By Lars Mytting
31 Dec 1996
Printed from SpeakerBuilding.com, 09 Sep 2010 06:05
URL: http:///content/diy/1006/
Let me point out that I am not a pro speaker builder. This was my fourth design, and I am convinced that it is possible to make improvements to it. The following might be a starting point for others who are planning to use the same drivers. E- mail me if you come up with something spectacular.
After several computer simulations of different crossovers, I decided to try a simple 12 db filter. This called for a low crossover frequency to mute the cone resonances enough. This puts great demand on the tweeter, which must handle lots of signal, also down in the midrange area. Vifa claims that their aluminium dome D25AG35-06 can be used down to 1500 hz, and I decided to let it be the companion.
--- 1,9 mh ----+---------------------- + | 10 uf ScanSpeak 8542-05 | ---------------+---------------------- - ---------------10 uf---+----3.9 ohm--- + | 0,8 mh Vifa D25AG 35-06 | -----------------------+--------------- -
The highpass has about the same values as a 1900 hz Butterworth filter, but the lowpass inductor is made way larger to mute the rise in midrange response caused by the baffle.
I found this crossover interesting because it only needed five parts, all with moderate values. Such "ascetic" crossovers usually is a good thing. Another theoretical advantage was to let the tweeter handle a larger part of the frequency range than usual. A tweeter has a way higher acceleration factor than a midrange/woofer, and the low crossover frequency could then lead to a more open and detailed midrange. This was later confirmed.
On the other side, the filter might not mute the breakup of the woofer cone enough, making its sound hard and ear- straining. Also, if the tweeter did not handle all the midrange signal, it would sound distorted and harsh. Anyhow, scientific value of this crossover was high, and since I love the smell of solder tin in the morning, I gave it a try.
I covered the inside with 10 mm thick pine wood. The boards were 10 cm wide, and fastened to the MDF with standard white glue. I then applied a layer of 1" Akustilux damping mats. (The usual stuff made of synthetic dead sheep)
Pine is a rather "soft" wood, and was chosen because I wanted to test if its different density from MDF would "eat" midrange resonances. MDF alone will get into resonance with midrange signals, especially female voices. I built another pair of similar enclosures, but without wood damping, to use as a reference for the tests.
A favourite track soon became Cowboy Junkies' "The Trinity Sessions", where Margo Timmins's voice was very present and involving, with lots of details and room information. Dynamics and transient response was good, but not exceptional. No really deep bass, but firm, with good control. -3db ca. 50 hz.
The drawback was that it did not have that refined homogenity which I would demand from a true high-end speaker, and some midrange harshness was present. I have no measuring equipment, but listening tests with the woofer alone convinced me that this was not distortion from the tweeter, but woofer cone breakup.
The tweeter did a excellent job even if it had heavy working conditions. Its qualities were confirmed when I later (yes, later..!) read in Stereophile (Jan. 1994) about the Audio Artistry Mozart speaker. It uses a Vifa metal dome crossed at 1800 hz - and thereby achieved "excellent midrange dispersion" (Quote page 107: "Its ability to suspend a wide, spacious soundstage in the front third of my Reference Room is quite remarkable".)
--- 1,9 mh ----+-----------+---------- + | | | 5.6 uf 10 uf | ScanSpeak 8542-05 | 0.3 mh | | ---------------+-----------+---------- - ---------------10 uf---+----3.9 ohm--- + | 0,8 mh Vifa D25AG 35-06 | -----------------------+-------------- -
This makes the slope steeper, and mutes the breakup at 2000 further down. It worked very well. The midrange was "softer", without losing detail. Some records may still sound a bit hard, but this is a result of the relatively loud midrange. It's a design choice. Some other crossovers were also tested, but my friend, which I built the speakers for, insisted on using the 12 db filter. "No, I want that folklorish sound, even if it sounds a bit hard on some records" he said. (Any better sound adjective than "folklorish" is rewarded $5!) He has a older and darker sounding equipment, and plays mostly classical recordings. The speaker cannot be said to be "neutral" (which speaker is?), but brings life into recordings which on his former system sounded dull an uninspiring.
I will suggest that you try different crossovers for yourself, but you should definately try out a low crossover frequency on this Vifa dome.
One quick'n dirty way of testing enclosure vibrations without measuring equipment, is to play different frequencies (signal generator or test CD's) and increase and decrease volume. Hold your hand to the box. Does it vibrate? Does the tone change charachter when played loud? If so, it is probably the "tones" added by the enclosure that you hear. With this, you will get a good indication on how much, and at which frequencies, the enclosure vibration adds to total output.
Back to my two pair of boxes: The differences between MDF/pine and MDF alone were NOT so significant, and I believe that better results may be achieved using another sandwiching method. Subjectively, the MDF/pine box did not sound any different than the MDF box.
The MDF/pine was generally a little bit more dead than MDF alone, and was brought into resonance in a bit smaller frequency range, but I believe this just was a result of added mass and weight to the box. Its heaviest resonances were at 180-200 hz and 400-500 hz, while the box of MDF alone had its heaviest resonances at 300-600 hz. At approx. 200 hz, the MDF/pine had a peak where it vibrated more than the heaviest vibration of the MDF box, but generally, MDF/pine was slightly more dead.
One surprising difference was at exactly 315 hz (You can find that freq. on the Sheffield/XLO Test CD). The MDF box went into heavy vibration, but the MDF/pine box was absolutely dead, no matter how loud the tone was played. All other enclosures that I have tried this on, has started to vibrate at this frequency.
The major reason for not recommending the described sandwiching method is because it is a bit unpredictable, not at least because the 200 hz peak. Maybe it was just because the pine boards went into resonance there, and transplanted the vibration into the MDF. Sandwhiching is no good if the second material used starts to vibrate and sends its vibration into the MDF!
My next experiment will be to fasten the second material with a thick, elastic adeshive, like a thick layer of silicone, to isolate the MDF from it. I don't know which material I will use yet, but it should be something with different density from MDF. Some attention should be sought to keep the sandwiching material from acting like a loose part in the box. This might cause buzzing and severe trouble.
I will share the results on Roy's web page if I end up with something useful. Right now, I'm gonna pop a beer and sit on the porch, playing Lynyrd Skynyrd, and then keep on popping beers until I am sitting in a convertible with Cindy C. Or Helena Christensen, I don't know yet.
Lars