B&C 15CXN88
15" Coaxial Two Way Design
Embedded Files
Design description and thoughts:
First I'd like to thank Mr. Delezene for lending me the B&C driver (for over a year) to finish this design.
The B&C 15CXN88 is a high quality pro sound coaxial driver with excellent dispersion behavior capable of extremely high output, very low distortion and essentially imperceptible levels of compression when used in a home setting.
The goal was to use this coaxial and have a high quality - high sensitivity speaker voiced for home use that could play down to 80-100Hz for crossover to subs. While you only need a few watts to rock the house you can still throw hundreds of watts into the speaker and performs without a hint of strain or struggle.
When I finally had the crossover dialed in I found the sound quality from this coaxial to be excellent, very clean midrange, the treble is clear and detailed without any coloration and the overall balance worked well with everything played through it. There is almost nothing I can fault the driver for and it will play as loud as you even need in a home so certainly no qualms if you are thinking about using these for a high end system capable of extreme output when used with equally capable subs.
As with most designs using a horn loaded tweeter the dispersion is tighter in the top octaves then a more conventional dome tweeter so I will say you can expect a more forward or out in front presentation of the soundstage with less ambient spill. Some may prefer this more direct type of sound signature but others may not.
Enclosure Design:
The intended use was to be with subs so in the end a rather compact sealed enclosure of ~1.5cuft was chosen over a slightly larger ported cabinet as it appeared it should provide just enough extension to use with an 80-100hz crossover and the sealed design has the advantage in midrange quality by eliminating potential port resonance and midrange leakage issues.
The cabinet dimensions I ended up building it out as were 20" wide x 23" high x 9.5" deep using 3/4" material which gives a net internal volume around 1.5cuft. I used a double baffle sandwiching 3/4" and 1/2" but it would likely be easier to do a double 3/4" baffle and make the cabinet 10" deep. Optionally you could extend the depth to 12" deep may be beneficial to squeeze out a touch more bass extension from the driver.
The driver should be flush mounted. I used 5/8" roundovers on the cabinet but those are not critical. Backside of the double baffle was chamfered around the driver cutout.
Damping; The sealed cabinet should be fully stuffed with damping material, I used 1.5" recycled denim damping to line the walls then stuffed the rest of the cabinet with polyfil because that is what I had on hand. Filling with fiberglass or rockwool would also work fine.
Triangle braces at the top/bottom are the cutouts from the middle braces. You will want to make sure one set is spaced just over 10" apart if you want to make mounting the Crossover PCB easier so it can fit between them. The crossover can then be mounted to either the back or bottom side.
Cabinet Cut List (20" x 23" x 10", double 3/4" baffle):
Baffle(x2)/Back - 20" x 23"
Sides - 23" x 7.75"
Top/bottom - 18.5" x 7.75"
Braces - (2x) 18.5" x 7.75"
Brace Dimensions:
Crossover Design:
The 15CXN88 woofer has a sizable dip in its response just under 1000Hz so that put an upper limit on the placement of the crossover frequency. Fortunately the large 3" diaphragm on the HF driver has no trouble reaching that low and I landed on an electrically asymmetric crossover just under 900Hz without having to force things from either driver.
The woofer filter is a fairly straightforward, a 4th order lowpass with a bit of damping on the first parallel leg from R6 was able to knock things in line quite well taking care of both BSC and general response shaping. Originally I started the design intending to use I-core inductors on the woofer but switched to 18 gauge air cores for two reasons. First the woofer distortion was exceptionally low so I wanted to use air cores that would not add additional distortion from magnetic hysteresis to that. Second the added DCR from the 18 gauge air cores actually helps extend the f3 because you effectively get a bump in output around the woofer/cabinet resonance where the spike in impedance prevents as much loss as you see above it.
The high frequency side required a bit more work to get the response shaped correctly as is the case for most coaxials since they often measure quite rough due to all the symmetric sources of diffraction along the HF path. The highpass is a simple 3rd order design but there are four LCR filters to help with response shaping. The first filter in series (R1/L2/C3) does most of the heavy lifting in general shaping pulling both the top end and knee of the crossover down some. R3/L3/C4 is there for impedance compensation flattening the impedance peak around 1500Hz so the highpass filter functions correctly. R4/L4/C5 knocks down a little bit of peaking from the driver between 3-4khz. R5/L5/C6 tackles the HF breakup peak around 16khz which was quite high on the driver I had for testing. Finally R2 is there to provide a just a bit of extra padding to bring the level down where is needs to be to match the woofer.
Since coaxials will often measure rougher on axis then off axis I normally shoot for a response shape that would be fairly level on axis without any major peaks while optimizing more so for smoothness of the listening window and general trend/shape of the sound power as I find this provides the best overall sound balance. The 15CXN88 was better behaved then most coaxials in that I did not need to sacrifice the on-axis response much at all to balance things with the off axis response.
-Optional crossover PCB I developed for this B&C 15CXN88 design-
Suitable Crossover Parts that fit the PCB:
C1(4uF) - C3(12uF) - C3(1.5uF) - C4(18uF) - C5(1.2uF) - C6(0.22uF) - C7(25uF) - C8(22uF)
L1(1.5mH) - L2(0.40mH) - L3(0.70mH) - L4(1.2mH) - L5(0.40mH) - L6(2.5mH) - L7(3.5mH)
R1 (5 Ohm) - R2 (25 Ohm) - R3 (9 Ohm) - R4 (12 Ohm) - R5 (2 Ohm) - R6(2.4 Ohm)
Crossover sim with individual driver responses and reverse polarity notch:
Crossover Filter Slopes:
B&C 15CXN88 Crossover Schematic:
All capacitors are designed to be standard metalized polypropylene.
For the inductors the 2.5mH and 3.5mH should be 18 gauge (1.0mm) air cores. The remaining inductors on the HF side should be 20 gauge (0.8mm)
Resistors should be 10w, standard wirewound are suitable.
In my testing the 10w resistors are enough for reference level playback in even very large home theaters with power levels up to 300-400w. If you wish to use the design for extremely high output levels beyond that continuously, you can use a pair of 10 ohm resistors in parallel for R1 and a pair of 5 ohm resistors in parallel for R6 as those two resistors dissipate the most power.
B&C 15CXN88 Design Measurements - Passive Crossover
Measurements taken 4pi at 2.83v/2m and scaled up 6dB to approximate 2.83v/1m (Nominal 1 watt / 1 meter)
Measurements gated at 14ms and blended to diffraction adjusted nearfield woofer response below 250hz.
1/24th Octave Smoothing Applied.
On Axis SPL & 10 Degrees Off Axis
CTA-2034 Style Spin Data
Harmonic Distortion @ 85, 95, 100, 105, 110 & 115dB/1m - (Measured at 50cm)
Very low distortion levels as expected for a large high quality pro sound coaxial, only just crosses 1% as you begin to exceed 110dB and remains impressively low in the 200-500hz range.
Compression at 85/95/100/105/110/115dB normalized against 75dB:
The 15CXN88 has exceptionally low compression, this test which covers four orders of magnitude with the 115dB sweep being 1000x the power level of the 75dB sweep shows just a touch over 0.5dB of compression at its worst.
Impedance minimum is just a tad under 7 Ohms, should be a very easy load to drive for most amplifiers especially considering its high sensitivity.
Page updated
Google Sites
Report abuse