Speaker driver efficiency, series vs. parallel etc


Okay Got a quick question...

Lets say you have 90 db 1w/1m drivers...

Now they are 8 ohm as well

You are to run lets say 2 of them in series and get a 16 ohm load, will the efficiency go up? I have heard anytime you double your driver that you will go up essentially 3 db..

So is this circuit now 93 db opposed to 90 db?

Next scenario--
Lets say we use the same drivers, 90 db, 8 ohm etc..

Now we use 4 of them in one speaker config so you are talking a total of 4 in one cabinet, and 4 other for the other channel..

We wire 2 of the drivers in series getting 16 ohm and than the other pair in that cabinet for another set at 16 ohm, than of course Parallel them together to come back to the full load being 8 ohm...

What is the efficiency on this? Will it go up to like 96 db, or higher?

Just curious..
matrix
Link...note that this is only true if you multiply the box volume by the number of woofers...otherwise you will lose some of the efficiency gain to acoustic damping.
As long as the drivers are located next to each other and reproducing wave lengths large relative to their dimensions, the efficiency increases with 10 log # drivers which is to say it goes up 3dB regardless in a pair regardless of how you wire them.

The configuration just affects the power delivered and therefore sensitivity. Wired in series each of a driver pair gets half the voltage so its output is -6dB down. But the two add in-phase so you get the same output for 0dB change in sensitivity. Since the load was doubled, current and therefore power is halved for the same output so efficiency is +3dB.

Wired in parallel you get the same +6dB in output, but power is +3dB from halving the load and efficiency has increased by the same 3dB.

Four drivers have +6dB efficiency regardless of how they're wired. In series-parallel they present the same 8 Ohm load and sensitivity is +6dB. Wired in parallel the present a 2 Ohm load, consume 4X the power, and sensitivity is +12dB - +6dB from quadrupling the power, and +6dB from efficiency.

Paired with an active cross-over that allows for different sensitivities across drivers and even the frequency band in one driver, a couple of bass drivers at +6dB sensitivity picking up another +6dB from floor mounting work really well.
Drew, seems I am following you but I might be getting a little messy on this so for the example above, my 4 drivers run series-parallel and original single drive at 90 db... Is in fact now a 96 db sensitivity seen by the amplifier? It would at least be seen as 96 db just like a good single full range driver lets say that is 96 db... the Amp would react just the same as it would with a high efficiency single 96 db driver is what I am getting at?

Thanks
>Drew, seems I am following you but I might be getting a little messy on this so for the example above, my 4 drivers run series-parallel and original single drive at 90 db... Is in fact now a 96 db sensitivity seen by the amplifier?

Yes, at wave lengths that are sufficiently long compared to the driver dimensions.

At high frequencies the output doesn't overlap as much due to increasing driver directivity, the phase shift becomes significant, and they don't add in-phase. Arrays built from stacks of 6" ribbon tweeters are no more sensitive than a single unit; they just have more uniform power response.

Jim Griffin's Line Array paper would be worth looking at.