"What can I say? I'm speechless! These speakers are awesome in every sense of the word! The cabinet design suits my furniture perfectly and is so smooth and beautifully finished. The sound is pure and clean. With a choral director and a drummer in the household, we listen to a wide variety of music styles and each one is heard with precise nuance and clarity"

-Terri Ashbaugh on her Topaz Towers

(Terri's Towers shown in Satin Pecan)

Original Concept

The Topaz was my first venture into a measured design using JustMLS and LSPCad. I wanted an economical sealed monitor with a small facade that could be used as L-R-C in a home theater or as part of a multimedia setup. The original customer for this system wanted a vertically rounded baffle with recessed drivers. (Not something worth doing with my Onyx drivers) The PE 299-712 clearance driver looked like a good candidate.

I've never been a fan of horizontal MTM center channels, so I needed a small footprint tweeter that would allow me to build a vertical, TM center channel. The Audax TMF0251 looked like a good choice for performance, small footprint, and ease of flush mounting. Since this tweeter also comes with a truncated flange, I plan on using that version for the center.

It took me awhile to iron out the kinks with JustMLS. (I never could get it to measure in two channel mode). Once I did my homework, cursed a bit, read some more, tried stuff etc. I finally got the measurements working. I setup at about 39 inches with my mike directly on axis with the tweeter. I took measurements for each driver, and for the two drivers wired in parallel. I also used JustMLS to measure impedance curves. After loading the measurements into LspCad, I verified that the combined curve looked very similar to my measured, two-driver curves by adjusting the position of the drivers. (John Kreskov has an excellent tutorial at his website http://www.geocities.com/kreskovs/John1.html) The design process took a few hours, a prototype Xover, some more measurements, and some more diddling. The "Def" woofer did not need a zobel, nor did a second order crossover improve it's response. The tweeter has a definite rise with increasing frequency. After some experimentation, I found that a 0.1 mH coil flattened the Audax response nicely.

Enhanced Xover Modifications (Added 2 October 2002)

The sealed version of the Topaz is now my office monitor. I've spent quite a bit of time listening to them. My wife was very happy with her Beryl towers but thought the Topaz was a bit better. She's asked me to build a pair of Topaz towers for her choir director. (Now we'll see if I'm any good as a designer! This woman can pick out a flat alto from 20 singers) The Topaz Tower is 22 liters. 8.5W x 7.25D x 36H (external) tuned to about 49 hz with two 1.5" x 4" ports (PE# 260-402)

I used this opportunity to go back and revisit the Topaz design. I wanted to smooth out the impedance a bit, and see if I could improve the imaging a bit. After a few hours of dinking I came up with this revision.

This version ups the Xover to 2600, and shapes the tweeter response a bit better. It also flattens the overall impedance response. The change in listening is subtle. It improves the timbre of vocals and adds a bit of depth to the soundstage. This version gently drops response about 1 dB in the 2000-3000 range. LspCad models also showed improved phase tracking.

I used a .67 mH coil in the final version but a .7 would be fine. Values of 3.7 to 4.3 on the tweeter "zoebel" do not change the response drastically.

Oct 2, 2002 - I ran a quick measurement on the tower version. This is a merged near + far measurement taken out on the driveway on a windy day. I did not add in port effects

Revised Xover for Enhanced Baffle Step (Added 17 July)

After trading emails with Darren K at PE and extended listening, I decided that the orginal crossover was just too thin for mains. (Though I may use it for the center channel). I spent some time with LspCad and came up with a revised crossover that ups the baffle step compensation to about 5 db.

I had a couple of 14 ga 1.2 mh coils in inventory so I used them as a basis for the woofer. The major changes on the tweeter network included increasing the cap to 11.7 mf and revising the LPad. This effectively reduced the crossover frequency to about 2400 hz.

I was afraid that these changes would loose that wonderful detail I heard with the initial design. The revised version pretty much mitigated most of the thin bottom end but left the high end detail I liked. The mids are now a bit more laid back, and I picked up a hint of sibilance. Overall I am now quite pleased with the performance. These are very capable standalone monitors. Bass heads will probably want a sub. The Topaz should be a fine HT monitor.

I believe that this Audax tweeter is the real deal. Very crisp detail, lots of air. Lovely on vocals and wind instruments. Very clear vocals, virtually no sibilance, excellent image and soundstage. Decent depth and placement.

July 28, 2002 - While I had my measuring gear set up, I did a ground plane measurement of the final sealed speaker with the BSC xover. The plot below is 1/3 octave smoothed. No nearfield merge was employed so I'm not sure about the bottom end. There's more bass in a real room with room gain.

Construction

The construction is fairly straight forward. 3/4 MDF with one small piece of bracing glued behind the baffle between the drivers. I use biscuits and clamps exclusively. After sanding the boxes square, I used my 3/4" radius bit to roundover the two front edges.

I used leftover 10 mil paperbacked cherry veneer with solvent based contact cement. I first veneered the back, then the wraparound front (always start from the center and wrap each side individually) I place the box on a 1/4" high scrap to set the veneer overhang. I allow 1/4" extra top and bottom with 2" extra along the wrap. For paperback, I trim the edges by hand with a utility knife. The router sometimes splinters the veneer. (See my Veneering Tutorial)

The finish was an experiment. I mixed a few drops TransTint Red-Brown dye stain with a 1 lb cut of dewaxed shellac and used that for a sealer. Followed by four light coats of Fuhr 355 waterbased Acrylic with my Walcom STM spray gun. I added a couple of drops of Trans Tint Amber to the Fuhr. It didn't completely eliminate the Cherry "splotching" but it did warm up the wood nicely. I've pretty much retired my Deft now. I left all the veneer intact when finishing. I then used my flush trim bit to trim the holes. Followed by my rabbeting bit to finish the recesses. I use closed cell self-adhesive window insulation (3/8 x 1/4) to seal the drivers with black #6 wood screws. The cabinets got dinged while I assembled. I'll probably spray them again.

The square cross section is far from optimal for internal resonance. To compensate, I lined the inside with acoustic foam and filled the rest lightly with Dacron. Parts list below for one speaker. If you choose to use PE for all parts, I would choose the 1.3 18ga (266-830) for the woofer coil. I buy from both.

Original Crossover (Minimal Baffle Step)

The original crossover looks a bit strange but it works out to about 2800 hz. I tried to work things around to use more or less "standard" values. (You can combine caps 5.1 + 5.6 = 10.7) Overall, it's a fairly economical crossover. You should be able to put the drivers & crossover together for less than $70 a speaker. Overall, with this xover, the bass is skimpy. For HT with a sub, this design would be more suitable for wall, or in-cabinet mounting.

The 0.8 mH coil provides about 2db of baffle step. If you are stand mounting the Topaz, my simulation suggests you can bump the coil to 1.0 mH for about 4.5 db of baffle step without drastically changing response.

The 3.5 ohm resistor could be replaced with a 4 ohm if you prefer your sound a bit more laid back. The large lug on the tweeter should go to ground.

The measured response curve below was taken on a stand about 38" off the ground in my driveway. I used JustMLS to merge farfield and nearfield measurements to obtain the curves below. The absolute dB scale is only an estimate. Some of the roughness is likely due to the surface mounting of the drivers. I also need to work with internal cabinet damping to cut down somewhat on internal reflections. My initial reaction: this is an articulate little monitor, decent bass. (Nearfield confirms F3 of about 89 db) Lots of high end "air" and definition. Solid mids.

Dec 4, 2002 MTM Crossover

Several people have asked me about an MTM version of the Topaz. I have not built one, but LspCad suggests this should work. Keep the same 8.5" baffle and same spacing of drivers. (5 inches from tweeter center to woofer center) You may have to adjust the tweeter padding.
I'd suggest a 50 liter box tuned with a single 4" port, 5" long.