Third Try

Project completed September, 2009.

Three's a charm.

After realizing the little ND20FS-4 tweeter was my main problem (not the driver's fault, mine for trying to use it this way), I decided to upgrade to a favorite among many builders, the SEAS 27TDFC tweeter. This is a good mid-priced 1 1/8" soft-dome tweeter with a successful history.

Then something came along while I was still thinking about my upgrade, Zaph introduced his ZA14W08. That pushed me over the edge, I bought a pair about a week after they became available at Madisound.

I suddenly realized that my desire to build a system that produced accurate, low distortion bass could be extended to to building a system that would produce accurate, low distortion sound across the full spectrum without spending a fortune. Starting with drivers designed to deliver good (i.e., flat) frequency response and good harmonic distortion (HD) performance through the midrange and treble, as well as using the 1/4 ft2 cabinets to hold the little woofer"s QTC at 0.71 for tight, accurate mid-bass response should yield better than good results. Leaving the power hungry/high displacement bass below 100 Hz to the sub rounds the sound out nicely and doesn't place any demands on the little satellite woofers they can't deliver.

I rebuilt the satellites using:

    1. The same pair of 0.25 cubic foot enclosures from Parts Express for about $120, I just replaced the baffles for about $10 each.

    2. A pair of Zaph's ZA14W08 mid-woofers, about $36 each from Madisound.

    3. A pair of SEAS 27TDFC tweeters, about $40 each from Madisound.

    4. Passive crossovers, designed using in-box measurements with SoundEasy.

    5. The speaker terminals and crossover parts were still about $50 for each system from Parts Express.

    6. All drivers were flush-mounted.

This time I started the design with a planed Fc of 2500 Hz. Vertical lobing Shouldn't be a problem.

The 27TADF is an established low distortion performer used by many builders in middle price rage designs. Zaph's design goals for the W14 were to produce a reasonably priced 5" midwoofer with low HD characteristics, good frequency response, good build quality and a distinctive appearance. He succeeded.

The tweeters are tested and reviewed on Zaph's site (tweeters), and his data sheet for the ZA14W08 puts factory spec sheets to shame. Both drivers are exceptional for what they cost. While these speakers are now about $200 each (no longer in the super- budget category), I think the added cost is worth every penny.

Armed with quality drivers that should deliver what I want, I pulled the W5-704s and ND20s (kept them for future projects of course) and plunged into the rebuild.

This is the arrangement I use for the speakers and my monitor. The speaker's position near large surfaces led me to believe that baffle step compensation wouldn't be necessary, but I learned the hard way that it helped tremendously,

The speakers are built in mirrored pairs. Both the woofer and the tweeter are offset horizontally in an effort to help deal with the first couple of response swings after the baffle step frequency and help control diffraction effects. The center to center spacing of the drivers is 4 15/16", suggesting a crossover frequency below 2743 Hz. The actual Fc wound up at about 2200 Hz. I didn't get a good polar plot of the vertical dispersion, but the center lobe is fat enough vertically for my head to never leave it.

This time I was better prepared to use SoundEasy effectively. My test set-up was properly assembled, tested to be working correctly, and actually measuring from 5 Hz to 45 kHz. This crossover design went much smoother. A year of following everything on the SoundEasy user's list, and reading every post I could find by people using SoundEasy was very beneficial.

I was worried about using steel screws so close to the coils when I mounted the crossovers, so I checked the effect of the proximity of a wood screw to a coil in different configurations. Now that I see the picture again, I realize I need to do the same test with driver motors (if they do affect the impedance of the coils, it wasn't enough to cause noticeable problems).

I even built prettier crossovers this time. Went with poly caps too. I'm still not bashful about using electrolytics for larger values at lower XO points, but polys on a woofer/tweeter XO at 2200 Hz aren't very expensive, and they are a more linear device than electrolytics.

Notes on screws around air core inductors:

  1. While it isn't illustrated, the biggest effect is a screw’s head sitting in the center of the coil.

    1. Holding the head of a #6 steel wood screw just inside the coil raised a 0.4 mH (nominal) inductor from 0.415 mH to 0.53 mH.

  2. The arrangement in A caused no measurable change in a 0.4 or 3.3 mH coil

  3. The arrangement in B caused

    1. The 0.415 mH coil to increase to 0.416 mH

    2. The 3.310 mH coil to increase to 3.312 mH

  1. The arrangement in C caused

    1. The 0.415 mH coil to increase to 0.417 mH

    2. the 3.310 mH coil to increase to 3.315 mH

With the exception of case #1, none of the other arrangements change the inductance of the coils enough to matter.

As near field monitors, they do quite well. They are very accurate, and I haven't heard any distortion under listening conditions below 100 dB. The lower midrange and upper bass start to get muddy with bass heavy passages, but I suspect it's the amps in the Logitech system. Listening to Caislean Õir (from Macalla, by Clannad) they'll play to 105 dB with no problems. The system will play Seven Bridges Road by the Eagles with peaks above 108 dB cleanly.

I like these two songs for testing. They're both recorded well and neither has any heavy bass content below 100 Hz, but both press a system very hard from 100 Hz to 6000 Hz when played at loud levels. If the vocals stay clear, the speaker's dong its job well. Caislean Õir is especially tough with its low pitched male chorus placing a lot of its energy from 100 Hz to 300 Hz range. Seven Bridges Road pushes a speaker system very hard from the low midrange into the treble range.

They reproduce delicate treble with better detail than anything I've ever owned, yet Rickey Lee Jones doesn't faze them with her hot 12 string and sometimes explosive vocals. They play Baba O-Riley like I remember The Who playing it in Denver (back in 1972). These things just work, really well. They also sound great from across the room when I relax in my easy chair. I've been using them for six months now, and I still like what I hear.

The details for the baffle layout, cutouts, and the crossover (including the values for the components) are in the following Web Album:

http://picasaweb.google.com/lhwidget1/ZaphZA14W8sAndSeas27TDFCsIn025FtDaytonEnclosures#

I've included most of my design and final system measurements from SoundEasy also.

Jay R. Taylor

March, 2010

Home