Does Efficiency Matter?

[Patrick Bateman]

I was reading a thread on this forum today, and the author dismissed a loudspeaker because the efficiency was "only" 95dB. For the past few months I've been doing a lot of loudspeaker measurements, and I've basically reached the point where I don't care at all about efficiency. Here's why:

• There are many high efficiency speakers which sound good, but I personally believe it's because they have low distortion, not high efficiency.
• Distortion and xmax are related; in general, it goes up with xmax. One of the reasons that prosound drivers sound good is that they use drivers with much greater surface area, which lowers excursion.

Basically, I am focused on low distortion and high displacement now. I don't care if a speaker has an efficiency of 80dB or an efficiency of 100dB, as long as distortion and displacement are controlled.

For instance, I have measured $15 computer speakers which can put out 100dB of clean sound with an efficiency of 85dB, because they have high excursion and underhung motors. At the same time, I've measured $100 prosound drivers that struggle to generate 100dB at 100hz, because they're severely displacement limited.

Am I misguided? Or is efficiency really that important?

[chris661]

Originally Posted by Patrick Bateman

For instance, I have measured $15 computer speakers which can put out 100dB of clean sound with an efficiency of 85dB, because they have high excursion and underhung motors. At the same time, I've measured $100 prosound drivers that struggle to generate 100dB at 100hz, because they're severely displacement limited.

Am I misguided? Or is efficiency really that important?

Different applications mean different outputs. When you say clean sound, did you measure this - distortion can appear in strange ways that aren't always noticable.

Have you read this? Power Handling Vs. Efficiency

The point where you're looking for >100dB efficiency is when you either have very little power, or like it extremely loud. For low distortion, high displacement, you'll probably want lots of very large drivers, not moving very far. As you have (correctly, IMHO) stated, distortion increases with Xmax.

I'd say efficiency can be a deciding factor - if you had 2 drivers, one at 82dB 1W/1m, and one that did 92dB/1W/1m, all other things equal, you'd pick the higher efficiency, simply because it means less distortion from your amp.

HTH
Chris

[John Sheerin]

You're ignoring one parameter to focus on others. Why?

What you need to specify is:
-frequency response
-polar response
-max output level
-power compression
-distortion

When you then make a few design choices like transducer type and size, all other requirements can are then defined such as efficiency, excursion, power handling, etc. Or you may find you made a bad choice and cannot arrive at your design goal, so you need to change that first design decision.

For example, I could use a speaker that had very low distortion but was very low efficiency and required large amounts of power to reach a reasonable output level. It would likely have large amounts of power compression. If you compared that to a speaker that reached the same output level with much less power input, it would very likely have less power compression and sound more dynamic. This speaker would be more efficient.

[SunRa]

Ok, my rambling will be totally unscientific here. I've read on this forum at least 2 authors insisting that beyond 103 dB at 1m across the band you will subjectively have a different sound presentation, with everything being more "fast", responsive, more dynamics and so on and so forth.. This seems to be valid no matter the path chosen to reach this target, be it horns or direct radiators in parallel.

disclaimer: deep down I believe this is related to distortion and power compression as well. Maybe this 103db is some kind of a subjective threshold?

[Don Hills]

"Oh no, not again..." - name that movie.

I recall going through all this back in the 1980s, due to the trend of making drivers with better and better frequency response and control of cone modes, which had ever decreasing efficiency due to the damping methods and cone materials used to achieve the performance. The end result was drivers that "couldn't get out of their own way", they needed so much power to achieve dynamic peaks that, even with big Krells etc to drive them, they couldn't reach realistic levels on large scale works. They sounded clean, but not dynamic. There's been a gradual return since then to paper cones, plus the availability of modern exotic fibres.

You appear to be suggesting that provided a driver meets your required distortion and SPL targets, you don't care how much power it requires to do it. But there's more to it than horsepower, as any car enthusiast can tell you. It's not just how fast it can go, but how quickly it gets there.

Applying that to speaker drivers, and assuming the distortion requirement is met, means you should look for high Bl, low moving mass, and large area. This combination is most often available in a (relatively) high efficiency driver.

[Glowbug]

I was reading a thread on this forum today, and the author dismissed a loudspeaker because the efficiency was "only" 95dB.

Sounds like a 45 triode guy who has a taste for headbanging

[nigelwright7557]

Originally Posted by Patrick Bateman

Am I misguided? Or is efficiency really that important?

I run a mobile disco so efficiency is very important.
My amp already weighs a ton to drive the efficient speakers so I really dont want t ohave to double the amps weight to get a decent sound.

[picowallspeaker]

Nigel , I thought you were a postman ,but that red box under your name isn't effectively a post-box !
Cheers!

[nigelwright7557]

Originally Posted by picowallspeaker

Nigel , I thought you were a postman ,but that red box under your name isn't effectively a post-box !
Cheers!

Thats a rear facing 18 inch speaker with forward facing outputs at the top and bottom !

It sounded good but was far too heavy for my disco so I put the driver in a simple box with a port.

[Pano]

Darn - I thought 95dB/W (real) was pretty good for a cone driver.
In an average sized domestic setting, you can get plenty loud with not much power.