12" Sonotube Subwoofer
Down Firing, Bass-Reflex Cabinet
Introduction:
Very few people have heard what a GENUINE subwoofer can do for their audio system. Those that have can't live without one because a subwoofer is arguably the single most important component in your home theater system. Until you have experienced what a genuine subwoofer can do, you don't know what you are missing! A subwoofer adds the visceral impact to today's action movies and provides a great deal of realism to the movie watching experience. A high quality subwoofer provides the tactile sensations (sometimes without being audible at all - frequencies like 15Hz) that add so much to the mood and tension of a good movie. A first class subwoofer needs to do two thing: 1) move lots of air at low frequencies and 2) do so without distortion. The more air a subwoofer can move, the greater its impact. The laws of physics dictate that to move lots of air, you need a big driver (12" or more) with high excursion (15mm or more, one way), a large box (probably 3+ cubic feet), and a big amplifier (on the order of 250+ watts). Finally, a real subwoofer needs clean and strong output in the 20-25Hz frequency range (and often lower!) without distorting. Sorry Bose fans, your Lifestyle system doesn't meet any of these criteria...
This subwoofer is my first honest effort at building a Do-It-Yourself project. (Back in the 1980's I built a pair of speakers boxes from pinewood and installed randomly chosen drivers from Radio Shack, but we don't need to go into that - suffice it to say that they lived a good life and were an excellent set of test speakers until they succumbed to foam rot...) Getting back to the subwoofer, I spent a great deal of time reading web forums (especially the DIY section of HomeTheaterForum) and using the WinISD subwoofer simulation software to learn about and to compare various drivers. The three drivers that offered the greatest promise included Shiva, Earthquake, and the MASS driver from AudiomobileInc (now defunct). Of these, the MASS driver offered the best performance from the smallest box, but also cost the most: $300 for the driver alone.
My goal was to design a subwoofer with strong response down to at least 20Hz in as small a box as possible that I would be happy to live with for some time. After what seemed like months of reading, simulating, and learning, I finally decided to purchase the MASS driver. The particular model I chose was the MASS 2012 S24 driver. Its a 12 inch driver with dual 4 ohm voice coils (originally intended for car audio) that featured a free air resonance (Fs) of 22Hz and a maximum linear displacement (Xmax) of 20mm one-way travel. This heafty driver lives in a 135L vented enclosure that is tuned to about 19Hz. The result is a subwoofer that is capable of delivering nearly 110dB right down to 15Hz - an absolutely frightening level of output!
One of the nice things about building a project like this yourself is that there are fewer restrictions for your design. The primary restriction that many commercial-grade subwoofers have to adhere to is a design that, when packaged for shipping, weighs less than 75 pounds. FedEx, UPS, DHL, and the USPS won't accept any single package that weighs more than 75 pounds due to OSHA regulations - anything that weighs more risks injury for the carriers and must be carried by a private shipping firm at substantially higher cost. My completed sub weighs in at 90 pounds. 90 pounds is relatively light-weight for this sub because I used Sonotube for the enclosure (see below). A properly braced six-sided box for this driver could easily weigh 150 pounds all by itself! Although the Audiomobile driver is no longer produced, it was the driver that ushered in a new breed of high-excursion drivers such as the Blueprint drivers that weigh in at nearly 50 pounds all by themselves. There is simply no way to make a commercially viable subwoofer using a 50 pound driver - it is impossible to make a 20 pound enclosure that would withstand the force of such a driver without eventually coming apart.
The Driver:
The Audiomobile MASS 2012 S24 is truly a remarkable driver! It features a 12 inch powder coated aluminum cone, dual voice coils, a massive motor structure, and a power handling capability of 1000 watts. An aluminum cone is desirable due to its low mass (approximately 160g) and high rigidity, while the powder coating helps prevent the cone from "ringing" when it is driven hard. The driver itself weighs in at nearly 30 pounds, 25 of it is the magnet structure alone! The name MASS stands for "Magnetix Apex Symmetrical Stasis" which refers to the two stacked magnets on the back of the driver. It features two 3" 8-layer aluminum voice coils, a rubber surround (as opposed to foam which literally rots away over time), a sturdy cast (as opposed to the cheap stamped variety) aluminum basket, forged gap plate parts, and a vented motor assembly (for improved voice coil cooling). Sadly, I did not take any pictures of my subwoofer while it was in progress and now AudioMobileInc is out of business - so the pictures of the driver itself (below) were taken by someone else who posted them to their web page some time ago. If you claim ownership, I'll be happy to provide credit for the photography... Judging by more recent pictures of other drivers I've seen floating around, it appears that the MASS driver was custom manufactured for Audiomobile by the legendary TCSounds Inc.
The MASS driver was a relatively new driver at the time I built my sub (Sept 2000) and it had received a great deal of positive comments from both the professional and DIY communities. One important characteristic of this driver is that the surround (materials surrounding the actual driver and joining it to the frame) is made of rubber, not foam. Foam has a tendency to literally rot away after about 10 years, rendering your driver useless! The Thiel/Small parameters for this driver (they describe its technical performance) also fit all of my design constraints and performance requirements of needing a modestly small box to produce enormous amounts of output at 20Hz.
Qts= 0.42, Qms = 5.80, Vas = 92L, Qes = 0.45, Fs = 22Hz, SPL = 90dB, Re = 7.6ohm, Pe = 600w, Le= 0.507mH, BL = 19.80, Xmax = 20.0mm, Diam = 12in, Z = 4.0ohm, Sd = 0.046m^2, Mms = 160g
One important characteristic to determine is whether this driver is well suited for use in a bass-reflex (or vented) type of cabinet. Bass-reflex enclosures tend to be larger but provide lower frequency response than sealed enclosures for the same driver. Often, a subwoofer designed primarily for home theater will feature a vented enclosure as they maximize the low bass (20-25Hz) reproduction. The general rule of thumb is that a driver is well suited for use in a vented enclosure if Qts <0.40, Fs < 45Hz, and Vas < 4 cubic feet OR if the ratio of Fs/Qes > 83. This driver passes the first test closely enough (Qts is slightly above 0.40 - close enough for me) but does not pass the second test so well as Fs/Qes works out to be a 48.8. Well, since its and either/or situation, looks like this driver will function reasonably well in a vented enclosure. (Its true heritage indicates that it should be placed in a sealed enclosure of less than 1 cubic foot and installed in the trunk of a car - hence the 4 ohm voice coils.)
Cabinet Design:
My intent was to build a subwoofer that would not visually "take over the room." I wanted it to blend in as easily as possible, yet still provide all of the fun a good subwoofer is known for while watching movies or listening to music. Since our couch was flanked with one round end table on each side, the natural idea was to replace one of the end tables with the subwoofer; thus, the size constraints had already been determined.
Since the MASS driver is suited for use in a vented enclosure, the next thing to determine was the appropriate characteristics of its enclosure. After playing around with the WinISD software for quite some time, I had settled on a final set of dimensions that would achieve my design goals. The enclosure would contain a net volume of 135L with a 6" diameter PVC vent that measured approximately 36 inches in length, yielding an enclosure that was tuned to approximately 20Hz. These numbers work out well, too, as the general rule is to keep the port length under 5x to 6x the port diameter - just squeaked in on this criterion... The response plot from WinISD along with the cabinet and vent parameters are below:
Note the green response line in the graph is "flat" from 20Hz to 100Hz (the response curve higher than 80-100Hz is meaningless because my pre-amp filters out anything higher). No single frequency is emphasized more than any other frequency, especially at the lower end of the graph.
The next trick was to design an enclosure that not only fit these performance specifications, but also would make the enclosure easily disguiseable as an end table that looked like the one at the other end of the couch. Further research turned up a material known in the construction industry as "sonotube." Essentially, its the thick cardboard tube that is used to pour round concrete footers. You set up the tube, pour the concrete in and when its dry, simply peel away the cardboard. The tube shape bestows several benefits that are especially attractive for building subwoofers: once capped on the top and bottom, the cardboard walls will not flex (due to its cylindrical shape) as a result of pressure from the driver. Because pressure is exerted evenly across the entire inside of the enclosure and the cardboard cannot stretch (remember, its designed to hold thousands of pounds of concrete until it dries...) it makes an ideal construction material. The tube-type of subwoofer construction was pioneered by Dr. Hsu of Hsu Research, in their very highly regarded TN series of subwoofers.
Using sonotube, however, places one additional constraint upon the construction of the subwoofer - the driver must be mounted to the bottom of the enclosure. Doing so means mounting the driver horizontally, thus exposing it to the effects of gravity (see set of photos below). Many subwoofers that are carelessly designed use drivers that are not able to stand up to the small, yet constant tug of gravity which results in driver sag. Driver sag decreases the Xmax of the driver which, over time, will cause the driver to more easily bottom out when pushed hard. Fortunately, Dan Wiggins - the owner of the now closed AdireAudio and all around nice guy - authored a white paper entitled Calculating Driver Sag and Suitability for Horizontal Mounting. The main jist of this article is that if, due to the effects of gravity, the driver will sag more than 5% of its Xmax, the driver is not well suited for horizontal mounting. Running the spec of the MASS driver, I calculated a sag of approximately 2.5%, nearly half of the 5% rule of thumb. Looks like a "thumbs up" for horizontal mounting! Pfew!
After determining the proper enclosure volume, tuning, and driver orientation, the next thing to plan is the actual physical dimensions of the cabinet. Remember when calculating the volume of the enclosure that the design parameters gave us the NET volume. The gross volume (actual completed physical dimensions of the enclosure) must also take into account the volume taken up by the vent and the driver. In this case, a 36 in long 6 inch diameter vent takes up nearly 20 liters. The driver itself takes up another 3 liters inside the cabinet. So we need a cabinet that will enclose a total of 135 + 20 + 3, or nearly 158 liters (approximately 5.5 cubic feet). This is a fairly large enclosure, but using a 24inch diameter sonotube cut to a 23 inch length fit the bill. The endcaps, or plugs, for the sonotube consisted of 2 layers of 3/4 inch thick high density fiber board glued together and cut such that one layer fit neatly inside the tube, while the second layer acted as a "lip" and remained outside of the tube. Thus, the 23 inch length of sonotube accounted for having 1.5 inches of the internal volume taken up by the endcaps.
Cabinet Construction:
So, this brings us to bit of a problem: the enclosure (at 21.5 inches tall) is too small to fit a 36 inch vent inside! Not so! In addition to getting a straight length of PVC, I also purchased two 90 degree elbows. When using an elbow, you have to be a little careful about how you determine its "length". The easiest method is simply to average the minimum and maximum lengths of the elbows, once joined. Just take your tape measure and measure across the "top" of the elbow and average that with the "bottom" measurement. I don't recall what my specific measurements were, but the elbows formed a (average) 26 inch long, 180 degree bend when they were joined. Next, I cut 14 inches of Schedule 40 PVC (the thick stuff) and inserted 3 inches of the straight section into the elbow that I had just created and glued it all in place. The end result was a candy cane shaped vent that measured 36.5 inches in length, yielding a calculated tuning frequency of just above 19Hz. With this configuration, the maximum velocity of the air moving through the port at Xmax works out to just about 0.07 Mach - the conservative recommendation is to keep the velocity under 0.10 to prevent compression of the air in the vent that will reduce the output and mess with the tuning of the enclosure when the subwoofer is being driven hard.
When building a subwoofer, it is also extremely important that all of the joints and places where materials come together are perfectly air tight. The only place where air is allowed to escape the enclosure is through the vent - not around the driver, at a seam in the wood, or at the speaker input cup. Thus, it is very important to use liberal amounts of a high quality glue and plenty of 100% silicon caulking around all of the seams and joints. Conveniently, the MASS driver came with an enormous rubber gasket that works well to form a seal when the driver is mounted to the cabinet. Any air leaks will quickly reveal themselves during testing - at high levels of output, even the smallest of air leaks will produce a suspicious "farting" sound.
In the picture above on the left, the driver diameter is 12" and the vent is 6" inner diameter PVC. The entire tube is just over 24" wide and the completed Subwoofer stands just over 28" tall with the legs attached.
Using my jig saw, I cut three holes in the bottom plate of the enclosure: one for the driver, one for the vent, and one for the input cup (see photo above). The input cup was carefully glued and screwed in place and then heavily caulked to prevent air leaks. The hole for the vent was purposefully cut too small so that I could widen it as necessary with a hand file until the vent just fit inside. Final positioning of the vent required a rubber mallet to push the vent through the hole one millimeter at a time - it was a very tight fit. After it was properly positioned, the joint between the vent and the endcap was securely glued and caulked. As it turns out, the combination of the pressure fit and glue are more than sufficient to hold the vent rigidly in place. To hold the driver in place, I used a number of "T-nuts" mounted to the inside of the bottom endcap. Fiber board does not respond well to screws (especially if you need to remove the driver for patching air leaks), so I placed the driver in its cutout and marked the location of the screw holes in the driver's frame. The holes were drilled out and then widened on the inside to accept the shank of the T-nut. The T-nuts were also glued and caulked in place to prevent air leaks. The length of the screws I used to mount the driver to the bottom endcap were selected so that they would provide enough grip on the T-nuts, yet not break through the caulk that sealed the holes.
Next, I sanded the wax coating from the inside walls of the sonotube and glued some polyfill to the inside of the walls to help prevent echo and reduce the impact of standing waves inside the cabinet. The wax is there to aid in the removal of the cardboard from the concrete when your construction project is complete, but its terribly difficult to glue anything to it so it needs to be removed. After the batting was glued in place, I joined the top endcap to the tube and applied a liberal amount of caulk.
The final step involved gluing the bottom endcap to the tube and then mounting the driver. Mounting the driver is always the last thing you do for two reasons: 1) you need to carefully caulk this final joint from the inside of the enclosure to prevent air leaks and 2) the fumes that escape from the drying caulk are known the dissolve the glue used to hold the speaker cone to its surround! Yikes! Be sure to allow the caulk to dry for 24 hours before mounting the driver! You've been warned!
Listening Impressions:
In some situations, it can be very hard to describe the results of your latest project. Some claim that the builder can never be objective because of the time and energy invested in the project. Well, this is one project that has very clear results. While I've never heard a T-Rex in real life (who has?), I feel quite certain that I know what one would sound like if it crashed through my living room! Prior to using my new subwoofer, I was using those old Radio Shack speakers with their 8 inch woofers as my makeshift "subwoofer". Well, there is simply no fair way to compare these two situations. Bass response was now amazingly deep and shook with authority - even being driven by an old 60wpc Fisher integrated amp that I purchased in the 1980s! The condo that we were living in at the time had a concrete slab under the kitchen floor. This subwoofer had no trouble at all making this concrete slab shake and resonate! I remember when my sister came to visit and I played a section of the Telarc disk containing Tchaikovsky's 1812 overture: "WOW!" she said, "They sound like REAL canons!" Movie watching would never the be same again! Or so I thought... Fast forward several months...
Part II: What a Difference an Amplifier Makes...
While running some test tones through my subwoofer while using my SPL meter (doesn't every sub owner do this?!?!), I wondered what it would sound like if I drove it with a pair of my Marantz MA500 monoblock amps instead of the cheap 15+ year old Fisher receiver I'd been using. The Fisher amp put out 60wpc into an 8 ohm load, but the voice coils in the sub were 4 ohms each. The MA500 is capable of 180w of output into a 4 ohm load, using two would deliver a maximum of 360w to the sub. So, I hooked up an RCA signal splitter and ran the sub out from my preamp into two MA500s, driving one voice coil with each amp.
The difference between the amps was immediately obvious! There were three primary benefits from using much better amplification. 1) There was a considerable amount of "chuff" or "noise" emanating from the subwoofer when running a 15Hz test tone through the Fisher receiver. I'd never noticed that it was there until I ran the same 15Hz tone through the MA500s. This time, there was no audible noise, just a thick, heavy feel to the air in the room while everything shook! Wow! What an improvement! 2) Using my SPL meter, there was also considerably more output in the 15-25Hz region through the MA500 amps. Thus, it seems that the Fisher receiver was really unable to reproduce sound frequencies as low as 20Hz, despite what the manufacturer's paperwork claimed. 3) This last improvement came when I played some CDs through the better amps. Specifically, I was listening to the opening track "Sweet Dreams" on the Greatest Hits CD from the Eurythmics. This track features an especially heavy bass line from both kick drums and a synthesizer. Here is where the differences between the two amps were easy to identify! The bass notes sounded much cleaner on the MA500s than they did on the Fisher. There was a "muddiness" with the Fisher amp that was no longer present - each hit of the kick drum sounded more "crisp" and "clean." It is kind of like the difference between a sharp "crack" sound and a dull "thud" sound. Also, the sound of the drum was much tighter through the better amps. Where the old receiver had allowed the sound of the drums to resonate and echo, the presentation of the drums was much better controlled through the MA500s. There was no "lingering" sound that didn't belong. The drum hit hard and then immediately stopped, rather than "fading" away as it did before. Guess this is the result of the substantially greater damping factor of the better amps.
Well, that's all I needed to hear! My next purchase was something more comparable to the Marantz amps: an Adcom GFA545 mkII amplifier that I found on E-Bay for a real bargain! The 545 is able to deliver 150wpc into a 4 ohm load and the mkII version of this amp was significantly re-designed and improved by Adcom for better performance while driving 4 ohm loads. The Adcom amp is able to control the subwoofer every bit as well as the Marantz amps and is such an improvement over my old receiver the difference is not even funny! Where the old unit could probably muster 100 watts total into the sub, the Adcom amp delivers 300 watts total with ease. Listening to music and watching movies is a new experience with a better amp behind the subwoofer! Subtle details creep out with greater refinement than ever before adding to the overall experience of watching movies. That old receiver is now driving the bass shakers attached to the bottom of my couch... It works just fine here!
To say that this subwoofer turns in a performance beyond my expectations is a gross understatement! In the commercial arena, I would put this subwoofer in head-to-head competition with any commercial subwoofer costing three times as much and expect it to come out even or on top. Remember that 10:1 ratio of parts cost to retails cost? This is another example of doing something yourself, doing it well, and saving a great deal of money - all at the same time.
Measurements:
Two items that are necessary to measure the response of a subwoofer (or any other speaker) are a CD that contains test tones (like the BassZone Test CD Volume 1 that is available at Acoustic Elegance - formerly Stryke Audio) and a sound pressure meter that can measure the output of the speaker.
Be aware, however, that the Radio Shack SPL meter (analog and digital models alike) is not entirely accurate. After you make your readings, it is necessary to add "correction values" to your measurements that are frequency dependent. The chart below contains the proper adjustment for each frequency range. To find the true measurements just add the figures below to your meter readings for each frequency. For measuring your subwoofer, only the first column of corrections will be needed. For example, if a 20Hz sound registers on your RS SPL meter as 70dB, you need to add 7.5dB to that reading, thus the actual sound pressure level for that 20Hz signal is 77.5dB.
Using my test CD, the in-room measured response (after taking into account the adjustments from the table above) reveals that the actual output of the subwoofer as measured from the seating position (green line) is quite strong well below 20Hz! You can see the output of the driver (blue line) dropping off below the resonant frequency of the driver (as expected) and the output of the port (pink line) hitting its maximum at the tuning point for the box (also expected). There is also a 3-4dB dip in the response of the sub that runs from about 50Hz to 80Hz, but this doesn't bother me at all as the mains will start taking over as the frequency increases. This dip is most likely caused by an interaction with the walls and the general shape of the room.
Overall, this is VERY respectable performance that varies just slightly from simulated results shown above
After properly calibrating the level of the subwoofer to match the level of the remaining channels (all channels were calibrated to 75dB using test tones from my preamp), I fired up the Pod Race scene from Star Wars I: The Phantom Menace. Hold on to your seat if you have a good subwoofer and run this scene at reference level (sound levels will peak just about 110dB)! This scene is well known to have particularly strong response down to 10Hz and below! Reference level on this scene pushes the Adcom amp into clipping - remember, that's 300 watts into the sub! (Perhaps I need a bigger amp because the sub has shown no signs of bottoming out even at this level of playback.) The second time I ran this scene at reference levels, several pictures "walked" off the edge of the fireplace mantle (that is 20 feet away and on a different wall) and fell to the floor due to the intense vibrations! These sound level measurements were made in my living room which is a shared space with the kitchen in the house. The two rooms together measure 24x15x8 feet. This is exceptionally impressive performance for a single 12" driver in a room this large!
Well, what else can I say except that "Bose fans need not apply..."