Subwoofer enclosure in steel?

I am new to the the DIY Hifi stuff , so this is my first attempt and I thought a Sub would be a good start. The choice of steel is a style decision, The decor in my house is industrial and the sub should compliment that

I like curves and I have access to a lot of pressure vessels and light pipe bends which got me thinking about making a sub enclosure in steel. Most of the steel I see is 4 to 6mm wall with plenty of compound curves which I think should stiffen the enclosure. I am not sure about resonances other than that the transverse modes can be quite low frequency but they require asymmetric loads to excite them in a pressure vessel, pressure loads should be symmetric where the wavelength is many times the enclosure size. Constrained layer damping just in case ? I will use fill as this is to be a sealed enclosure operated below resonance.

The driver I am looking at is a PA driver 15" 98 db/w good for 500W RMS with an xmax of +/- 11mm Vas around 170l The intended use is home hifi to get 32 hz organ pipes and the thud of large Asian drums. Asian drumming is often fast paced and off beat I am thinking about crossing over at 55 hz I have about 300W RMS to play with.

Has anybody made a steel enclosure? Thoughts about using high sensitivity PA drivers? Good speaker design books? Should I use an 18" driver instead or maybe two 15" drivers.
since your goal is purely asthetical and will be a first attempt why not just follow a conventional
and tested design out of wood/mdf then use the steel just to plate it. a metal veneer if you will..
That way you will have something that works well and looks like it has been entirely made out of steel.
The steel design will be either cylindrical with ellipsoidal ends or a conical elbow, due to the difficulty of making curves of any type in wood let alone compound curves I never considered it, even concrete or FRP would be options except making the mould is no fun. The steel comes preformed to the shape I desire, I know I am lazy.

Unless you are suggesting build a square wood box inside a curved steel shell in which case I might as well use wood bracing to dampen any resonance in the steel shell and skip the box.

Use a proper subwoofer, unless you planning on huge subs you will find
no gains in using high efficiency drivers, bass efficiency will be poor.
I cannot see the point of CLD at such low frequencies.

Two mono subs spread around the room are better than one bigger one.

download a similator, WinISDpro or Unibox and play with them.

rgds, sreten.
Last edited:
Have a look at

What you have in mind sound very good to me, the material choice is not an issue. Try to ensure that the wave front from the back of the woofer does not strike untreated metal.

As for your choice of drivers, most Pro drivers are not good at very low frequencies. Look at decent subs designed specificaly to go low.

Being a sculptor perhaps you could do what it takes to make sure you have perfect cylinders and symmetrical ellipses, since that is what is required to load the vessel material with tensile and compressive force rather than bending. If you did that, there would be absolutely no way to match the volume or distortion performance of a 1/4" thick steel vessel with a wooden box subwoofer for a single 15 or 18. I would make a pair.
Don't know what your room is like..but if you have a fireplace, I made subwoofer
in the last house I lived utilising the old chimney breast.. That Thing couldn't have
been Any better for the low organ frequencies you require .. or have been any more industrial :)

No need for a super duper woofer, excellent results can be had from a less capable, cheaper driver, since it will be
in it's own natural 20-30 foot solid concrete resonance chamber. which corresponds perfectly to
low 30 hz wavelength.

My setup was just plated underneath the chimney.. yours could then have an elbaroate
steel sculpture to hold the driver in place, leaving you much creative freedom..
Just an idea.

if not maybe some inspiration could be had from the RAAL requisite 'Eternity' OMNIS

Last edited:


2008-01-08 11:51 pm
I wouldn't agree that the material choice is not an issue. I suspect that steel is too lively a material to lend itself to loudspeaker enclosure construction. Wood and MDF are deader and more likely to provide good damping. 'Course it all depends on the thickness ratio, there's a big difference between a 6 inch cube with half-inch-thick walls and a 3-foot tall floorstander made of sixteenth-inch sheet.

wakibaki I would recommend thick steel sheet in typical pressure vessel construction is nearly ideal only for a subwoofer. At higher frequencies things get more complicated.

As an aside I think when you're dealing with high performance pressure vessels the ends are usually spheric sections except for an interface bend so that high stress doesn't develop at the weld. For the relatively low pressure differentials of a woofer volume the stress there can be ignored and spheric/cylindrical figured optimum from a minimum surface distortion under drive standpoint.
Last edited:
For the relatively low pressure differentials of a woofer volume the stress there can be ignored and spheric/cylindrical figured optimum from a minimum surface distortion under drive standpoint.
Exactly what I was thinking pure tensile load if possible Even thin walled (1.2mm) 2 stroke exhausts driven with pulses around 200db don't deflect much. A lot of industrial pipework is subjected to internal pressure pulsations with minimal sound radiating from the pipe surface.

As for the use of a high efficiency driver, is the problem due to the chance that the voice coil may leave pole piece area as the excursion increases? I could measure cone deflection vs drive current over the range of excursion but that would mean buying the driver just to find out. I have used loudspeaker motors as current to force transducers in the past and ended up using hifi ones because of their greater linearity. Are automotive subwoofers any good? low Xmax and plenty of snake oil doesn't inspire confidence.

Aurality I love that idea I bet no birds sat on your chimney Unfortunately I do not have a fireplace to convert. The Raal units are very attractive.

I will download a simulator, I am new to Ts parameters but understand the physics.
I'm not sure what you're looking for in a woofer driver but if you get a simulator what the TS parameters do for you will become quite apparent and there are drivers available to hit any practical level of performance, for a price. High efficiency does pretty much mean reduced excursion capacity, though, especially for a given dollar.

One major difference between a spontaneous internal pulse pressure load in any enclosure and a speaker cabinet is the driver basket is bolted to the baffle. At higher frequencies basket drive can easily excite resonances in a thin steel sheet, but you're not dealing with that here. You can pretty much figure that the entire enclosure is going to move, if any part of it moves, in the low tens of Hertz. You could use an opposed driver configuration to cancel that as well but then need a larger diameter cylinder to avoid making a long tube which at some point will develop some resonant response of its own. Size would come to the aid of surface resonances. Ultimately if you didn't like how the tank rang when you were done you could always wrap it with all kinds of things. Layers of aluminum flashing, tar paper, lead, bent MDF, with some kind of glue or goop, whatever. You could place the damping on either side, but for a sub I don't think you'll really need it. Just some fill if sealed to keep cone harmonics and whatever upper frequency energy makes it through the crossover from having a field day in there. You'll probably want at least a thin layer of something or to make a minimum internal surface roughness just because it's cylindrical and hard which can keep things bouncing around "forever" at high frequencies.
Last edited:


2006-01-09 7:04 pm
Hi metalsculptor,

According to your initial post you are intending to build a seal enclosure, and-as has been pointed out already-a PA driver is not really suited to produce 32Hz in a sealed enclosure; I do kind of wonder which 15" PA driver has a Xmax of 11mm, 98db/w, and is rated for 500W RMs. My recommendation would be more along the lines of a CSS Trio 12, if for no other reason than that it's T/S parameters make it suitable for a wide variety of enclosures, including sealed, and that it has a Xmax of 20mm.

A fun project might be a coffee table build around planet10's concept:

with a stainless steel enclosure, and one, or two or even four drivers exposed into the center space, the whole thing suspended on big brass cones, and visible through a glass table top. With the Trio 12

there would really not be any need for an isobaric mounting, and the internal drivers would not be visible anyhow (unless the whole top would be a glass/see-through construction). Another way to do the same thing would be something like a single turn (coil) of a spring made from large cross-sectional area tubing with the woofer(s)-exposed-added to the ends.

I'm definitely going to hang around to see what you come up with. It seems a nice excursion from the usual.

I am starting to get an idea now, 12" drivers would only be of interest in pairs and I have to find pressure vessels of suitable diameter or add a lot of filler to occupy the internal volume armaflex sheet would probably do as it it is closed cell and low Q also it costs nothing. Also my amp is rated into 4 ohms and there is no way I will series connect drivers so 4 ohm 12" drivers are out of the question unless I run another amplifier module off the same power supply. Australia is not the best place for getting stuff so choice is limited.

As for the PA driver I was looking at, the Shenzen SB15-5 the 40Hz resonance would not have been a problem as I would have bumped the resonance up to 55Hz with the enclosure and used electrical compensation for the roll off in sensitivity.

Interesting that spiders are still the main method of voice coil location, when using the speaker motor as a current to force transducer I machined a lightweight bearing carrier with guide pin in the centre of the magnet, rigid location and very high compliance, the voice coil would hover over the gap with no load. There may be stick slip issues but large 12" hard drives used a similar technique for head positioning

I will get busy with a simulator and driver catalogues to get approximate volumes and bring home a few pressure vessels and see how they look. Stainless is a problem because stainless pressure vessels are scarce and often thin walled, Toroidal pressure vessels are available but in a very limited size range.


2010-07-21 11:21 pm
Hi metalsculptor,

Why not cast your own enclosures, if you use Bells as an example they can be tuned to any frequency by shape & just as easily detuned to eliminate unwanted resonance etc by shape & or wall thickness.
For what you want you could have some great brass enclosures of your own design & polish them up to a high finish, scrap brass etc is everywhere, even aluminium is easily cast as you will know.
The mind runs riot when you think of what you could actually do by casting your enclosures. Sand casting is easy & finishing would be well worth the effort.
Before you finish them try them out & if you want something different just melt the suckers down again & recast a different shape.

You only need timber as the pattern for the sand, that's easy.

Might be worth thinking about, maybe use some of the Bell technology to head you in the right direction for tuning the perfect enclosures.

They would look pretty spectacular & sound great when you get them right, you could recess them into anything as well for looks.

There's a heap of information about harmonics & bell shaped castings, here's just a quick site I found.
Australian Bell

I may give this a go myself when I have a bit more time.

Last edited:


2008-09-04 1:47 pm
You could get away with a 1.2mm skin if you designed it so that there was no assymetry in the way the pressure is distributed so that it tries to stretch the skin rather than trying to bend it. If you have pressure variations along a curved axis, pressure differences in adjacent sections will cause it to deform. A good shape for this is a cylander with the woofer at one end, because it loads evenly around the circumference, the pressure is the same all the way around. There are longitudinal differences, ie the pressure varies down the length of the tube, but they are relatively gradual because the waves involved are longer than the cabinet, and the skin is stiffer in that direction anyhow.
Here's an example of a T-line, using stainless steel air conditioning pipe and a JL Audio high end car subwoofer (these are also used in the Fathom HT subs).

JL Audio 13W7 Sub in new Stainless Steel Tube 147dB! - YouTube

The reason why more speaker boxes aren't made of metal? it's hard fabricating metal, other than that I would highly recommend steel. You could add the damping ratio by using something like green glue, or filling cavities with lead shot or sand. Steel is very good in that it's stiff, and it's unlikely that the resonant modes of a very stiff material will be low in frequency – ideal for a subwoofer.
Last edited: