The Triune
An All Dayton LCR Speaker
Click on drawings for a larger image
Woofer network
Tweeter Network
The crossover topology:
I am a firm believer in the KISS principal in crossover design.
That and the budget nature of this design resulted in a very
simple crossover network utilizing only 6 components. The
woofer net is a series inductor and shunt cap with the Q
limited by a resistor. This results in an electrical filter gain of
-8 dB / octave, and when summed with the driver response,
provides an average acoustic response of approximately 3rd
order BW. The tweeter uses a series cap to derive the 2nd
order BW acoustic response, and a series inductor to slightly
roll off the extreme high end at 2 dB / octave.

Parts notes:
All caps and resistors are specified as Parts Express parts,
but certainly you can substitute your favorite brand. The
inductors specified are Jantzen 18 gauge air core.

Listening impressions:
While I found the prototype crossover design quite acceptable
and perhaps even preferable for a center channel speaker,
after extensive listening I found for music reproduction it was
slightly too forward for my tastes. Looking at the responses, I
suspected the shelf in the stop band might have been
responsible. I modified the crossover to move the shelf an
additional 5 dB down, and mitigated the hump at 1400 Hz. The
crossover differences are small, with only the value of the
inductor and resistor in the woofer network changed, and the
addition of a resistor in the tweeter network. The new
crossover has much more natural voicing with excellent detail,
and even better off axis response. I'll have to say I am quite
impressed with the capabilities of these drivers, and the
sealed system response. I believe the word I used when I first
heard these with the new crossover was "WOW!".

A note about the design process:
The crossover for this project was designed utilizing the CLIO
files as provided by Parts Express. These files were modified
using programs available at the
FRD Consortium to simulate
the diffraction effects of the baffle, and were also modified
using the Hilbert transform to derive minimum phase. These
modified files were then imported to LspCAD for the design.
While I did not use actual measurements in this design, I have
confidence in the validity of these modified files, and of the
design.

Suggested enclosure designs:
While the original design as pictured is sealed, Eric Y has
came up with a vented design for those who want the added
low end extension a vented design offers. The slot ports only
slightly add the the baffle width and calculated to provide an
f3 of 47 Hz. compared to the f3 of 90 Hz of the sealed
design.The deep enclosure would be ideal for undermonitor
placemement, but could be used above the monitor as well

Stuffing suggestions:
I stuffed the sealed version with R11 fiberglass, although I
would think poly fill or similar would work as well. I filled
roughtly the back 75% of the enclosure, leaving room for the
drivers to 'breathe'
Eric lined the back, sides and top on his vented version with
eggshell crate acoustic foam with good results.

Tweaks:
Thanks to a fellow PE poster, it has come to my attention that
in some situations there may be too much baffle step
correction in the Triune when used as a center channel. This
would likely be cause by adjacent surfaces extending the
baffle size.  Since this was to be a L/R/C design, I was
somewhat in a quandary as how to voice the Triune and
ended up voicing it as a L/R main. His solution was to remove
the 1-ohm resistor in the tweeter network, as I had done in the
TL version of the design. If you find the tweeter a bit recessed
in your situation, this should be the solution.


Copyright 2004 Curt Campbell
Parts List
Vented Enclosure Design
Courtesy of Eric Y
Sealed Enclosure Design
Courtesy of Sandy H
1 < Page > 0