Efficiency of Box

[Mos Fetish]

Hi

Here's my situation.

I'm contemplating a vented box with 2 x 6.5" bass drivers each rated at 87db. The tweeter will is rated at 91db.

So, prima facie, I'm going to have to attentuate the tweeter. But here's my question:

Does the fact of having two drivers at 87db in the same box alter the simple fact that there's 4db difference between the tweeter and the bass units, that is, as far as my attenuation calcs go. I sense that I'm missing something important here.

Mos

 

[roddyama]

Hi Mos,

You’re going to get approximately +3db from the pair of 6.5” woofers in proximity to each other. So the sensitivity of the pair will be ~90db.

Rodd Yamashita

 

[Mos Fetish]

Thanks, Rodd.

"Approximately +3db"? So, does the actual efficiency boost depend on the proximity of the bass drivers to each other or some other factor.

You see, if the efficiency of the bass drivers ends up at +90db, then do I really need to attentuate the tweeter by <1 db?

Mos

 

[Kanga]

Plus if you wire the speakers in parallel, you will gain another 3dB of voltage sensitivity, so add 6dB overall. So you will be attenuating the bass-mids!

What impedence are the mid-bass drivers?

Mick

 

[Mos Fetish]

Hi, Mick.

Really? 6db? Yes, they will be in parallel. And, it's a 3 way system.

I read somewhere that it's better to attenuate the tweeter than the bass because the tweeter receives the least power in the system. Correct?

Actually, the mid driver is also 87db, so that means I'd have to pull the bass drivers back by 6db to match it. Seems kinda counterproductive. Gee, this throws the whole thing out of whack!

So, this is guaranteed? Those two bass drivers wired in parallel are going to give me a 6db increase in efficiency, effectively raising my bass section sensitivity to 93db (87+6)?

Mos

 

[Mos Fetish]

Sorry, Mick: the bass drivers are 8 ohm.

 

[Kanga]

Mos - you will get 6dB extra. The free 3dBs are a result of correlated sound waves adding together. I don't really understand why, but there have been previous threads about this. Do a search.

It would be better to pad the tweeter, however if you want to incorporate baffle step compensation, then you will lose 3dB in the bass-mid above the baffle step frequency, which will bring you back to 90 dB. In theory the baffle step is 6dB, but in practice most people seem to find that 3dB is sufficient, due to room reflections. There is a good article on Rod Elliots ESP site about this.

Mick

 

[Doug]

Mos - you will get 6dB extra. The free 3dBs are a result of correlated sound waves adding together. I don't really understand why, but there have been previous threads about this. Do a search.

It would be better to pad the tweeter, however if you want to incorporate baffle step compensation, then you will lose 3dB in the bass-mid above the baffle step frequency, which will bring you back to 90 dB. In theory the baffle step is 6dB, but in practice most people seem to find that 3dB is sufficient, due to room reflections. There is a good article on Rod Elliots ESP site about this.

Mick

When you double the number of speakers in an enclosure, you double the radiating area over just one speaker, hence, the 3 dB increase in output ( leaving the power input the same ).

When you have 2 speakers in parallel, you can now double the power input for another 3 dB increase in output. 3 dB + 3 dB = 6 dB increase.

You don't loose 6 dB above the baffle step frequency as stated, you loose 6 dB " below " the baffle step frequency.

The theory of baffle step is this: If you use an audio oscillator and amplifier and connect it to your speaker, start at the highest frequency and start sweeping down. The wavelenghts at the higher frequencies are shorter than the width of the front baffle so all of the energy is radiated forward. It is said to be radiating into a " full space ".

Now, as you sweep down in frequency, the wavelenghts get longer. Eventually the wavelenghts get as long as the width of the baffle and once they get longer than that, they start wrapping around the edge of the baffle so that all of the energy is no longer radiating forward. It is now said that the speaker is radiating into " half space ". You loose approx 6 dB in SPL. You usually have to attinuate the higher frequencies above the baffle step frequency to bring them down to the level of the lower frequencies for a flat frequency response.

Hope this helps;


😀 Doug

 

[Mos Fetish]

Thanks, Doug.

So, if I crossover my woofer at 500hz say, that freq has a wavelength of approx 688mm. Since my front baffle will be less than 300mm then ALL frequencies put out by the woofer will radiate into halfspace and be attentuated by 6db?

Right?

So, effectively I'm back to my original 87db woofer sensitivity and will need to attenuate the tweeter by 4db as per my original post.

Of course, this raises the problem for the mid freqs. By my calcs all freqs between 500 and approx 1150 will also by attenuated by the baffle and thus will reduce the mid driver's sensitivity in that range well below it's nominal 87db. Eek!

 

[Doug]

Baffle step

Mos;

Don't say eek!! just yet. It's really a good thing cuz you don't have to attinuate the mids as there will be less components to worry about, just the tweets as is usually the case.

BTW the baffle step we are talking about is also known as diffraction loss.

Are you using a computer program to do the simulation for your crossover? Do you have equipment to measure the results when you are done?

We will talk later as it's 7 AM here in Canada and I haven't been to bed yet!! I have to be up at the crack of noon.

Later; Doug Z Z Z Z Z Z

 

[Doug]

crossover point

PS.

As a general rule when selecting crossover points, and you may already know this, try to keep the crossover points at least a mininum of 1 octave away from the resonance peak of the driver you are using.

Doug

 

[cocolino]

originally posted by Kanga
Mos - you will get 6dB extra. The free 3dBs are a result of correlated sound waves adding together. I don't really understand why, but there have been previous threads about this. Do a search.

I believe You`ll find some useful information in this thread:

Multiple Drivers to Increase Efficiency

 

[planet10]

When you have 2 speakers in parallel, you can now double the power input for another 3 dB increase in output. 3 dB + 3 dB = 6 dB increase.

This is only true for a perfect SS amp (which cannot exist). SS amps with very stiff power supplies get close. A tube amp will gain you no extra power. With some amps you will actually get less power (ie i think the Zen is an example -- if we go from 8 to 4 ohms, and certainly an OTL tube amp -- the latter will put out more power if you series the speakers).

So in practice you get anywhere from almost a 3 dB gain to a loss due to the different amp outputs into a lower impedance -- it is very amplifier dependent.

So what amp do you have?

dave

BTW, Doug's analysis (other than the note above) is very good. It is always important to consider baffle-step -- there is a comprehensive article in the Wiki on this subject.

 

[Mos Fetish]

Dave:

Hi. I have a Perreaux 2150.

Doug:

Um, are you saying it won't matter if the mid freqs are attenuated by the baffle step? I'm not clear on this.

Thanks.

 

[Kanga]

You don't loose 6 dB above the baffle step frequency as stated, you loose 6 dB " below " the baffle step frequency. The theory of baffle step is this: If you use an audio oscillator and amplifier and connect it to your speaker, start at the highest frequency and start sweeping down. The wavelenghts at the higher frequencies are shorter than the width of the front baffle so all of the energy is radiated forward. It is said to be radiating into a " full space ".

Now, as you sweep down in frequency, the wavelenghts get longer. Eventually the wavelenghts get as long as the width of the baffle and once they get longer than that, they start wrapping around the edge of the baffle so that all of the energy is no longer radiating forward. It is now said that the speaker is radiating into " half space ". You loose approx 6 dB in SPL. You usually have to attinuate the higher frequencies above the baffle step frequency to bring them down to the level of the lower frequencies for a flat frequency response.

My turn for a correction..

Above the baffle step frequency you are radiating into half space, and below it you are radiating into full space.
Given that driver sensitivity and frequency response is most often measured into half space (but depends on the manufacturer), Doug is correct that you lose 6dB in theory below the baffle step frequency. However if you compensate for the baffle step in the XO (say by 3dB), then this will reduce sensitivity by 3dB above the baffle step frequency. So the net sensitivity of your mid-bass drivers will be 87 +6 - 3 = 90.

Mos - it sounds like you are planning to cross over to a bass speaker at 500 Hz. Is this correct? If this is right, and this frequency is above your baffle step frequency, then you are correct that you won't need baffle step correction, and so your sensitivity will just be 87 + 6 = 93.

Mick

 

[Mos Fetish]

Hi Mick

No. My bass driver will handle 40-500hz, therefore it's whole freq range will be wider than the baffle and will require compensation.

Correct? Man, this thing's getting complicated.

Anything to say about my mid freqs (and the diffraction loss) which will be from 500 up to around 4000? As I said, my baffle will be <300mm wide which means diffraction loss will affect the mids up until around 1100hz. Effectively, my lower mid freqs will be down--not a good thing--but my upper mids should be ok. Am I getting this?

Mos

 

[Mos Fetish]

I said:

"My bass driver will handle 40-500hz, therefore it's whole freq range will be wider than the baffle and will require compensation"

What I meant was: "will suffer from diffraction loss"

 

[Mos Fetish]

So, not to belabour the point but this is what I seem to have now:

Woofer: 87db, 8ohm -- Required range: 40-500hz

Coupling and parallel connection will provide a 6db gain in sensitivity. 87+6=93

However, all of the woofer's freqs fall within diffraction loss which will attenuate them by 6db. So 93-6=87.

I'm back to where I started. I assume a zobel isn't going to mess this up further?

Now...

Midrange: 87db, 8ohm--required range: 500-4000hz

Assuming a baffle of width 300mm, the mids will suffer a 6db diffraction loss from 500hz up to around 1100hz, but from 1100 to 4000hz they will not.

So, sensitivity is now 81db (!!) from 500 to 1100hz (87-6) but remains at 87db above 1100hz.

****, is this correct?

Mos

 

[Nappylady]

Okay, I'm only addressing the attenuation of various components being "better" than attenuating others: if you attenuate tweeters, you can use lower power rated components because there will not be as much power flowing through them. If you're going to deal with 100w signals, you need a 100w+ L-pad for the woofers, but you could get away just fine with perhaps a 25w L-pad for the tweeters, because only a small portion of the sound will be going through them.

Someplace there's a formula for the percentage of wattage rating required for a given crossover freqency and slope; I don't have it or know it though, so all I can say is that the curve will look like the top-right hand side of a circle, slowly dropping off at first, then massively dropping off as you get close to 20khz. You could use a 0-watt L-pad to attenuate frequencies above 24khz any amount you want. 😀

 

[planet10]

I have a Perreaux 2150.

I'm not really familiar with that amp. What is its rated power into 4 ohms vrs 8?

dave