A Question on Speaker Driver Efficiency


I have been tweaking my guitar amps, by upgrading the speakers.

I installed a larger speaker (was 8" now 10") in my bass amp, but I made sure it was very efficient - net result
- not only is the bass much deeper sounding,
- but because the new driver was more efficiant I now play at a lower volume.

So I am now considering upgrading my other amp (i.e. used for my 6 string) and got to thinking about building a new cabinet that houses two speakers.

I know that connecting the speakers in ...
- series will double the impedance, i.e. 2 x 4 ohms would have an onverall impedance of 8 ohms
- parallel will halve the impedance, i.e. 2 x 16 ohms would have an onverall impedance of 8 ohms

But what I have not been able to get my head around is...
- what will each connection method (i.e. series or parallel) have on the "combined" sensitivity rating?
- e.g. if both speakers are rated at 96db sensitivity, will the overall sensitivity change due to the connection method or remain at 96db?

Since I can get 4 ohm or 16 ohm drivers - which connection method would be best? series or parallel?

in case it is a factor
- the amp is 15 watts into 8 ohm
- I am looking at employing two identical drivers each rated at 96db sensitivity
- 96 db (or higher) is the target for the combined sensitivity

Any help is appreciated - Many Thanks Steve
williewonka
"Again, current (and not voltage) is what moves the coil." - kijanki

That statement is utterly clueless and contradicts a widely held definition that has been born out in countless measurements of voltage and current in coils. Once again, the definition of voltage as it RELATES to current in any given inductor coil is:

V(t)= Ldi/dt

where  voltage across an inductor at any given time is directly proportional to the time rate of change of current running through said inductor. Voltage and current in an inductor are INEXTRICABLY LINKED to one another and the precise mathematical representation has been given earlier in this thread. If you disagree with the above equation, good for you. You might be on to a new discovery in electronics and physics. Somehow, I doubt it. You could know something that no one else on the planet knows and we're all in for a new, amazing discovery. But then again, you just might be another clueless individual trolling around on the internet in search of a pointless argument....
This bit is problematic. The drivers are going to sound the same whether in parallel or in series. What **won’t** sound the same is the amplifier, which reacts very differently to impedance depending on the amp.
Hi Ralph, 
I believe that I did elude to the fact that it is the amp that sound differently, but I want to stress that in this case, compensation circuitry should sound better with tubes that the speaker without the circuitry.
Using 2 - 16 ohm speakers, depending on frequency, you will see impedances normally rise to 30 or 40 ohms if not higher. You can obviously go into why and amp sounds better on 8 ohms than 4. Impedance compensation circuitry normally only lowers the peaks (stated for others on the thread), it never raises an impedance dip. So when we can make a speaker consistent from say 8 to 9 ohms, in my experience A tubed amplifier putting out consistent power across all frequencies has always sounded better than an amp that sees a speaker that strays radically. Again, this is only my experience, today, I seldom build a speaker without compensation, there are a few exceptions, but Few.
I encourage anyone wanting to follow along, to grab XSim crossover simulator, and examining the Frequency Response, Watts Dissipated and Impedance charts. 

http://www.diyaudio.com/forums/software-tools/259865-xsim-free-crossover-designer.html

The "blank" driver is an ideal 8 Ohm driver with an 80 dB efficiency. This will put to rest any questions about power, efficiency and sensitivity you may have. 

Best,

E