I have several drivers, both 4 and 8 ohms of 82 dB sensitivity, to use together.
If, for example, we use four 8 ohm and two 4 ohm together in series/parallel,
will the 4 ohm drivers play louder than the 8 ohms?
I looked up " power sharing" and saw the distribution of amplifier power (but determining the sensitivity/output level are still uncertain).
If, for example, we use four 8 ohm and two 4 ohm together in series/parallel,
will the 4 ohm drivers play louder than the 8 ohms?
I looked up " power sharing" and saw the distribution of amplifier power (but determining the sensitivity/output level are still uncertain).
Attachments
Output will follow power received by each unit.
Speaker sensitivity per se will remain the same, now depending on how you array them, that may focus them and have somewhat higher SPL than expected at a certain point.
Lots of people claim that by using 2 speakers side by side sensitivity rises by 3 dB (doubles) and so on, indefinitely.
I find it nonsense, there is no physical justification for that and personally never could make that work.
I know I will ruffle some feathers.
Speaker sensitivity per se will remain the same, now depending on how you array them, that may focus them and have somewhat higher SPL than expected at a certain point.
Lots of people claim that by using 2 speakers side by side sensitivity rises by 3 dB (doubles) and so on, indefinitely.
I find it nonsense, there is no physical justification for that and personally never could make that work.
I know I will ruffle some feathers.
Sensitivity and array gain is hard to predict and there is no firm rule as to how much gain you end up with or in which frequency band it occurs. This is easier to predict with LF drivers operating in very close proximity and the gain can be substantial in the right situation.
In theory you get a max of 3 dB sensitivity gain as well as 3 dB power doubling gain with two drivers connected in parallel close proximity. The reality is more like 1-2 dB coupling gain with smaller mid drivers in a limited bandwidth and the distance/wavelength relationship dictates the gain. The more drivers you have operating in the same bandwidth, the higher the coupling gain. In theory it will double (3 dB gain) with every doubling of driver count, but again this isn't what happens in reality. The gain will diminish quickly past 4 drivers, especially in a line array. A clustered array is more effective here but that configuration isn't very useful in most audio settings.
In theory you get a max of 3 dB sensitivity gain as well as 3 dB power doubling gain with two drivers connected in parallel close proximity. The reality is more like 1-2 dB coupling gain with smaller mid drivers in a limited bandwidth and the distance/wavelength relationship dictates the gain. The more drivers you have operating in the same bandwidth, the higher the coupling gain. In theory it will double (3 dB gain) with every doubling of driver count, but again this isn't what happens in reality. The gain will diminish quickly past 4 drivers, especially in a line array. A clustered array is more effective here but that configuration isn't very useful in most audio settings.
“Output will follow power received by each unit.”
Ok,
so do the 4 ohm drivers receive more (or less) power than the 8 ohm drivers in the above diagram, and by how much? (I saw a few examples circuits on the power sharing webpage, but none resembling my array).
Thanks for your expertise.
Ok,
so do the 4 ohm drivers receive more (or less) power than the 8 ohm drivers in the above diagram, and by how much? (I saw a few examples circuits on the power sharing webpage, but none resembling my array).
Thanks for your expertise.
Each 4 ohm gets 25%, each 8 ohm 12.5% of the power, but.... remember those are only 'Nominal' impedances, not actual....
Thanks for the estimate.
So if the four 8 ohm units in series/parallel put out 6db more than one 8 ohm driver(each is 82db), that’d put them at around 88db for 2.83v from the amp.
But does that mean those two 4ohm units together will play 3db louder still (say 91db), or would they’d also be at around 88db for 2.83v amp output?
(Or am I missing something?)
Thanks again.
So if the four 8 ohm units in series/parallel put out 6db more than one 8 ohm driver(each is 82db), that’d put them at around 88db for 2.83v from the amp.
But does that mean those two 4ohm units together will play 3db louder still (say 91db), or would they’d also be at around 88db for 2.83v amp output?
(Or am I missing something?)
Thanks again.
Just simplify and look at them as plain resistors with a DC voltage across the lot of them and then it's a simple case of working out the V^2 / R (i.e. the power) on individual drivers (resistors) to work out the share of power.
In this example it's very easy to visualise this without the use of a calculator.
Thx. I believe I understand the power sharing. My real question is about how the sensitivity
of the two 4 ohm drivers compares with the four 8 ohm ones when wired as shown. I realize now that because the 4 ohms are in series that together they may be the same sensitivity as the four 8 ohm ones. I was just hoping someone could corroborate, or correct me on that.
Thanks
of the two 4 ohm drivers compares with the four 8 ohm ones when wired as shown. I realize now that because the 4 ohms are in series that together they may be the same sensitivity as the four 8 ohm ones. I was just hoping someone could corroborate, or correct me on that.
Thanks
..at 82
I couldn't answer this until you stipulated that sensitivity was being measured per Voltage.. However, you contradict that in post #7, which needs clarifying.
Did I contradict myself? (If so I apologize) Perhaps I’m merely confused.
Both the 8 ohm & 4 Ohm drivers are specified by the manufacturer to be 82db/2.83V@1M,
Although on another sheet they specified 82db/W/1M for the 8 ohm driver.
Doesn’t that mean they have the same sensitivity? If not, please explain.
I’m just trying to determine what happens to sensitivity when mixing these two models of the drivers as shown, 4vs8ohm, but Especially overall sensitivity.
Thanks so much for your patient help!
Both the 8 ohm & 4 Ohm drivers are specified by the manufacturer to be 82db/2.83V@1M,
Although on another sheet they specified 82db/W/1M for the 8 ohm driver.
Doesn’t that mean they have the same sensitivity? If not, please explain.
I’m just trying to determine what happens to sensitivity when mixing these two models of the drivers as shown, 4vs8ohm, but Especially overall sensitivity.
Thanks so much for your patient help!
2.83V@1M is the equiv of 1W@1M - for an 8 ohm load
For a 4 ohm load that 2.83V would produce 2W, so you need to reduce the sensitivity by 3dB to get the same
Power is V^2/R so 2.83^2 =8V, divided by the load (8ohm) = 1W
If you plug 4 ohm load into the calculation that spec produces 8/4 = 2W
If both 4 and 8 ohm drivers are essentially the same model, than the correct "sensitivity" is 82 dB/1W/1m for both models, or 82 dB/2.83V/1m for the 8 ohm driver and 85 dB/2.83V/1m for the 4 ohm driver.
Power distribution per schematic in your post #1:
Each 4-ohm driver get 25% of amplifier power, each 8-ohm driver get 12.5% of amplifier power.
Sensitivity of two 4-ohm drivers in series = 85 dB/2.83V/1m.
Sensitivity of parallel/series combination of four 8-ohm drivers = 88 dB/2.83V/1m.
Group all four 8-ohm drivers together (in-line), one 4-ohm driver at top and other 4-ohm driver at bottom, like this:
4-8-8-8-8-4 (rotate this line 90 degrees)
Total sensitivity of all 6 drivers = 89.8 dB/2.83V/1m.
Total impedance loading: 4 ohm.
Power distribution per schematic in your post #1:
Each 4-ohm driver get 25% of amplifier power, each 8-ohm driver get 12.5% of amplifier power.
Sensitivity of two 4-ohm drivers in series = 85 dB/2.83V/1m.
Sensitivity of parallel/series combination of four 8-ohm drivers = 88 dB/2.83V/1m.
Group all four 8-ohm drivers together (in-line), one 4-ohm driver at top and other 4-ohm driver at bottom, like this:
4-8-8-8-8-4 (rotate this line 90 degrees)
Total sensitivity of all 6 drivers = 89.8 dB/2.83V/1m.
Total impedance loading: 4 ohm.
Yes, Mike answers the first of your questions and Sonce answers the second.
Note in the 4 ohm case the sensitivity of both is the same as for one, whatever that is. In the 8 ohm case the total is +6dB compared to one.
Note in the 4 ohm case the sensitivity of both is the same as for one, whatever that is. In the 8 ohm case the total is +6dB compared to one.
Great! So we’re around 90db and 4 ohms load. Almost done-
Now…(ready for this?) if we take TWO of the above 4-8-8-8-8-4 arrays, and put them together in series, it looks like the sensitivity would still be 89.8db (or does it change?)
I hope to place each array, vertically in line, flanking an 8 ohm, 89db tweeter of 40mm dia,(above & below) to balance the high treble of the “full-range” drivers (which get choppy & directional above about 4 to 6 kHz). Any advice or caveats?
Thanks for any informed opinions here.
-Charles
Now…(ready for this?) if we take TWO of the above 4-8-8-8-8-4 arrays, and put them together in series, it looks like the sensitivity would still be 89.8db (or does it change?)
I hope to place each array, vertically in line, flanking an 8 ohm, 89db tweeter of 40mm dia,(above & below) to balance the high treble of the “full-range” drivers (which get choppy & directional above about 4 to 6 kHz). Any advice or caveats?
Thanks for any informed opinions here.
-Charles
Last edited:
Sensitivity will be again 89.8 dB/2.83V/1m - but only at low frequencies (say, below some 500 Hz - depending on the driver diameter). Higher in frequency it will be less.
If you want to use two such line arrays, you are entering complicated and dangerous territory, where nothing is so simple as it appears. Bearing this in mind, then:
1. For open space, outside - "J" line array: 4-4-4-4-8-8-8-T-T-8-8-8-8-8 (curved part of "J" below, with 4-4-4-4 group)
2. Inside, in room: inverted "J" line array, like Don Keele's CBT (google this). It needs shading. Tweeters near, but not at the top.
If you have a larger number of those drivers, do not mix different drivers - make two pairs of line arrays - one pair from 4-ohm drivers, another pair with 8-ohm drivers.
If you want to use two such line arrays, you are entering complicated and dangerous territory, where nothing is so simple as it appears. Bearing this in mind, then:
1. For open space, outside - "J" line array: 4-4-4-4-8-8-8-T-T-8-8-8-8-8 (curved part of "J" below, with 4-4-4-4 group)
2. Inside, in room: inverted "J" line array, like Don Keele's CBT (google this). It needs shading. Tweeters near, but not at the top.
If you have a larger number of those drivers, do not mix different drivers - make two pairs of line arrays - one pair from 4-ohm drivers, another pair with 8-ohm drivers.
If you put two drivers far apart from each other and measure between them you should get +6dB.
“Sensitivity will be again 89.8 dB/2.83V/1m - but only at low frequencies (say, below some 500 Hz - (depending on the driver diameter). Higher in frequency it will be less.”
The drivers are 2” diameter diaphragms, in 2.5” frames. Would the sensitivity
still be anywhere near 89db at 1, 2, or 3khz? I. Have enough 8 ohm drivers to make line arrays of 8 or even 12, but would need to add a few 4ohm units for an array of 16 (I had hoped to boost sensitivity, by incorporating some 4ohm units).
The original plan was to use them “full-range” (above about 300hz), but then learned of the rough and directional behavior above 4-5 kHz. I thought sneaking a 40mm AMT in between the line (at ear level) might be a good idea.
Any way to get away with this- or is it la lost cause?
Thanks.
The drivers are 2” diameter diaphragms, in 2.5” frames. Would the sensitivity
still be anywhere near 89db at 1, 2, or 3khz? I. Have enough 8 ohm drivers to make line arrays of 8 or even 12, but would need to add a few 4ohm units for an array of 16 (I had hoped to boost sensitivity, by incorporating some 4ohm units).
The original plan was to use them “full-range” (above about 300hz), but then learned of the rough and directional behavior above 4-5 kHz. I thought sneaking a 40mm AMT in between the line (at ear level) might be a good idea.
Any way to get away with this- or is it la lost cause?
Thanks.
You are good up to 2 - 3 kHz.
Stick with 8-ohm drivers and make 12-unit line array with 4 parallel drivers in 3 serially connected groups. Total impedance = 6 ohms (nominal) and sensitivity = 94 dB/2.83V/1m.
Is it for home hi-fi, or for professional sound reinforcement (live concert, etc)?
Stick with 8-ohm drivers and make 12-unit line array with 4 parallel drivers in 3 serially connected groups. Total impedance = 6 ohms (nominal) and sensitivity = 94 dB/2.83V/1m.
Is it for home hi-fi, or for professional sound reinforcement (live concert, etc)?
Believe what you want, you are assuming sensitivity doubles with 2 speakers side by side, quadruples with 4 together:
Your array will NOT give you 91dB/W/m , just-test-it.
You will have somewhat more SPL than expected, maybe a couple dB at most for the full array, with two important caveats:
1) you are not increasing sensitivity by the classic definition, you are not increasing efficiency, the electromechnical conversion of energy, but you are focusing whatever you already have at the microphone position, 1 meter away, not the same.
2) most SPL vs distance calculations assume a point source or at least a very small source approaching that.
Don´t remember it by heart but Classic books such as Olsen´s, assumed a 1" to 2" diameter piston as a source, which by itself is more realistic than an impossible point source.
You didn´t give us speaker size so I´ll assume they are all 8".
Why?: it´s a smallish size, quite popular, smaller ones (6", 4") won´t change things MUCH and larger ones will only separate centers more, so let´s go on.
2 side by side 8" speakers now are 42 cm wide; 4 of them in a square, edge to edge, now 42 by 42 cm.
For a microphone placed a mere 100 cm away, they are not a point source by any means, you do not have a spherical waveform but something intermediate between a spherical wave and a flat front one.
Definitely NOT losing 6dB by each distance duplication, so louder "than it should" at that point".
Does it prove "efficiency/sensitivity duplication"?
No, this describes a flawed experiment (an experiment based in flawed assumptions).
Now to the main assumption behind "sensitivity/efficiency doubling 3dB per each pairing" and starting with the simplest case: 2 speakers side by side,edges touching.
The common explanation which is offered, without analysis, just repeating it "because everybody says so" is:"efficiency doubles because of cross/mutual coupling" and they leave it there, as if a simple enunciation were enough, without proof/measurements/Physical backing.
If you search for it, the only mechanism to achieve that is that by each speaker contributing sound pressure in nearby space, so locally increasing air density, it helps its partner.
Does it happen?
Yes ... a little.
Does it double air density? Which would justify +3dB.
NO WAY even in the wildest dreams.
That would require double air pressure, 2 full Atmospheres of it.
Just for reference, to reach that point:
I don´t see that happening, so ......
That said, not an Evangelist, each one is entitled to his own opinions, mine is (as always) based on "basic principles" , rooted in Physics.
PS: what if speakers are "separated by a larger distance?"
Well, self answered: even less mutual coupling.
Best case, and it would mean "cheating", you have even less of a point sound source.
I told you that simply does not happen,which is easy to verify but apparently nobody cares to test in the flesh.
Your array will NOT give you 91dB/W/m , just-test-it.
You will have somewhat more SPL than expected, maybe a couple dB at most for the full array, with two important caveats:
1) you are not increasing sensitivity by the classic definition, you are not increasing efficiency, the electromechnical conversion of energy, but you are focusing whatever you already have at the microphone position, 1 meter away, not the same.
2) most SPL vs distance calculations assume a point source or at least a very small source approaching that.
Don´t remember it by heart but Classic books such as Olsen´s, assumed a 1" to 2" diameter piston as a source, which by itself is more realistic than an impossible point source.
You didn´t give us speaker size so I´ll assume they are all 8".
Why?: it´s a smallish size, quite popular, smaller ones (6", 4") won´t change things MUCH and larger ones will only separate centers more, so let´s go on.
2 side by side 8" speakers now are 42 cm wide; 4 of them in a square, edge to edge, now 42 by 42 cm.
For a microphone placed a mere 100 cm away, they are not a point source by any means, you do not have a spherical waveform but something intermediate between a spherical wave and a flat front one.
Definitely NOT losing 6dB by each distance duplication, so louder "than it should" at that point".
Does it prove "efficiency/sensitivity duplication"?
No, this describes a flawed experiment (an experiment based in flawed assumptions).
Now to the main assumption behind "sensitivity/efficiency doubling 3dB per each pairing" and starting with the simplest case: 2 speakers side by side,edges touching.
The common explanation which is offered, without analysis, just repeating it "because everybody says so" is:"efficiency doubles because of cross/mutual coupling" and they leave it there, as if a simple enunciation were enough, without proof/measurements/Physical backing.
If you search for it, the only mechanism to achieve that is that by each speaker contributing sound pressure in nearby space, so locally increasing air density, it helps its partner.
Does it happen?
Yes ... a little.
Does it double air density? Which would justify +3dB.
NO WAY even in the wildest dreams.
That would require double air pressure, 2 full Atmospheres of it.
Just for reference, to reach that point:
I don´t see that happening, so ......
That said, not an Evangelist, each one is entitled to his own opinions, mine is (as always) based on "basic principles" , rooted in Physics.
PS: what if speakers are "separated by a larger distance?"
Well, self answered: even less mutual coupling.
Best case, and it would mean "cheating", you have even less of a point sound source.