Sub amps

September 30, 2014
 by Paul McGowan

We’ve covered much about the crossover workings of a subwoofer in these posts. We will cover more, such as phase, when we get to the setup portion of our efforts. For now let’s move on to the second category of a subwoofer, the power amplifier. For those just catching up with this series, you can go here to start reading.

Unlike the majority of loudspeakers on the market, which are passive devices requiring the user to supply the power to drive them, modern subwoofers are active with built in power amplifiers.

These internal subwoofer power amplifiers vary in wattage, topology and design. Most active subs today use what we refer to in the industry as a ‘plate amp’. This means the amplifier is mounted to a plate of metal and comes to the manufacturer as a pre-built unit. Look at the rear of most subwoofers and you’ll see this plate I am speaking of.

In the early days, these plates included a rather large heat sink to dissipate the heat generated by the more traditional Class AB power amplifier topology. Today, most subs require smaller heat sinks, while generating far greater power, because they are typically designed with high efficiency Class D amplifiers or a hybrid of a class D amplifier, called a tracking amp, whose brand name is BASH. Whatever their topology, power amplifiers in modern subwoofers can vary from 50 watts up to many thousands of watts.


Here is a picture of the rear panel plate amplifier of my REL T9 subwoofer in Music Room Two. This is a great subwoofer using a more traditional class A/B amplifier design producing 300 watts. Many other designs, as mentioned earlier, can produce far greater watts with even smaller heat sinks.

One thing about subs you may have already figured out. If you include a subwoofer in your system you are, by default, running a bi-amplified stereo system.

The choice of how many watts are needed in a design depends greatly on the design choices made by the design team. Parameters include the size of the enclosure, ported or sealed, servo or EQd and overall performance expectations.

I’ve been involved in a number of subwoofer designs: those of Genesis technologies and Martin Logan’s Descent series come to mind. In both these design challenges the subwoofers were servo controlled and required enormous amounts of power; a topic we’ve yet to cover but will shortly.

Enclosure size plays, perhaps, the second biggest card when it comes to required amplifier power. The general rule of thumb is the smaller the cabinet, the greater power you need to make it properly woof. A good example of this was Bob Carver’s impressive work on his original subwoofer called the True Sub or the Cube. It featured a tiny eleven inch cubed enclosure with dual woofers, and was powered by a 2,700 watt amplifier. Talk about tiny! This less-than-a-cubic-foot of subwoofer rumbled down to 18Hz and, if you could keep it from walking across the room when it got going, produced prodigious bass.

The point is, you really cannot make an adequate judgment on the quality of a subwoofer based on wattage alone. Properly designed subs should have the power needed to produce their rated results. What’s the lowest it goes, what’s its crossover features, size, etc.? They all matter.

Tomorrow, let’s look at connection. How do we connect a subwoofer to our systems.

Subscribe to Paul's Posts

9 comments on “Sub amps”

  1. I still have the sub that you mention. It required spikes and 100 pounds of lead shot on top to hold it in place. It was somewhat of a novelty to me mainly because I really appreciated the enginuity that went into its design. However, I found that I could build my own and easily beat its performance with a more traditional sized design, partly because of the great drivers that are available to the hobbyist.

  2. Today’s post mentioning power and small box volume implies futhur discussion about equalization, I assume is coming later. Small size needs equalization for two reasons. The obvious one is bass extension. The other one is the transfer function which has a lot to do with bass quality.

  3. Paul,

    All of this focus on subwoofers gives one reason to speculate if there might just be a future horizontal expansion of the PS Audio product offerings?

    Simple to assemble, reasonable cabinets can be purchased off shore, 3 finishes 9 sku’s; don’t you have an ar.ti.san in the woodworking business : )

    Foundation 1
    Foundation 2
    Foundation 3

  4. As the box gets smaller, the resonant tuning frequency gets higher. This is because the resistance of air being compressed and rarified inside becomes greater, that is the change in pressure for a given movement of the cone increases. This can be offset by increasing the mass of the moving element, that is the cone and voice coil assembly and/or by increasing the damping. Damping in an acoustic suspension speaker system is provided by the stuffing. The speaker must work to push and pull air between the fibers. The collective surface area is enormous. This aerodynamic drag is related to the velocity of the cone and is the frictional loss damping force. Tuning an acoustic suspension speaker is therefor a matter of adjusting these three parameters. The choice defines the F3 and the system Q. One problem is the more stuffing you put in a sealed box, the less air will be left in it. In other speaker types, the restoring force and damping are in part or in whole the result of the speaker’s mechanical suspension. In systems with tuning ports, at the lowest frequencies the restoring force and damping are strongly influenced by the ease or difficulty the speaker has pushing air through the tuned port. For a specific design, it will be easy around the tuning frequency, difficult elsewhere.

    The parameters of the AR1, granddaddy of all acoustic suspension woofers are: Free air resonance 18 hz. System resonance 42 hz. enclosed air volume 1.75 cu ft. Q 0.707 stuffing weight 20 ounces. Power handling capacity 150 watts. Sorry I don’t have the rest of the Theil Small parameters.

    In most subwoofer/enclosure systems, the speaker is merely equalized for among other things its falloff below resonance for AS designs, probably its falloff above resonance for ported designs. This is where Bose got it wrong in the bass IMO with his original 901 which was an acoustic suspension design. The system resonant frequency was pushed up above 180 hz, the equalization was insufficient to compensate for the falloff (6db/octave in his design instead of 12) and the system Q was over 0.707 (critical damping) and was uncorrected. This resulted in a peak above resonance and too rapid a falloff below resonance.

    The limit of how much power a speaker can withstand depends on the heat generated and the maximum excursion before the voice coil is deformed or the cone is torn away from the suspension.

    A servo woofer does not need equalization to compensate for its FR. That is the job of the servo system itself. However, its job is made much easier and the results will be better if it is tuned to a low F3 and has a Q of 0.707 to begin with. Equalization then consists of limiting the high end and adjusting for room resonances. That equalization comes ahead of the amplifier. One problem for subwoofer amplifiers is that they not only need to have sufficient power, they need robust power supplies to keep the quiescent operating point of the output devices stable when confronted with the woofer’s back EMF.

    Looking at the IRS and Genesis 1.1 and 1.2 woofer towers it seems to me that the enclosure for each woofer is fairly small. Paul, have you got any data for the open loop FR of the system? The Genesis 1.2 and 1.2 Dragon double the number of woofers per tower from 6 to 12. Has the enclosure volume been increased? In the IRS only one subwoofer per tower has an accelerometer. It is assumed that each woofer/enclosure system will preform the same way so the waveform correction for one of them will be the same correction for all of them. This is a reasonable assumption. BTW, with the woofers in opposition to each other as a dipole, the recommend most effective placement is against the front wall facing sideways. This configuration was pioneered in AR9 and was the result of work by Roy Alison. However, due to the need to couple to the ribbon midrange which presumably will be away from the wall, that might create a phase problem at the crossover point.

    1. I think you mean a Q of 0.707 is maximally flat. Maximally damped is a Q of 0.50.

      I find the specs on the AR1 interesting since it was known for how low it went and your specs indicate it’s 3 db down at 42 hz, at least in an anechoic chamber. That’s a full octave left over. Of course there’s still room lift.

      As to equalized small boxes, I’m familiar with two, the Pipedreams and Scaena speakers. The original Scaena has a resonance at about 180 hz and is equalized passively to be flat to below 20 hz and then the response is reduced quickly to stop over excursion and running out of power since that means 12 db/octave equaliztion which is a ton of power. The transfer function of the equalization is chosen to be well damped and from what I’ve heard it plays very loud down to sub 20s and tightly. I was present at Harry pearson’s of The Absolute Sound with a Scaena designer playing a CD with a real volcano on it and both of them had bulging eyes and were waiting for flying woofers, by far the most impressive bass demo I ever heard.

  5. The Sunfire sub is an excellent example of the ills of the subwoofer market. Because the Federal Trade Commission does not deign to tap inside a subwoofer and measure the actual amplifier power, the manufacturers are free to LIE and put any number the marketing department likes on the box, Carver being the worst offender. No, they did not invent a perpetual motion machine that draws 2.7KW from a branch circuit that provides 1.9KW of electricity before the breaker trips. Their rationale is something like “if there was no impedance rise, the power required to produce the maximum excursion of the woofer required to equal the swept volume of the woofer and passive radiator taken together” or some such folderol. Yes, these subs are locationally unstable. They shake, rattle and roll anything they make contact with generating spurious sounds. If you you attach them to the foundation with concrete anchors they will generate substantial bass, but the group delay is enormous and if there is more than one bass note the intermod is horrible. I can’t stand to listen to them on music. I worked in a shop that sold them for Home Theater where the LFE and sound effect bass is all synthesized for speakers like this. Velociraptor sounded just like the real thing! (NOT)

    1. Actually the specs supplied with my unit indicated that they would draw a maximum of 300 watts from the wall at any time. I assumed that to get the 2700 watt output they drew 300 watts from the wall and stored it up until needed.

    2. You are right about the FTC ruling which was enacted in the 1970s to end the outrageous claims of power amplifier output. FTC required both channels driven, 1/3 full output for 20 minutes of “preconditioning”, and then the full power output with the stated bandwidth and distortion. Instantaneous peak power is not allowed although “headroom” or “maximum current output are sometimes used to suggest amplifiers can produce more power in short bursts. In theory a 120 volt 20 amp single pole circuit breaker can deliver 2400 volt amps (watts if the power factor is one) before tripping however it can only do that for 3 hours continuously, then it must reduce to 1920 watts or less to cool down. Class AB amplifiers are typically around 40% efficient (not counting heating filaments in tubes.) Class D amplifiers are much more efficient. Still, you can’t get 2700 watts output continuously from 2400 watts input.

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram