation.

If all of the above could be achieved, then it would be an interesting choice for some browsing this forum: "will I buy the beach house mansion or get the sound system?"

But since you can't achieve all this, you will have to pick your compromises. A good balance will be better than something that is excellent in one area and very poor in others.

How loud will my sub play in my room?

This is not possible to accurately predict with software alone. If you have a sub and measurement equipment, you can measure the transfer function of the room (SPL gain vs frequency plot). You can then simulate the max SPL of your sub and then add the effect of the room, and get a good indication. If you don't have the means to do this, then some guess work is needed. You may have to start with one subwoofer, and be ready to consider adding more later. Room modes can easily add 15 dB peaks and dips into the response and room gain can also boost the low end.

If your room is large, bass response may be less than predicted. Also if it is open to other rooms, then it is effectively a very large room. Room construction will also be a factor. Rooms in which the floor, walls and ceiling can vibrate noticeably will have less severe room modes, and less room gain than a room constructed of masonry and concrete.

To simulate max SPL independant of the room, you need to use a program such as WinISD pro. You can use it to determine max SPL, which will be limited by either the thermal power rating of the voice coils or maximum excursion (xmax), usually the latter. With some ultra high excursion drivers, you will reach thermal limits before all of the xmax is used up. If you want to push limits, a driver can handle more power than is usually quoted as the thermal rating, as long as you use it for short bursts and not continuously. The driver may be rated at 500w, but you might safely use it with an amp rated at 1000w if the continuous RMS level doesn't exceed the thermal rating. So to find peak SPL you might simuate in this case with 1000w. WinISD will show excursion and SPL for a given power input.

Keep in mind that with high power and excursion, you won't get as much as predicted. High power heats up the voice coil, and will usually give you power compression. You may lose 3 dB or more. Also if the suspension and BL curves are not flat at the range in which you are operating, you will have excursion related compression.

Work in progress - to be continued!

I'm thinking of building two subwoofers, is it worth going stereo or just dual mono?

Which driver is better - driver XYZ or driver ABC?


Do I need a subwoofer for music?