shavlogo2.gif (3085 bytes)        

        2-way, 12" PA cabinets

7/03 - Larry Mundy -

Since one of my initial articles here on the Shavano site, A Speaker Cabinet Project in Words and Pictures, I’ve had quite a number of emails from people who want to build something with a little less expensive drivers than the $400-$500 worth of woofers that went into that project. A lot of the inquiries go something like "could I substitute 100-watt speakers for the ones you used?" Well, sure you can, but my usual response is "hey, if you’re only going to put 200 watts worth of power-handling capacity into a box, why build such a giant box?" That is, if you don’t need the 800-watt power capacity of the mains described in that article, why not build a smaller box with a single powerful speaker, and have less bulk and weight to haul around? You can indeed make 2, or 4, or 8 cheaper speakers equal the power-handling capacity of a lesser number of expensive ones, but you end up with a huge box that has to be lifted and transported. My rule is, build the smallest box, with the most powerful speaker(s), which will accomplish the job at hand.

Actually, we’ve had the same thought more than once, when we’ve been called to perform in a small venue, cart those 75-pound main speakers with us, and then use them at a level far below what they’re capable of. And like so many bands, as we got larger and more powerful equipment, we disposed of some older, smaller stuff. So we decided to build "PA Lite" for those occasions when our main system would be overkill – and since "PA Lite" is exactly what most groups probably need for typical small-venue performances, to document the construction process.

Driver Selection.

We intended these to be bi-ampable with small subwoofers but capable of reproducing pretty much full-range by themselves. The design has a single 12" woofer, high-frequency horn and crossover in about a 2 sq. ft. box, a very typical setup for medium-power PA use. [For a single-speaker enclosure I’d consider a 12" driver the "sweet spot" for its radiating efficiency vs. high-frequency response capabilities. 10" speakers can reproduce a full range of PA sound, but their cone surface area ("Sd" in the specs, if it lists that) at about 54 sq. in. is 32% smaller than the 80-square-inch cone area of a 12" speaker, so they move slightly less air for the same voice-coil excursion, and it’s difficult to find truly heavy-duty 10" speakers with greater efficiency or excursion at reasonable prices. 15" speakers (132 sq. in) are 60% larger than 12’s, but usually don’t reach as far into the midrange frequencies and thus have to be mated with a high-frequency driver that can cover that part of the midrange – which can get expensive; 15’s in full-range cabinets are more often found in 3-way designs that lend more midrange help.  And, a cabinet for a 15" speaker has to be pretty large for pole-mounting. These are generalizations and the frequency response and efficiency characteristics of a given speaker depend on many more variables than mere cone size, but with the average pro speaker in the average 2-way configuration I think the laws of physics point to a 12" driver as the best compromise]. There are literally hundreds of 12" speakers to choose from; my recommendation is to buy the highest-rated speaker you can afford for this purpose, because a pair of them when used alone will have to handle your entire PA.  In this project I will use a pair of Delta Pro 12’s from Eminence – not their highest-capacity model but still rated to handle 300+ watts apiece, with a 2.5" voice coil, a 56 oz. magnet, and superb efficiency. You can find these at a "street price" of about $120 apiece, or their steel-framed non-"Pro" cousins with basically identical specifications for about $80. Or there are a world of other choices; I’d just recommend you read this article and look for "pro sound" speakers with the largest voice coils and magnets, and particularly with the highest efficiency ratings, you can afford.  I wouldn't recommend a "subwoofer" design with extended "Xmax" or low-end like the Eminence "LF" models, because those will poop out too early in the midrange and require too much midrange coverage from your tweeters; speakers designed to reach to 3000 or 4000hz, at least, in the upper end of their response curves work best for 2-way PA uses.   Most appropriate speakers will have an 8-ohm impedance rating, which will work with any amp and still allow daisy-chaining of multiple speakers in parallel with most amps.

The Deltas have useful response to nearly 5000 hz, and the more of that midrange I can squeeze out of my woofers, the longer and easier life my tweeters will enjoy, so I will design a crossover that lets my woofer work to at least 3,000 - 4,000 hz before the tweeter starts kicking in. I will use a conventional "compression driver" with standard 1-3/8"-18 threads, screwed into an inexpensive plastic horn lens (i.e. the "horn part" of the unit). If you’ve read any of my other articles, you know that I generally use piezo drivers for high-power applications, mostly because I’m cheap and truly high-power compression drivers are fearfully expensive, even though they allegedly sound "smoother" than piezo drivers. For these cabinets and with a proper crossover in the upper midrange, I can get by with reasonably-priced compression drivers rated in the 75- to100-watt range.

At the crossover frequency I’m using, the size of the horn lens probably doesn’t matter, but I prefer to use a horn lens wider than the woofer I’m using and let that dictate the cabinet width. That way I can use a smaller grille over just the woofer, and leave the horn exposed for looks. So my horn lenses are 15" x 5" molded-plastic models available at Parts Express and elsewhere in the $10 range. Attached to these will be Selenium DH200E compression drivers, which are rated to handle around 100 watts, but only above 2000hz ($35 or so each).  This is the first time I've used Selenium products, which I understand are made in Brazil, but so far I've been pretty impressed by the sound and performance of these relatively-inexpensive little drivers.

[As a side note, it is typical in two-way hi-fi or even DJ speakers for the tweeter component to have a lower power rating than the woofer by a factor of 4, 6 or even more, because prerecorded sounds are usually "compressed" and free of random high-frequency peaks, and because most popular music has a frequency distribution that requires far more power capacity from the woofer than the tweeter.   When you are producing (as opposed to reproducing) sound, this is not always the case, and I have burned up quite a number of compression drivers proving that point.  For PA cabinets, I would recommend no greater than a 3:1 power-rating ratio between the woofer and tweeter if possible.  You may use your PA system only for vocals (I would not recommend routing both vocals and instruments through a small PA at high volumes), but your microphones are probably capable of picking up frequencies up to 15,000 or 16,000 hz.  In a typical band, high-frequency instruments (cymbals, for example, and overtones of instruments with high-frequency content) will "bleed" through the microphones into the PA.  It takes a pretty hefty tweeter to withstand that sort of treatment, so try not to skimp on tweeter-power ratings.   And try not to leave live, unattended mikes sitting onstage in front of drummers or instrument amps during a performance!  I would never buy a performance microphone without an on-off switch for just this reason.]

My crossover will be the simplest kind imaginable, a "first-order" or 6 db/octave design as discussed in this article.  The steeper cut-off curves of 12 db/octave or 18 db/octave crossovers protect tweeters better if you choose a crossover point dangerously close to the lower end of the tweeter's capabilities, because those designs roll off the lows much faster (put another way, the "rolloff" curve is much "steeper").  But, they are more complex and can add phase shift and other weird nuances to the sound; I prefer to set a higher crossover point and use a simpler crossover.  I will design for a crossover point of 4000hz, using a 5.0uf bipolar capacitor and a .32mh inductor coil, both with 250V ratings, which is about as heavy-duty as they come (you could get by with less, maybe even 100V ratings, but when something blows inside a cabinet it stops working, and it's a major pain to disassemble, troubleshoot and repair/replace.  My favorite cabinet designs are the one's I've never had to open up after the initial construction).  They are wired as in the diagram in the article, with the capacitor in series with the tweeter and the coil in series with the woofer.  Simple and effective.

Matching Efficiencies.

After you have decided on a woofer-tweeter combination, dig out the manufacturers’ specifications and compare the relative efficiencies of the two, sometimes expressed in decibels of output when fed with one watt at the distance of one meter (db/1W/1M), or sometimes abbreviated "SPL" or "efficiency" and expressed in decibels (db). What you want for smooth response is for one watt of power fed to your woofer at 1,000 hz to be approximately as loud as one watt of power fed to your tweeter at 10,000 hz – and that almost never happens. It’s common for a tweeter or horn driver to be much more efficient than a woofer, and play louder at a given power-input level, and it’s particularly common for compression drivers to have insanely high efficiencies. For example, your woofer may be rated to produce 95 db from one watt of input – not a bad specification. But your horn driver may produce 105 db from one watt of input – a 10db difference which will make your horn sound twice as loud as the woofer (see discussion of efficiency ratings in my speaker selection article). If both your woofer and tweeter have close to the same efficiency ratings, skip the rest of this section and have a beer or something instead.

You can’t make your woofer more efficient, so what you have to do for smooth response is make your tweeter or horn less efficient by adding electrical resistance to the tweeter circuit.  There are two ways to do this, either an adjustable resistor known as an "L-pad" which lets you adjust tweeter levels on the fly, or a simple resistor which fixes the relative power of the woofer and tweeter.   L-pads have the advantage of adjustability, but the disadvantages of fragility and complexity. An adjustment knob has to stick out of your cabinet somewhere where it can be set wrong or broken off. But if you do use an L-pad, get the large "100-watt" model for maximum burnout resistance.

My usual course is to use a couple simple, fixed resistors instead, which cost about 50 cents, have no moving parts and are essentially burnout-proof. What rating do you need? Well, as an approximation, a resistor with a rating (in ohms) equal to the impedance of the speaker it’s hooked to, placed in series in that speaker's circuit, will reduce the relative volume of that speaker by about 3 db. So if you have an 8-ohm horn driver that’s 3db more efficient than your woofer, an 8-ohm resistor in the "+"-side wiring, after your crossover but before the horn or tweeter "+" terminal, should smooth the relative response of the two.

However, if you’re using a typical crossover, just adding a single resistor will also change the crossover frequency, which you probably don’t want. Check the chart of crossover values for a first-order design here and you’ll see that doubling the impedance of a driver, with a particular set of crossover values, lowers the crossover frequency by a factor of 2. So with a single in-line resistor intended to cut down the relative loudness of a tweeter, you can push the crossover point to that tweeter down into the "danger" range prety quickly. The better approach is to use two resistors, one across the voice coil (i.e. between the + and – terminals of the tweeter) and one in the "+" line between the crossover and the tweeter, like so:

                                 tweeter resistors.jpg (24047 bytes)

The resistor across the voice coil actually lowers the impedance of the tweeter by giving the signal an alternate path. If you think of the voice coil as just another resistor, then by putting a resistor across the coil you lower the total resistance just as if you had hooked another speaker in parallel with the tweeter, or as if you had hooked two resistors in parallel. The total resistance of the tweeter-resistor combination is expressed as (R1 x S1) / (R1 + S1) where S1 is the tweeter’s nominal impedance, and R1 is the value of the resistor hooked across the voice coil. For example, if you have an 8-ohm tweeter and hook a 15-ohm resistor across its voice coil, the resistance of the result is (8 X 15) / (8 + 15), or 5.2 ohms.

But resistors hooked in series are additive. So if you then added 2.8 ohms of resistance (R2) in the "+" line between the crossover and the tweeter, the crossover would be "seeing" an 8-ohm load just as before and crossing over at the proper frequency, but the tweeter would be playing at a lower level relative to the other speakers because there is an extra 2.8 ohms’ passive resistance in the circuit, and roughly 1/3 of the power that’s getting to the tweeter is being "shunted" through the 15-ohm resistor.

To attenuate the tweeter just slightly, use a larger value for R1 and a smaller compensating value for R2. To attenuate the tweeter drastically, use a smaller R1 and a larger R2. Just so the total resistance of the circuit is what your crossover expects (usually 8 ohms), this sort of circuit will keep the crossover point where you presumably designed and intended it to be. Resistors are cheap and lend themselves to experimentation. Play a recorded song you’re familiar with through the completed speaker and aim for a natural sound with the amp tone controls set "flat." Or you can get an inexpensive test CD which puts out test tones at a constant level at different frequencies, and try to smooth the response that way. If you really want to be scientific, you can use a sound-pressure-level meter and test CD, and take your ears out of the equation (if your ears are as old and abused as mine, they may be losing high-frequency sensitivity anyhow, and what sounds good to you may be overly bright-sounding for your audience). Perfect frequency response is neither necessary nor achievable, and you can always use equalizers or tone controls to "tweak" the balance electronically, but it helps to start with a speaker system that’s reasonably well-balanced in the first place. In my case, the woofers are rated 100 db and the compression drivers 105 db, but I like a bright sound from PA cabs, so I will use the setup described above with 15 ohms ("R1") across the voice coil and 3 ohms (couldn’t find a 3.2, but it’s close enough) in series as "R2", figuring that will cut the tweeter output about 3 db. This will also help the power-rating mismatch between the two, as the resistors will cut down on the relative power going to the tweeter. Use the biggest, heaviest resistors you can find in the ratings you need – the 10-or 20-watt, big rectangular ones work great.

Box Design.

When I don’t use it in a bi-amped configuration, I want decent low-end response from my woofer, which can be helped (or hindered) by the size and design of the box I put it in. There are a number of shareware or freeware programs which will help you design the optimum enclosure for your woofer – if your woofer is carried by Parts Express, the free "PEBox" program downloadable from their web site is ugly but serviceable for this purpose; otherwise you will need a program that lets you key in woofer parameters. But the optimum enclosure size is not always something you want to carry around, and often you can get by with a much smaller and lighter enclosure at the expense of just a few hz of low-end response. So typically I design enclosures (other than subwoofers, where low-frequency response is critical) "backwards," coming up with an enclosure size and volume, and then querying the computer program as to the results I can expect.

The width of my horn dictates that the inner walls of the enclosure be about 15" wide. The depth of the horn-driver combination, plus a recess of at least 1" from the cabinet front to the horn face, plus the thickness of the rear panel, plus a little margin for error, dictates that my cabinet be at least 14" deep overall. The height of the woofer, the horn, and some extra space on the cabinet face for ports and grille attachment, dictates a cabinet at least 21" high at the baffle, or 22" high overall if I use " material. And I will recess the woofer about 2" from the front face of the cabinet so that the grille will be slightly recessed when I’m done. So the interior of the box will be about 15" wide, 11.5" deep, and 21" high. That’s 3,622 gross cubic inches, or roughly 2 cubic feet. I tell the program that I am putting my particular woofer into a 2-cubic foot box, and it will tell me what size of ports to use, and what low-frequency response I can expect from the result. If the low-frequency cutoff is too severe, I will design a bigger box, but by changing variables within the program I can usually come to a good compromise. In this case I found that within a 2 cu. ft. box ("F8" key if you’re using PEBox) I could achieve a low-end rolloff (F3, or the point where the woofer output is down 3db from average) in the mid-to-high 50’s and no really egregious peaks or dips in the frequency response. Since these speakers will be used for PA, I don’t really care how they might reproduce the 41hz low "E" of a bass guitar, so the projected response in a 2 cu. ft. box is fine with me. In a pinch they could also be used for regular guitar speakers or any other instrument which doesn’t have significant low-bass output.

The program recommends tuning the enclosure to 57.42hz. I can get port tubes in 1-7/8" diameter, and if I use two of them, can fit them on the face of the baffle at the "corners" left by the round woofer. The program tells me ("f3" key in PEBox) that to tune a 2 cu. ft. cabinet to 57.42hz using two 1-7/8" diameter ports, each of them should be about an inch long. The port tubes come in a longer length but I have a hacksaw that will cut plastic easily, so this will all work. I order the parts and/or put them away while I visit the lumber yard.

Box Construction.

Since these are small cabinets and I’m not trying to reproduce heavy bass, " material will do fine if it’s well-braced and the finished box will be relatively light. I will use plywood, not particle board or MDF, for the reasons harped upon at length in my prior articles.  Now I need to fine-tune the sizing before actually firing up a saw.

My horn lenses, now that they're sitting in front of me, are actually 15-3/16" wide. Plus, carpet will wrap around the sides of the enclosure to the edges of the horn, so I’ll add another " to 5/16" for that, and aim for an interior width of 15-1/2". This means the top and bottom plates should be 16-1/2" wide (adding the width of two " thick sides), and the 14" depth I decided on for my overall cabinet depth. The sides will be 21" x 14". The baffle and back, nominally 21" x 15-1/2", although I know I’ll have to trim those a bit later to fit inside the carpet wrap-around (if you’re having this precut at the lumber yard, you can have them reduce the back and baffle dimensions by " in each direction and be pretty darn close). In my experience, if you have the panels precut at the lumber yard with their giant saw, the cuts will be very straight but the dimensions will probably be off a little because they’re in a hurry. If you cut them yourself, any ill-fitting panels will be your own fault. But in either case, you will probably need access to a circular saw for trimming panels to size as the box is assembled.

Here’s the best part: you can get two cabinets out of a single sheet of plywood, with not much wastage:

112cuts.jpg (66501 bytes)

Actually, if you needed to, you could even bump up the height dimension to as much as 24" and still fit within a 4’ x 8’ plywood sheet; I just don’t want to carry weight and volume I don’t need from gig to gig.  Since my box grew 1/2" in width from my first approximations, I rerun the modeling program and learn that my port tubes should be cut a little bit longer.  No problem.

As in my previous articles here and here and here (any of which may have useful tips I've forgotten to mention in this article), I will assemble the box with drywall screws, wood glue and interior bracing. The glue should join both the plywood edges and the bracing, but the screws should go into the bracing only; you can’t put a drywall screw into the edge of a plywood plank successfully. As usual, I will use 1" x 2" bracing (actually " by 1-1/2"). I recess the rear bracing (to which the back will attach) by " from the rear lip of the cabinet – to accommodate the " back-panel thickness plus the thickness of the carpet that will be attached to it. In front, I recess the bracing a full two inches. The baffle will stick out another " or so after it’s carpeted, and I’ll need the resulting 1-1/4" of depth to fit the woofer (with its big front gasket sticking out; I hate to deface the speaker by removing that) and the grille. Here are the raw boxes with top, bottom sides and bracing in place:

12boxes.JPG (68328 bytes)The horn doesn’t need to be recessed as far as the woofer, so there are two ways you can proceed. One is to cut the rectangular hole for the horn into the baffle (it will come very near the edges of the baffle) and let the horn be recessed to baffle depth. Another is to trim the baffle so you’re sure it will fit after the carpet is installed (you’ll need 1/8" of clearance all around, or in other words " of "slop" in both baffle height and width), and then make a "frame" for the horn out of additional bracing strips attached to the front of the baffle to move the horn forward, roughly flush with the cabinet edge. This makes the horn/driver depth (usually the deepest thing mounted in the cabinet) a little less critical – plus, I think it looks better to have the horn exposed and nearly flush with the cabinet front. Here's a shot showing a projecting "horn frame" on a baffle:hornframe.JPG (133916 bytes)

I want to space the woofer at least 1" to 1-1/2" from the bottom of the horn because I will use rubber cabinet feet as "standoffs" to hold the grille away from the face of the woofer. Otherwise I want the woofer as close to the horn as possible, both to maximize the space at the bottom for ports, and to provide clearance for the pole-mount or "top hat" that will be intruding into the cabinet interior from the bottom – I don’t want it trying to occupy the same space as the woofer magnet. Then with a saber saw I measure and cut the hole for the woofer; if it ends up a little small you can always enlarge it, but if it’s too big the woofer mounting bolts won’t have much to grab onto, so do this step slowly and carefully. The ports will need their own baffle holes, located where they won’t contact the bracing strips on assembly and far enough from the corners that there is room for whatever "standoffs" you might use to attach the grille, if you use a full-face metal grille, or far enough from the speaker that there is room for the attachment clips if you use a round "waffle"-type grille. Pull out the speakers and ports and test-arrange them on the baffle before cutting holes; it’s a lot easier to visualize the "fit" with the actual items than to transfer measurements from one to the other. Here’s a rear view of the assembled cabinets, showing how everything fits:itfits.JPG (74262 bytes)

On the cabinet back, I saber-saw the hole for the terminal plate. Various supply houses have terminal plates, connectors, crossover components, etc. for your soldering-and-screwing pleasure, and there are articles all over this site about crossover design and speaker connection methods.

At this point, I rough-mount the baffle, back, and drivers and try to determine the front-rear "balance point" for locating my pole mounts in the cabinet bottom using the "broomstick method".  I described this process in a previous article, and it’s crucial for any pole-mounted cabinet. Only then do I cut the hole for the bottom pole mount. Then I paint the whole thing (the exterior, at least) with flat-black latex paint so that any carpet tears later won’t expose shiny white wood. If you are using recessed side handles (a good idea; makes the cabinet easy to hoist onto a pole), now is the time to figure out the top-to-bottom placement of these so they won’t contact either of the speaker drivers, and then cut the appropriate holes for them.

The carpet-covering process is just as described in that same article, using latex-backed "Ozite-type" black carpet, 3M "77" spray adhesive, a big Arrow stapler with fairly short, 3/8" staples, and the sharpest razor knife or carpet cutter you can find. The perimeter of the box is 77" so you will need at least 2-1/2 yards of carpet for that, plus another 1-1/2 yards for the front and back if the material width is 3 or 4 feet. From a 54" wide roll, you should be able to get the boxes, baffles and backs out of 3 or 3-1/2 linear yards of material.  Install the baffle with a few drywall screws from the bracing (inside) side, making sure your screws aren’t so long that they’ll stick through the carpet in the front. With the smaller bracing, it may help to predrill these holes with a bit slightly smaller than the screw diameter, so the bracing won’t split when the screw goes into it.  But with skinny little drywall screws and soft pine bracing, that's rarely a problem.

At this point I line the cabinets with eggcrate foam. You can use fiberglass insulation or other absorptive material instead, but make sure that nothing loose and fluffy can plug the ports from the rear (If you have a factory-made cabinet that sounds sort of boomy and "honky" at certain frequencies, chances are it has no or inadequate inner sound-absorbing material). If I will be using screws or fasteners visible from the cabinet exterior, I generally stick them into an old cardboard box and give them a coat of black epoxy spray paint (I should have told you this earlier, because it takes a long time to dry. Sorry).

Now it’s just a matter of attaching the parts – corners, handles, pole mount, speakers, handles, terminals, grilles, etc. – wiring the speakers appropriately, and testing the system before installing the cabinet back. It’s easier to install the woofer if you install the handles last, because that gives you a big hole in each side of the cabinet to reach through when threading T-nuts, etc. 95% of the time, when you’re done everything will work fine; if not, check your wiring, crossover and jack connections, and check using another speaker to make sure the amp you’re using is working. Most of the time, speakers are easy to troubleshoot, because they either work or they don’t. If you need to readjust the relative volume of the woofer and tweeter with fixed resistors as described above, now is the time to do it, because it will be a major pain to remove the screwed-on back each time you need to adjust something. recess.JPG (60447 bytes)

I would predrill the holes in the back into the bracing as well to avoid splits. I usually attach the back with drywall screws and black finish washers. Or, you can use fat, dome-headed finish screws of an appropriate size. This cabinet is small enough you can probably pick it up and shake it to make sure you didn’t leave any tools or small animals inside.


The finished product:

complete.jpg (67051 bytes)These speakers aren’t insanely loud and we wouldn’t use them (except maybe as monitors) at a large venue, but for a small room they’re fine. Actually, right after these were completed we were booked on a very small (50-100 people, tops) rooftop of a nightspot in downtown Dallas’ West End – the catch was, there was no elevator and everything had to be hand-carried to the roof. We left all the big, heavy stuff at home and took the smallest amps and lightest equipment we could, so these little 12’s got a workout that night as the main PA speakers, pole-mounted on tripod stands. At the moderate levels we were playing, they worked fine, very clean and without noticeable distortion. With a relatively compact size and for a total investment of around $400, these are another useful bit of equipment in our arsenal of "band stuff".

For about the same price, you can find various factory-made 2-way, 12" cabinets, and most of them in this price range are pretty darn good.  I think these are just a little bit better in terms of sound and driver quality, and I know what's in them and how to fix anything that might go wrong.  For me, the slight edge is worth it, and building speakers keeps me off the streets and out of trouble! 

Publications by the same author: Design and Build Your Own Live - Sound Speakers - A book about How to build your own Pro-Audio Speaker systems - 117 pages of excellent information. ISBN 1-4120-2998-8.


Questions or Comments about this article? Write me.

Questions? Comments? .

Return to Shavano Music Online Home page