loudspeaker sensitivity

[rv1890]

I have a question. I am planning on building a speaker cabinet that is going to house a 8" woofer, a 5-1/4" mid, and a 2" tweeter. When Im choosing my loudspeakers should they all have the same sensitivity? I am going for sound quality. Also does anybody know what the ideal way is to choose speakers for a cabinet?, or where I can read about it. Im having difficulty when it comes to choosing the 3 speaker sizes i need. What specs should i look for so that all 3 will sound good together. Any specific frequency ranges all 3 should have ?.....thanks i would appreciate any help.

[Tim Moorman]

Good questions. I would look at the overall goal first, as far as cabinet size and bandwidth needed, as that is dictated primarily by the bass unit - 8" woofer, in this case. Then at the mid you desire for the proper match in efficiency and bandwidth. The woofer efficiency may be slightly higher to somewhat offset baffle step losses that will occur. Often, two drivers are used in a floor standing speaker to gain output here, and match better to the mid efficiency. The impedance will be halved with twin bass and parallel wiring. There will also be a small amount of gain, a dB or so, passed on to the mid, depending on how they match up, so it is not uncommon for the mid to be down a dB or so less in efficiency. Look at Peerless, Vifa, for a quality 8". Scan Speak or Seas may entail more money, and difficulty with the crossover is understood with the metal cone Seas, but the quality is there. Sonically, very good.

Crossover points are determined by the behavior of the driver in the intended passband to be used. Look at an 8" woofers output from 1000 Hz on up, and most will start to get non-linear at some point above. You want to avoid these regions effecting the sound, and you should stay an octave below these at the crossover. So, if the woofer is starting to break up above 1k Hz, go no higher than 500 Hz for the crossover. If the degree of break-up is large, you will then need to use steeper crossover slopes to keep this from being audible in the final design. A notch filter may also required on particularily bad resonances(metal cones), even if it is out of the intended passband.

Dispersion of the individual drivers needs to be the same at the crossover region when designing a multi-way. If you look at the response of the 8" woofer, the 30 degree off axis response will usually be well below that of the on axis measurement at higher frequencies. A lower crossover point is needed to avoid drop outs off axis at these higher frequencies.

When choosing the mid, look at the Fs of the driver. If you intend to cross over at 500 Hz, as in our example, then the mid unit should be an octave above its Fs at the crossover. So, 250 Hz Fs and below will be suitable for the 500 Hz crossover, so long as the driver's response is smooth at that point. You may need to build a separate enclosure to house the mid, if not in a sealed rear chamber, to prevent negative interaction from the woofer.

You mentioned a 5 1/2" mid, which may be more than you need. More of a midbass really. I would look at the Seas 4 1/2" MCA 11 FC, the Morel MDM 55 2" dome mid, which is quite good but expensive, and pair these up with the small Vifa XT 19(4 ohm) above a 3K crossover or higher, or one of the Seas 27 (TTFC ?) models. A 2" tweeter will serve no purpose when crossed in above a mid range, unless it has some particular quality you were after.

Tim

[pinkmouse]

It's also worth checking the phase of the drivers at the crossover point, and making sure that it is not changing too much.

[wintermute]

Originally posted by rv1890
Also does anybody know what the ideal way is to choose speakers for a cabinet?

You don't choose speakers for a cabinet as such.... you may want to limit yourself to a particular cabinet size, in which case yes you need to choose a woofer that doesn't require a cabinet bigger than X litres. But once you have decided on a driver you must design the cabinet to suit the driver. (simulation software such as WindISD pro http://www.linearteam.dk/

or Unibox (requires microsoft excell) http://www.pvconsultants.com/audio/boxmodel/unibox.htm Will help in this regard)

often the spec sheets have recommended box volumes too, so you may be able to discount a particular driver early if it looks unsuitable....

find the specs of drivers on your shortlist, and put them into one of the above simulations . You will need to decide on an allignment too (sealed, vented, TL, etc) This will/may affect your driver decisions too (or the other way around).

Good starting point for driver research is http://ldsg.snippets.org/idx.php3

Tony.

edit: note that the linked to simutaltion software won't model TL (Transmission Line) speakers.

[simon5]

What is baffle step losses?

[Tim Moorman]

Bass is said to be omnidirectional below 100 Hz, although I don't subscribe to this entirely. That is, when you play music or tones at say 60 or 70 Hz, the sound does not eminate from the cabinet in a beam, like a wide focus flash light, but rather wraps completely around the cabinet in a circle. As the frequency goes higher, the sound starts to become more directional. The transition from omnidirectional to directional is based on the width of the front baffle, and normally occurs over several hundred Hz, even extending up to the tweeter.
So, no matter how a driver looks on paper, the cabinet will effect the performance, particularily through this baffle step transition.
The loss from going from omnidirectional to directional is normally around 4-6 dB when the speakers are well out into the room, although this varies somewhat with the driver response.
To avoid or reduce the effect, you may use a dipole arrangement, mount the speakers in wall, or crowd them back against the rear wall. Many use a second woofer up to the transition area, then roll this off electrically, and independently from its twin, with a coil/cap.

Tim

[Tim Moorman]

You will need to do some homework on the crossover circuit needed to compensate for this (BSC) and other concerns for the final design.

http://www.pvconsultants.com/audio/frdgroup.htm

Also, search here for the Baffle Step simulators available from some of the members.

Tim

[Tim Moorman]

Originally posted by Tim Moorman
[The loss from going from omnidirectional to directional is normally around 4-6 dB Tim [/B]

That should read " loss from directional to omnidirectional...."

Tim

[rv1890]

thank tim...u helped a lot..appreciate it

[Grandma´s_SUB]

Hello!

Am I wrong about sensitivity(pic):





Dean