 |
     |
 |
|
|||
|
The most serious con is the need for some sort of equaliser to account for a 6dB/oct drop in frequency response due to acoustic shorting of the waves coming out from the front and back. So what is the dipole radiation? Well, since the speakers don't have a box the sound is radiated both in front and in back of the speaker. The waves to the back are radiated in opposite phase whiche means that the response to the sides (where they meet) drop conciderably. This reduces the standing waves and room interaction that can appear in a room which means a neutral bass reproduction and detailed midrange. Although the sound is "beamed" at you the beam is larger than monopole speakers which results in seamless integration between the speakers. The result is that the speaker more or less dissappears and instead only a seamless wall of the most natural sound is heard together with superb stereo-imaging. This is quite different from a normal speaker where the bass is radiated in 360 degrees. But, as the wavelength gets closer to the width of the baffel the sound is only projected to the front of the speaker with an increase in response as a result. This baffle step can be up to 10dB and there are various techniques to remedy it (cut-off edges, omni-tweeters etc.). Even if you adress it in some way it will still be there along with the coupling of bass to room renonances making a box speaker inferior to an open-baffle. The other major benefit of making it this way is the use of bi-amplification (actually tri-amping) and active crossovers. Take a look at Rod Elliot's article "Benefits of bi-amlification - not quite magic but close". Basically the biggest benefits is that the crossover can be made totally phase-coherent and that the signal is not influenced by physical properties in the crossover components such as temperature, resistance, inductance and so on. By using separate amplifiers (a must in active x-overs, option in some passive ones) the problems an amp can encounter (clipping, heavy loads and so on) is isolated and does not affact the other drivers. The voice-coil is a very easy load for any amplifier. Furthermore there's no stored energy in the box at all, which is always a goal when designing enclosures. Some manufacturers use closed boxes as a first way to deal with it (the reflex port is simply a renonance pipe) and goes to extremes in order of internal bracing and exotic materials. This is what really separates ultra-highend from quality speakers since there's often the same elements in them.
I'm designing and building my own version of a Linkwitz style dipole using my old drivers and the one's bought for my concrete speaker project. This gives me a little less maximum SPL than his designs, it's around 102dB (at 1m) for the midrange and 94dB at 35Hz for the woofer which works for me since I live quite compact. It's not recommended to drive it lower than around 40Hz when only using 8" drivers since the cone excursion demands will be to big and they'll bottom out. According to Linkwitz (read why here) the optimal setup for surround sound is a 2/2/2.2 setup or 2/2.1 as an alternative if there's no place/budget for a eight speakers setup. The point is to skip the center channel since it will do more bad then good. For a uniform sound the speakers need to be identical and it's not easy to place the center at an optimum position (behind a totally acoustically transparent screen). So I'm skipping it and mixing it into the front channels. So, my first version were built like this:
The paper cone on the midrange is extremely light which makes the driver fast (less stored energy) and the sound very transparent and detailed. The non-linear distorsion is also low which makes it a better choice over the Scan-Speak 18W8545 when the bass capabilities are not being used. The tweeter is Seas cheapest high-end tweeter that gives very much bang for buck. The baffles are made of 20mm MDF painted grey. I made curved sides and a bottom to make it more rigid. Two surround speakers with MGR elements, these will have the same kind of sonic character as the front channels but will not have the same kind of detailing and openness due to lower quality elements:
All of them have bass redirection below 125Hz to 2 H-baffle woofers with 2 8" drivers each. Each speaker needs equalisation and crossover. This is done with active filter modules with an integrated amplifier. This is how the signal flow looks on one of the front channels:
The next step is to make them as free standing panels. The front speakers will have dual 8" bass drivers and a 6.5" mid while the surrounds will have a single 6.5" woofer. This will hopefully be made during the spring. Below is a concept image of one of the main speakers. Here's a cad drawing of a front speaker.
Related pages: Related articles: Linkwitz comments surround sound requirements Rod Elliot's "bi-amplification- not quite magic, but close" Links of interest: Linkwitz Lab: One of the best places for loudspeaker theory |
||||
| © Anders Zakrisson 2003. Please contact me if you have any questions |
||||
Even
though I recently settled on making two way monitors I changed my mind
after I heard my dipole prototypes, I'll build open baffle loudspeakers
instead! If you need to be convinced why open baffle is the way to go
take a look at Sigfried Linkwitz's 
Take
a look at Burmesters range-topping 
The
company that has taken the dipole approach the furthest is 
