Speakers work by modifying a magnetic coil inside a magnetic field created by a permanent magnet. The problem is that the field is only based on the incoming signal and the force on the speaker cone is limited by the power of the incoming signal. In the ideal case where the cone and coil are light weight this system works very well, but once you add mass to the system, it gets sloppy. The first part of my idea is to modify the signal to the force needed to move the speaker cone. This would require the electronics to find the rate of slope change of the incoming signal and then adjust the signal so that the required force is applied to accelerate the cone correctly. The second part of my idea is a censor feedback loop that measures the position of the cone and adjusts the correcting circuitry so the motion of the cone matches the incoming signal.

So in a worst case scenario where a square wave signal is fed into a speaker, the coil would not see a square wave, but a modified square wave punctuated by massive spikes at the changes necessary to first accelerate the cone and then decelerate the cone to follow the original square wave signal.
-- MisterQED, Nov 08 2007

Servo Speakers http://www.aes.org/...rowse.cfm?elib=5831
[Wrongfellow, Nov 08 2007]

Servo Speakers 2 http://www.danmarx....ation/servosub.html
[Wrongfellow, Nov 08 2007]

First part is a really good idea [+]

Regarding the //censor feedback loop//, would that be something to record audience responses to beeped portions?
-- vincevincevince, Nov 08 2007


No, it is so the mass of the cone does not have to be programmed into the curcuitry, and accounts for changes in humidity and such that would also affect cone acceleration.
-- MisterQED, Nov 08 2007


re: censor: watch the penny... ... ..... ....... just might drop!

Interesting point you've missed out is accounting for the mass of the microphone's mechanics. The signal from the original recording relates to the response of the microphone to the sound waves, and not to the sound waves themselves.
-- vincevincevince, Nov 08 2007


I think I've seen this idea somewhere before, but I'm not sure. It sounds good (Ha!). But I'm unsure of the possibility of actually doing it.

Let's consider a 20kHz signal. That's zero, up to a peak and back to zero in 0.025mS (half cycle). For the feedback loop to have any success, it ought to be correcting that waveform, let's say, 20 times during its formation. Those are 1.25 uS intervals. Whew. I assume that some kind of laser position sensor would be used.
-- Ling, Nov 08 2007


//Let's consider a 20kHz signal//
Why? I assume we're talking about audio humans can hear, here.
You could use a capacitive effect to measure cone displacement.
-- AbsintheWithoutLeave, Nov 08 2007


Yes, that's right; but it depends on the person (the audio range, not the capacitor).
-- Ling, Nov 08 2007


This is baked - see links.
-- Wrongfellow, Nov 08 2007


d&b have a very nice series of amps and speakers where the amps model the behaviour and, crucially, the coil temperature of the drivers, this enables them to run at very high SPL's without blowing drivers. As far as I know this is done by modelling, not sensing.

PS Hope you've got a good amp, as in order to accurately recreate a square wave, you will require infinite acceleration for an infinitessimally small amount of time.
-- wagster, Nov 08 2007


Baked - Servo speakers, oh well, I thought of this in 1984 so I guess I wasn't to far behind. ~10 years.
-- MisterQED, Nov 09 2007



random, halfbakery