As the voltage going to a speaker increases, the excursion increases, but not in a linear manner; it goes on a curve and eventually hits a wall of speaker bottoming-out and suspension destruction.
So, you have a power rating for how much the voice coil can take instantaneously without boiling its epoxy, and you have a power rating for the same over a period of time, and you have a power rating for how much power it can take without the suspension buying the farm, and how much power it can take without actually reaching the limits for the suspension, and then power ratings for various THD figures at 10%, 1%, .1%, .01% (if you're picky) and then you have a power rating that's just there to "look cool" (1000 watt speakers at the computer store, and they're only 4" paper drivers! Augh!) and I'm thinking, you know, this sucks.
Obviously, you can't overcome the physical limitations, such as the voice coil exiting the gap, or the spider coming unglued, but to some extent, couldn't you use a computer to apply a corrective waveform to fix the distortion?
Computers are powerful enough and cheap enough that I'll bet you could make a computer/crossover/amp with custom software and EQ, and with the power out there, you could probably do EQ at like 1/16 octaves, and then apply your corrective measures....
What does everyone think? Is this a worthwhile thing to pursue? Could I make millions? 😀
So, you have a power rating for how much the voice coil can take instantaneously without boiling its epoxy, and you have a power rating for the same over a period of time, and you have a power rating for how much power it can take without the suspension buying the farm, and how much power it can take without actually reaching the limits for the suspension, and then power ratings for various THD figures at 10%, 1%, .1%, .01% (if you're picky) and then you have a power rating that's just there to "look cool" (1000 watt speakers at the computer store, and they're only 4" paper drivers! Augh!) and I'm thinking, you know, this sucks.
Obviously, you can't overcome the physical limitations, such as the voice coil exiting the gap, or the spider coming unglued, but to some extent, couldn't you use a computer to apply a corrective waveform to fix the distortion?
Computers are powerful enough and cheap enough that I'll bet you could make a computer/crossover/amp with custom software and EQ, and with the power out there, you could probably do EQ at like 1/16 octaves, and then apply your corrective measures....
What does everyone think? Is this a worthwhile thing to pursue? Could I make millions? 😀
I think the Whole World is waiting for such a device.
However, how will the computer predict the
speakers outcome under real world conditions?
Worst Case Scenerio's?
However, how will the computer predict the
speakers outcome under real world conditions?
Worst Case Scenerio's?
It is possible, but it's not trivial. This is called non-linear inverse or mirror filtering. The problem is that speakers are a pretty complex dynamic. First you have to model the linear dynamics of the speaker, which would be easy except for the fact that voice coil resistance depends on temperature. Correcting for displacement related distortion products is pretty easy too. Klippel has a bunch of papers in the JAES on that topic. Modeling eddy currents, flux modulation, and hysterisis ain't so easy, though. I've seen models of each effect in isolation, but never a unifying model--something I started to work on, but this damn work thing keeps getting in the way. Perhaps the hardest thing of all, and probably the one that ultimately kills this good idea, is that so far we have only modeled the system as a one-way causal system: voltage on terminals = diaphragm movement = SPL. The problem is the SPL coming from other sources (other speakers, other drivers in the same speaker, reverberant sound, closing doors, etc. work in the opposite direction: SPL = diaphragm movement = emf. Oh and then you have to worry about the impedance of your amplifier/wire/etc, which can vary considerably from case to case. And so far we haven't worried about the distortion of the amplifier itself.
You probably get the picture by now--it is pretty much impossible to correct for every error in your system using a mirror filter. I've learned the hard way that if you predistort the input signal (mirror filter) to correct one source of distortion without taking into account the other sources of distortion, you end up making things worse off rather than better off. The exception to this case is when one source of distortion is at least a magnitude higher than the other sources. There are two cases where I have seen useful applications in such instances. The most common one is inverse filtering of linear distortion, ie. equalization. The other is throat distortion in compression drivers (due to non linear properties of air at high pressure and high frequencies). Once again, Klippel has some nice papers on the subject in JAES.
As you can see, I've spent some time thinking about this topic and so far the only thing I have to show for it is that I can answer some crazy questions on a diyaudio board. But hey, it is interesting...
John
You probably get the picture by now--it is pretty much impossible to correct for every error in your system using a mirror filter. I've learned the hard way that if you predistort the input signal (mirror filter) to correct one source of distortion without taking into account the other sources of distortion, you end up making things worse off rather than better off. The exception to this case is when one source of distortion is at least a magnitude higher than the other sources. There are two cases where I have seen useful applications in such instances. The most common one is inverse filtering of linear distortion, ie. equalization. The other is throat distortion in compression drivers (due to non linear properties of air at high pressure and high frequencies). Once again, Klippel has some nice papers on the subject in JAES.
As you can see, I've spent some time thinking about this topic and so far the only thing I have to show for it is that I can answer some crazy questions on a diyaudio board. But hey, it is interesting...
John
The answer to your question is feedback control system. You sense the position of the coil and compare it with the input signal. the error is used to apply a correction signal.
Velodyne uses this method, and so does Paradigm. A new company is releasing a feedback subwoofer (not quite as sophisticated as velodyne) for a very attractive price.
http://www.rythmikaudio.com/
I am in no way affiliated with this company, but someone I know has visited the designer and says the product appears well thought out.
Velodyne uses this method, and so does Paradigm. A new company is releasing a feedback subwoofer (not quite as sophisticated as velodyne) for a very attractive price.
http://www.rythmikaudio.com/
I am in no way affiliated with this company, but someone I know has visited the designer and says the product appears well thought out.
there is an electronics australia article on motional feedback
they used a dual voice coil sub,with some circuity,-
reduced harmonics and distortion..
ithink it was a small sealed box..
they used a dual voice coil sub,with some circuity,-
reduced harmonics and distortion..
ithink it was a small sealed box..
- Status
- Not open for further replies.