Background

When I had the initial idea for this project, my speaker projects had involved "kits", i.e., someone else's design, or crossovers that I had built using standard crossover formulae.  What I noticed was that kits tended toward very expensive components and standard crossover values rarely sounded anywhere near good enough to justify using them with expensive drivers.

Enter the Dayton 2-Way Bookshelf Project (aka the Budget Bookshelf Project). I had heard some great things about the Dayton 1 1/8" silk dome tweeter and that it compared very favorable with a Morel tweeter costing around $45. The Dayton tweeter cost $15.50 making me want to try it out. If it tanked, so what?

But what woofer to match it up with?  Then the project idea hit me... Set a maximum price limit of $100 for the speakers (without enclosures) with good quality crossover components.  All things considered, I decided to go with the Dayton 295-300 5 1/4" woofer.  In retrospect, I think I made a pretty good choice, although woofers in this price range aren't all that impressive. The 295-300 has excellent bass extension for such a small driver. Also, it's very clean bass.  The midrange is adequate (although commendable for a woofer in this price-range), although the rolloff of the driver isn't exactly picturesque.  But this driver costs a whopping $13.50.  I think for the prices we're talking here, we have to put all things in perspective and wait for the final product before we judge any of the parts.

My feeling about a loudspeaker is that it must be designed for a room and with a purpose. You don't shop randomly, right? No, you go to the store with a particular amplifier, room, and placement in mind before you shop.  You might decide you want a tower or a bookshelf, but at least you know before you start buying. Same thing with a loudspeaker... and in this case I wanted to build a small bookshelf speaker that would sound good in a small room (in other words, plan on a fair amount of room gain and sacrificing efficiency for clean, accurate sound.) Finally, to meet the spousal 'eye-roll factor' (S.E.R.F.) I decided that a friend of mine would receive these speakers as a christmas gift. A hundred bucks was just about the right price for this gift. Now I had to deliver.

Crossover Design

Since the time I published the original write-up for this project, I redesigned the crossover to account for changes in the Dayton 275-070 tweeter (it's since been changed back to have properties that are essentially identical to the original version).  Although both crossovers are included with this revised write-up, please use the later version, unless you are simply experimenting.  The objective for the revised crossover was to improve the overall quality over the midrange performance, using measurements obtained with the CLIO measurement system and Calsod Pro 3.10g -- a loudspeaker simulator.

The first part of any design is to break-in and measure the woofers.  Although some will dispute the value of breaking-in woofers, I believe that heating up the voice coil and exercising the suspension, is a good idea.  After all, the glues and other components in the driver may be a bit "stiff" when it first comes out of the box, so why not run the driver with a low-frequency signal for 24 hours, or so, to make sure that the measured parameters represent the performance of the speaker? 

If you follow the various message boards for amateur speaker enthusiasts, you may have read how certain amateur designers believe that small-signal analysis is not a worthwhile measurement of the driver's performance and that large-signal analysis should be used.  Don't believe everything you read on! While these folks do have a good point, in that large signal analysis is useful, it does not supplant the use of small-signal parameters for enclosure design -- and advanced modelling software, such as SoundEasy (which I use to model things like non-linear distortion, power handling, port noise, etc.) take large-signal performance into account, without discounting the value of industry-accepted small-signal parameters. After all, who should you and I believe?  World-renowned speaker designers who have studied these things as a career and had their finding published in peer-review journals and accepted to such an extent that Thiele-small parameters are now the industry standard?  Or some guy who posts out whatever sounds good to him, at the time, on some obscure message board?  Worse yet, some of the people touting these large signal parameters don't even design enclosures -- check their designs to make sure that everything isn't designed around a "Madisound Woodstyle Enclosure" before you put much stock in what they say.  As for me, I trust the people who have stated their findings to the Audio Engineering Society and had those findings accepted by the world-wide speaker design and manufacturing industry.  On top of that, I've built hundreds of enclosures based on small-signal analysis and it hasn't let me down yet!

Small-signal parameter of the Dayton 295-300 woofer resulted in a .48 ft^3 vented enclosure tuned to 47.5hz.  I then mounted the woofer in the box and began testing the frequency response both on-axis and off-axis. The off-axis -3db point for the driver was around 4khz, as determined by a plot of the relative rolloff rates at 0 degrees and 30 degrees off-axis.

Although you don't see them often, a plot of the relative rolloff rates is a great tool for determining your optimum crossover frequency.  There are competing factors in making this determination.  On one hand, we want this frequency to be as high as possible, to make tweeter selection easier and to avoid using a crossover in the portion of the audible spectrum where human hearing is most sensitive (in addition to various other, far more technical, factors).  On the other hand, the off-axis performance of the speaker (which determines, in part, how well the speaker produces a sonic image) is improved by using a lower crossover frequency.  By comparing the rolloff rates of the on-axis response (actually an average of the response of the driver in a +/-15 degree window) and the 30 degree off-axis response, lets us determine where the woofer's off-axis response will be off by about 3 db (in this case, a hair under 4khz) -- this is generally considered to be the optimum crossover frequency for balancing the competing factors.

Initial testing showed that the bass was clean, tight, and as deep as I've ever heard from such an inexpensive woofer; and the high end was impressive. When you open the box, the 275-070 has the appearance and feel of a quality tweeter. When you put it in a box, it has the sound of one.  This definitely gets my 5-star award for one of the best values in DIY speakers. Accordingly, the high end in my enclosure was surprisingly crisp and accurate.

After evaluating the impedance response of the woofer, I decided to add a zobel filter, to flatten the impedance profile of this woofer, due to its fairly high voice coil inductance.  This helps to improve the crossover's ability to control the rolloff of the woofer and to adhere to the target 2nd-order Linkwitz-Riley acoustic slope.  The tweeter's impedance response was measured, as well, as this data is required for use in crossover modeling in Calsod or SoundEasy -- both of which are excellent and powerful loudspeaker system simulators.

After simulating the design of the system, building various test crossovers for critical listening, and shaking a magic "happy speaker voodoo doll" in the general direction of my CD-player, I decided on the previously mentioned 2nd order rolloff for the woofer and 4th order rolloff for the tweeter, with an in-phase crossover at 2500hz.  Although this is slightly lower than necessary for this woofer, I decided to cross a bit lower, due to the low-order acoustic target slope I chose for the woofer.  More importantly, this sounded better than crossover frequencies of 3khz or higher.  As Vance Dickason suggests in his excellent book "Loudspeaker Recipes: Vol. 1", mixed order filters of this nature can help compensate for the offset between the acoustic centers of the woofer and tweeter.  In this instance, there was about a 1" offset and between the use of mixed-order filters and computer-optimizations of the overall system response, excellent overall system response and impedance was achieved.

The Enclosure

Internal dimensions: .48 cubic feet, W: 7.5", H: 12", D: 9.5" (this over-volumes the cabinet slightly to account for the port and speaker volume, as well as a 1/2" birch plywood 'h-brace' mounted horizontally between the woofer and tweeter to support the side f/r baffles. All construction is 3/4" mdf with 1/2" quarter round edges. The front is 4 coats of hi-gloss black enamel spray paint. The edge veneer is 3/4" white-birch banding and the top, sides, and back are red-oak veneer. All wood surfaces have 6 coats of minwax wood-oil. The vent is center-mounted on the rear of the enclosure and measures 1.5" diameter x 2.75" long. The crossover is mounted on the bottom of the inside of the enclosure and the top, rear, and sides have egg-crate acoustical foam lining.  You, of course, can design your own enclosures to taste.  I suggest, however, you maintain the suggested internal volume and front baffle width (although users have reported excellent results using baffles as wide as 9", so that they could reduce the height of the box for aesthetic reasons.  I once built a set of these for a friend with a 9" wide baffle and the system measured well and sounded great).

Final Thought

When you design a project like this, it's hard to separate your ears from your mind -- you want it to sound good and every little bump and dip in the frequency response graphs seems to weigh on your mind. Going by the opinions of the people who have listened to the speaker, I'd have to say it was a success. I built the original DBP as a christmas gift for a friend and my wife said she was sorry to see them leave the house -- she wants a set too! My 12 year old son thought they sounded great... and at least two visiting neighbors (who didn't know I had designed and built the speakers) inquired about where I got them and how much they cost.  Also, at the recent "DIY DC 2002" event, where many DIYers gather to show-off their creations, the DBP received many positive comments -- and that's one picky bunch to please!

---