Speaker sensibility
If I have a speaker that have 90dB SPL and I put two in a box, will my system have 93dB SPL? and if I keep doubling the number of speakers will I have an exponencial increase in sensibility?
If I have a speaker that have 90dB SPL and I put two in a box, will my system have 93dB SPL? and if I keep doubling the number of speakers will I have an exponencial increase in sensibility?
Theoretically yes, but practically no. The drivers have to be in such proximity to each other that their outputs add at the point in space where you are making the measurement. As you can imagine, as the number of drivers start to increase, the smaller the chance that the drivers outputs are going all sum at a single measurement point.
I asked that because I saw some loudspeakers bigger than I am, with loads of speakers and wonder why they are built that way... they have to be conected in series/parallel in a way that it acts like a single 8ohms speaker and uses a single crossover...
I have a pair of JBL N24II 86dB and a acoustic research 108PSB subwoofer, I like its sound but it doesnt play too loud.
I have a pair of JBL N24II 86dB and a acoustic research 108PSB subwoofer, I like its sound but it doesnt play too loud.
Rodd:
My understanding is that 2 10" woofers placed next to each other will raise the SPL for any given drive voltage level 6 dB. Isn't that true? Of course, adding one more speaker in parallel converts a single 8 ohm speaker into a 4 ohm pair, which will draw twice the current at any given voltage level, which means that the pair draws more watts from the amp. Still, there is 3 dB advantage for the pair over the single driver, I thought.
My understanding is that 2 10" woofers placed next to each other will raise the SPL for any given drive voltage level 6 dB. Isn't that true? Of course, adding one more speaker in parallel converts a single 8 ohm speaker into a 4 ohm pair, which will draw twice the current at any given voltage level, which means that the pair draws more watts from the amp. Still, there is 3 dB advantage for the pair over the single driver, I thought.
Hi Kelticwizard,
The advantage from two drivers over one should be qualified bexause it is both acoustical and electrical. If we keep the delivered power constant and double only the drivers, the gain is 3db. If we, as you suggest, maintain a constant voltage, and wire the two drivers in parallel, the current through the pair will double therefore doubling the delivered power. The gain would then be both acoustcal and electrical for a total gain of 6db.
The advantage from two drivers over one should be qualified bexause it is both acoustical and electrical. If we keep the delivered power constant and double only the drivers, the gain is 3db. If we, as you suggest, maintain a constant voltage, and wire the two drivers in parallel, the current through the pair will double therefore doubling the delivered power. The gain would then be both acoustcal and electrical for a total gain of 6db.
Dear Aunty DIY, I have a problem with this..
Dear Aunty DIY
I have a problem and would like your help.
Connecting two drivers in parallel halves the impedance and increases the sensitivity by 3dB. This we can be sure of and is immediately obvious when we have to turn the volume control down for the same SPL (loudness).
Regarding the increase in efficiency, I'm not at all convinced.
If we take another two drivers and connect all four in series/parallel so the impedance is the same as a single driver then we should get a 6dB increase in efficiency relative to the single driver. Now (and here's where I get a bit hazy) shouldn't this give us a 6dB rise in SPL for the same volume control setting? Please help me here Roddyama - it's still a bit early in the day for me.
Anyway, when I compare one drive unit with four in series/parallel, the four are NOT 6dB louder than the one. Nothing like. Maybe 2-3dB. Perhaps Roddyama's "proximity" thing comes into play here but, as you all know, I use 2" drivers and they're quite close together.
I think that something's not right with the conventional wisdom here. I've thought so for a while actually but have been too ashamed to mention it. I could be ostrasized from the entire audio community.
Please advise
"Worried" of Surrey
Dear Aunty DIY
I have a problem and would like your help.
Connecting two drivers in parallel halves the impedance and increases the sensitivity by 3dB. This we can be sure of and is immediately obvious when we have to turn the volume control down for the same SPL (loudness).
Regarding the increase in efficiency, I'm not at all convinced.
If we take another two drivers and connect all four in series/parallel so the impedance is the same as a single driver then we should get a 6dB increase in efficiency relative to the single driver. Now (and here's where I get a bit hazy) shouldn't this give us a 6dB rise in SPL for the same volume control setting? Please help me here Roddyama - it's still a bit early in the day for me.
Anyway, when I compare one drive unit with four in series/parallel, the four are NOT 6dB louder than the one. Nothing like. Maybe 2-3dB. Perhaps Roddyama's "proximity" thing comes into play here but, as you all know, I use 2" drivers and they're quite close together.
I think that something's not right with the conventional wisdom here. I've thought so for a while actually but have been too ashamed to mention it. I could be ostrasized from the entire audio community.
Please advise
"Worried" of Surrey
Dear "Worried":
This has been covered before in this forum. I will try to dig up the thread.
My understanding is that the 6 dB in sensitivity-those who deal with it actually say it is "near 6 dB"-only occurs if the the 4 speakers are put into a square arrangement:
OO
OO
When you put them into vertical arrangement,
O
O
O
O
the increase in sensitivity becomes considerably less as the "mutual loading" concept lessens.
I'll try to locate the thread. I do remember that P. Lacombe mentioned that a pair had only an increase of 1.3 dB, and therefore a vertical arrangement of four speakers would have the 2 or 3 dB increase you mentioned.
Edit: Here it is:
http://www.diyaudio.com/forums/showthread.php?threadid=1319&highlight=increase
This has been covered before in this forum. I will try to dig up the thread.
My understanding is that the 6 dB in sensitivity-those who deal with it actually say it is "near 6 dB"-only occurs if the the 4 speakers are put into a square arrangement:
OO
OO
When you put them into vertical arrangement,
O
O
O
O
the increase in sensitivity becomes considerably less as the "mutual loading" concept lessens.
I'll try to locate the thread. I do remember that P. Lacombe mentioned that a pair had only an increase of 1.3 dB, and therefore a vertical arrangement of four speakers would have the 2 or 3 dB increase you mentioned.
Edit: Here it is:
http://www.diyaudio.com/forums/showthread.php?threadid=1319&highlight=increase
Speaker Array Sensitivity
While ultimately the SPL of the speaker array would best be measured, in the development process you can compute the system sensitivity (or efficiency) impact of an array. First, assume that an individual driver in the line has a known SPL value. Next assume that the drivers in the array have overlapping acoustical radiation patterns and are spaced within a wavelength center-to-center from each other. Thus, the acoustic improvement in efficiency gain at 1 watts, 1 meter is given by:
Efficiency Gain = 10*log (Number of Drivers Driven)
while the sensitivity gain or loss at 2.83v, 1 meter is:
Sensitivity Gain/Loss = 10*log (Nominal Driver Impedance/Nominal Array Impedance)
If the nominal array impedance is less than the nominal driver impedance, the sensitivity increases or is a gain. If the array impedance is greater than the individual driver impedance, then the sensitivity decreases or becomes a loss.
Hence, for the overall system
System Efficiency Gain = SPL + Efficiency Gain
System Sensitivity = SPL + System Efficiency Gain + Sensitivity Gain/Loss
As an example, consider a case wherein we have 4 drivers connected in 2 parallel groups of 2 in series. Each individual driver is 8 ohms impedance and has an SPL of 85 dB. Hence, the efficiency gain, total impedance, and sensitivity gain are:
Efficiency Gain = 10 log 4 = 6.02 dB
Total Impedance of the Combination = 1 / (1/8 + 1/8) = 8 ohms
Sensitivity Gain = 10 log (8/8) = 0 dB
Hence, the sensitivity increase of the array is:
System Sensitivity = 85 + 6.02 + 0 = 91.02 dB.
Botom line though is that I find that measurements are preferred as the theoretical system sensitivity calculation tends to overstate the actual sensitivity for a large array.
While ultimately the SPL of the speaker array would best be measured, in the development process you can compute the system sensitivity (or efficiency) impact of an array. First, assume that an individual driver in the line has a known SPL value. Next assume that the drivers in the array have overlapping acoustical radiation patterns and are spaced within a wavelength center-to-center from each other. Thus, the acoustic improvement in efficiency gain at 1 watts, 1 meter is given by:
Efficiency Gain = 10*log (Number of Drivers Driven)
while the sensitivity gain or loss at 2.83v, 1 meter is:
Sensitivity Gain/Loss = 10*log (Nominal Driver Impedance/Nominal Array Impedance)
If the nominal array impedance is less than the nominal driver impedance, the sensitivity increases or is a gain. If the array impedance is greater than the individual driver impedance, then the sensitivity decreases or becomes a loss.
Hence, for the overall system
System Efficiency Gain = SPL + Efficiency Gain
System Sensitivity = SPL + System Efficiency Gain + Sensitivity Gain/Loss
As an example, consider a case wherein we have 4 drivers connected in 2 parallel groups of 2 in series. Each individual driver is 8 ohms impedance and has an SPL of 85 dB. Hence, the efficiency gain, total impedance, and sensitivity gain are:
Efficiency Gain = 10 log 4 = 6.02 dB
Total Impedance of the Combination = 1 / (1/8 + 1/8) = 8 ohms
Sensitivity Gain = 10 log (8/8) = 0 dB
Hence, the sensitivity increase of the array is:
System Sensitivity = 85 + 6.02 + 0 = 91.02 dB.
Botom line though is that I find that measurements are preferred as the theoretical system sensitivity calculation tends to overstate the actual sensitivity for a large array.
Ooops on Impedance Calculation for Example
For the example in my previous I screwed up the impedance calculation. Two 8 ohms drivers in series would be 16 ohms and then if you parallel two groups of these 2 in series, you get 8 ohms. The calculation is thus,
Z (overall) = 1/(1/16 + 1/16) = 8 ohms
For the example in my previous I screwed up the impedance calculation. Two 8 ohms drivers in series would be 16 ohms and then if you parallel two groups of these 2 in series, you get 8 ohms. The calculation is thus,
Z (overall) = 1/(1/16 + 1/16) = 8 ohms
Being math challenged and simple minded, and since 'sound is round' as hitware on the FR forum likes to say, I try to envision the impact (or lack thereof) of two or more relatively transparent balls having a 'coming together' at some point in space. When they just barely touch, this ~represents pi*WL, the frequency where particle density summing begins and according to (proven) theory will be +3dB when they ~completely overlap (some much lower frequency), becoming as one. This implies that beyond pi*WL, any gain is due to reflections and varying rates of decay over distance of the surrounding BW so will vary depending on where in the polar response it's measured if all drivers are on the same plane.
Not very scientific, but has worked well for me in 'calculating' a real world approximation.
GM
Not very scientific, but has worked well for me in 'calculating' a real world approximation.
GM
Re: Speaker Array Sensitivity
Once more I'm forced to immerse myself into my text-books.
This is generally not a good idea for an 'ear-man' like myself. Last time I went this way I was researching the possibilities of a Helmholtz resonator to absorb the top/bottom standing wave of a tall, slim column loudspeaker. I still blame you for this, kelticwizard.
And where did all this theory lead me? Well, according to keltic I came up with a "Helical Silicon Technology creating a synergistic interaction between enhanced structural stability and suppressed negative reflectivity." I described this miracle of technology as "putting a sock in it". The sock was filled with sand and sat, coiled like a snake, at the bottom of the column. I should add that I got much of the inspiration for this sock from keltic, zozo and other members of the diyAudio.com forum.
But I digress ...
On this occasion, I'm delving into "Loudspeakers" by EJ Jordan (1963 Focal Press). This is fitting as I heard my first mini line array, many years ago, at Ted Jordan's house when he was helping me to design a stereo system for my new car.
First, a few pertinent words from the Preface of the book (when I see a book has a Preface I already get a headache) ...
"Whenever there is a difference between theoretical and practical results, one or both of them must be wrong. There are no exceptions to ths rule."
Hmmm ....
Ted Jordan's conclusion on the efficiency of multiple unit systems (N similar units mounted, close together, on the same baffle) was as follows ...
"It will be seen that, provided the total radiation mass is much less than the mass of the cone and coil system, there will be a gain in efficiency below the frequency where kr=2. If the radiation mass were negligible, the gain in efficiency would be N. In practice it will always be less than this and will decrease as N increases, ultimately becoming independent of N. Above the frequency where kr=2, the efficiency is constant.
The frequency where kr=2 is proportional to 1/root N.
An additional effect of the radiation mass is to decrease the resonant frequency of each unit as N is increased."
I assume that if N is, for example, 2 units and the gain in efficiency would therefore be 2, this is the equivalent of +3dB.
So ...
Was Mr. Jordan correct?
Does the resonant frequency of each unit increase as N is increased?
Are Jim's equations the same as this or an update and can the differences between theory and practice be accounted for by the "total radiation mass not being much less than the mass of the cone and coil system"?
Come to think of it, what the heck is the "total radiation mass" and should we make sure that Saddam Hussein is not given access to it?
I eagerly await your views.
Still "Worried" of Surrey
Once more I'm forced to immerse myself into my text-books.
This is generally not a good idea for an 'ear-man' like myself. Last time I went this way I was researching the possibilities of a Helmholtz resonator to absorb the top/bottom standing wave of a tall, slim column loudspeaker. I still blame you for this, kelticwizard.
And where did all this theory lead me? Well, according to keltic I came up with a "Helical Silicon Technology creating a synergistic interaction between enhanced structural stability and suppressed negative reflectivity." I described this miracle of technology as "putting a sock in it". The sock was filled with sand and sat, coiled like a snake, at the bottom of the column. I should add that I got much of the inspiration for this sock from keltic, zozo and other members of the diyAudio.com forum.
But I digress ...
On this occasion, I'm delving into "Loudspeakers" by EJ Jordan (1963 Focal Press). This is fitting as I heard my first mini line array, many years ago, at Ted Jordan's house when he was helping me to design a stereo system for my new car.
First, a few pertinent words from the Preface of the book (when I see a book has a Preface I already get a headache) ...
"Whenever there is a difference between theoretical and practical results, one or both of them must be wrong. There are no exceptions to ths rule."
Hmmm ....
Ted Jordan's conclusion on the efficiency of multiple unit systems (N similar units mounted, close together, on the same baffle) was as follows ...
"It will be seen that, provided the total radiation mass is much less than the mass of the cone and coil system, there will be a gain in efficiency below the frequency where kr=2. If the radiation mass were negligible, the gain in efficiency would be N. In practice it will always be less than this and will decrease as N increases, ultimately becoming independent of N. Above the frequency where kr=2, the efficiency is constant.
The frequency where kr=2 is proportional to 1/root N.
An additional effect of the radiation mass is to decrease the resonant frequency of each unit as N is increased."
I assume that if N is, for example, 2 units and the gain in efficiency would therefore be 2, this is the equivalent of +3dB.
So ...
Was Mr. Jordan correct?
Does the resonant frequency of each unit increase as N is increased?
Are Jim's equations the same as this or an update and can the differences between theory and practice be accounted for by the "total radiation mass not being much less than the mass of the cone and coil system"?
Come to think of it, what the heck is the "total radiation mass" and should we make sure that Saddam Hussein is not given access to it?
I eagerly await your views.
Still "Worried" of Surrey
Interesting thread.
P.Lacombe seemed to be pretty much 'on the money'.
The main difference between E J Jordan's explanation for the difference between theory and practice and the explanation of P.Lacombe and others is that Jordan had this "total radiation mass not being much less than the mass of the cone and coil system" thing whereas Lacombe et al claim that the loss over theoretical values is caused by incomplete acoustical coupling of the drivers. Ted Jordan also mentioned this when he said "N similar units mounted on the same baffle close together".
Could both be true?
Steve
Re: Re: Speaker Array Sensitivity
That actually got published as a book? I got a loose leaf ring binder from him with that book in it.
as far as i can see.
dave
kelticwizard said:
That actually got published as a book? I got a loose leaf ring binder from him with that book in it.
as far as i can see.
dave
phase_accurate said:
Yes, yes, ok Charles, I meant decrease.
Increase - decrease - increase - decrease, what's the difference?
God, you people are SO fussy.
Re: Re: Speaker Array Sensitivity
LOL, don't look now, but I think I am about to give you something else to blame me for.
What's k? I saw the same variable in Small's paper on ported boxes, and it had something to do with efficiency. Never did figure out exactly what it meant, but Small gave examples so that you got the general idea. As an aside, even though efficiency is considered a good thing, the most efficient ported box alignments are high in distortion compared to other ported box alignments.
If you can explain k, it would greatly help.
I don't think you have to be worried about being ostracized from the audio community. I am getting the impression that quite a few people are only in the audio community in the first place because all the other communities have given them the heave-ho. They are in no position to ostracize anybody else at all. 😉
7V said:
LOL, don't look now, but I think I am about to give you something else to blame me for.
What's k? I saw the same variable in Small's paper on ported boxes, and it had something to do with efficiency. Never did figure out exactly what it meant, but Small gave examples so that you got the general idea. As an aside, even though efficiency is considered a good thing, the most efficient ported box alignments are high in distortion compared to other ported box alignments.
If you can explain k, it would greatly help.
I don't think you have to be worried about being ostracized from the audio community. I am getting the impression that quite a few people are only in the audio community in the first place because all the other communities have given them the heave-ho. They are in no position to ostracize anybody else at all. 😉
Re: Re: Re: Speaker Array Sensitivity
where, f = frequency
and, c = velocity of sound in air.
The r mentioned is the radius of the cone.
However, if it all goes horribly wrong, don't blame me, I'm just a speaker designer. The only way I'd get anywhere near the speed of sound in air is if I was dropped from a great height.
The contributor formerly known as "Worried"
I believe that k = (2 x PI x f)/ckelticwizard said:
where, f = frequency
and, c = velocity of sound in air.
The r mentioned is the radius of the cone.
However, if it all goes horribly wrong, don't blame me, I'm just a speaker designer. The only way I'd get anywhere near the speed of sound in air is if I was dropped from a great height.
So we're all "pre-ostracized". That's handy. Phew.
The contributor formerly known as "Worried"
When did the thread change from sensibility to sensitivity
I quite liked the idea of rating speakers on a scale from sensible to silly. Maybe something like:
Turn that thing off it's putting me to sleep
Dull as dishwater
Extremely sensible
Very sensible
Sensible
Slightly sensible
Slightly silly
Silly
Now that's just plain silly
Very silly indeed
Extremely silly
Welcome to the forum
Steve (more than slightly silly)
I quite liked the idea of rating speakers on a scale from sensible to silly. Maybe something like:
Turn that thing off it's putting me to sleep
Dull as dishwater
Extremely sensible
Very sensible
Sensible
Slightly sensible
Slightly silly
Silly
Now that's just plain silly
Very silly indeed
Extremely silly
Welcome to the forum
Steve (more than slightly silly)
Gurardian Spelling Angel
7V:
I do believe that this forum has a benign Higher Force which changes the most egregious spelling mistakes in thread titles.
I found that out when I started the thread about the Frankfurt High End Show. I spelled it in the thread title and opening message as "Frankfort".
The reason for this is that the capital of the state of Kentucky is Frankfort, KY and it gets into the national news occasionally. Threw me off.
You can imagine my relief when I realized that in the thread title, at least, some kind soul had changed the spelling to Frankfurt.
Somebody up there is watching out for diyAudio.
7V:
I do believe that this forum has a benign Higher Force which changes the most egregious spelling mistakes in thread titles.
I found that out when I started the thread about the Frankfurt High End Show. I spelled it in the thread title and opening message as "Frankfort".
The reason for this is that the capital of the state of Kentucky is Frankfort, KY and it gets into the national news occasionally. Threw me off.
You can imagine my relief when I realized that in the thread title, at least, some kind soul had changed the spelling to Frankfurt.
Somebody up there is watching out for diyAudio.
- Status
- Not open for further replies.