30-11-2016, 12:09 PM
(This post was last modified: 30-11-2016, 12:18 PM by Mike Watterson.)
Loudspeakers are very strange loads ...
if you lock voice coil, then it's a rubbish sort of inductor, as the metal will have eddy currents.
If the cone can move, then there is back EMF. This can be quite out of phase depending on baffle, cabinet design / filling, room size / shape / reflectiveness. Oh and crossover network too! A large cabinet loudspeaker is a different load on an open hill top, stage in large hall and a small room. All varies with frequency.
Mostly the assumption is that you should drive from a very low impedance relative to speaker dynamic load. I've wondered what difference it makes:
a) Constant current (i.e. infinite impedance drive). Probably terrible idea.
b) A transducer measuring speech coil movement and putting speaker in a fast enough analogue servo loop (I think this has been tried)
c) Measure sound near listener (decent electret?) and apply DSP. Initialise for room and occupants using maybe white noise and then apply feed forward dynamic transformation. I've a feeling something like this has been tried.
One problem is non-linear behaviour of the cone vs frequency, transients and level.
The basics of the electro-mechanical-acoustic system can in theory be modelled* and stuck in spice as a purely electrical model. You can then examine what is happening on the amplifier output or if you are very expert, at any location in your chosen room.
*Deriving an accurate model may be difficult.
It's amazing what folk with slide rules and hand made parts achieved in 1930s.
if you lock voice coil, then it's a rubbish sort of inductor, as the metal will have eddy currents.
If the cone can move, then there is back EMF. This can be quite out of phase depending on baffle, cabinet design / filling, room size / shape / reflectiveness. Oh and crossover network too! A large cabinet loudspeaker is a different load on an open hill top, stage in large hall and a small room. All varies with frequency.
Mostly the assumption is that you should drive from a very low impedance relative to speaker dynamic load. I've wondered what difference it makes:
a) Constant current (i.e. infinite impedance drive). Probably terrible idea.
b) A transducer measuring speech coil movement and putting speaker in a fast enough analogue servo loop (I think this has been tried)
c) Measure sound near listener (decent electret?) and apply DSP. Initialise for room and occupants using maybe white noise and then apply feed forward dynamic transformation. I've a feeling something like this has been tried.
One problem is non-linear behaviour of the cone vs frequency, transients and level.
The basics of the electro-mechanical-acoustic system can in theory be modelled* and stuck in spice as a purely electrical model. You can then examine what is happening on the amplifier output or if you are very expert, at any location in your chosen room.
*Deriving an accurate model may be difficult.
It's amazing what folk with slide rules and hand made parts achieved in 1930s.
