In the 1970's Servo Sound of Belgium marketed a speaker with correction by negative output impedance ( or variant , Professor Korn ) . Philips followed with Motional Feedback Speakers ( similar goals ) . I seem to remember Meridian used it with their active speakers . I think it is claimed to halve box size due to ability to damp resonant peak ( sorry I forget the exact reasons ) .
RCA had a valve amp before 1960 using it .
Belgisches Wunder - Das Servo Sound-System - ..:: Bandmaschinen - Forum ::..
meridian-audio[.info] | Meridian Boothroyd Stuart
Philips Motional Feedback (MFB) loudspeaker
RCA had a valve amp before 1960 using it .
Belgisches Wunder - Das Servo Sound-System - ..:: Bandmaschinen - Forum ::..
meridian-audio[.info] | Meridian Boothroyd Stuart
Philips Motional Feedback (MFB) loudspeaker
The circuit can be made to work.
bentoronto said this in these pages some years ago:
"I don't know if using a second voice coil is more trustworthy than doubling back an error signal from the first voice coil. But I've been using a very well-performing MF system with my Klipschorn for 20 years solid-state and 20 years before that with tubes. It is a bridge design where the speaker VC is one leg of a bridge - not unlike some current feedback circuits.
Can't recommend it highly enough to bass lovers for cleanliness as well as tight bass.
Basic reference was, I think, J. Audio Eng. Soc, maybe 1954, RCA labs.... pre-internet, I know."
Thinking of the circuit in terms of "negative resistance" is not helpful. Instead, if you write the equations for the common mode voltage at the amp in terms of the resistors in the circuit, the blocked impedance of the speaker and its motional emf, it becomes clear that with the right choices of resistors the amp will produce an output voltage such that, Vmotional = constant x Vin - ie the velocity of the cone is proportional to the input signal. Note that this will be low in bass in a sealed box, because the air load is a almost all massive at low frequencies and hence the pressure generated is the proportional to the derivative of cone velocity. The problem is easily fixed by integrating the signal post crossover.
Errors introduced by heating of the voice coil at high power, and voice coil inductance etc are easily modelled. However, non-linearities in the speaker drive are not, so the method only works well when excursions are in the liner region - but is that really a problem for quality sound?
I think the best way to think of this circuit is that it nulls out the DC resistance on all or part of the voice coil/cabling. This effectively reduces the Q whole thing. Again, it is fairly easy to model the effect of a reduced VC DCR on the Q of speaker-in-a-box and if that is done it soon becomes clear that even partial reduction in the apparent DCR is very effective at reducing cone mass/enclosure+suspension resonances. This allows the use of a small box for a big speaker - which is a good thing for many of us.
I would't write it off, but then, like everything, it's got to be done right...
I can post the analysis if anyone's interested.
bentoronto said this in these pages some years ago:
"I don't know if using a second voice coil is more trustworthy than doubling back an error signal from the first voice coil. But I've been using a very well-performing MF system with my Klipschorn for 20 years solid-state and 20 years before that with tubes. It is a bridge design where the speaker VC is one leg of a bridge - not unlike some current feedback circuits.
Can't recommend it highly enough to bass lovers for cleanliness as well as tight bass.
Basic reference was, I think, J. Audio Eng. Soc, maybe 1954, RCA labs.... pre-internet, I know."
Thinking of the circuit in terms of "negative resistance" is not helpful. Instead, if you write the equations for the common mode voltage at the amp in terms of the resistors in the circuit, the blocked impedance of the speaker and its motional emf, it becomes clear that with the right choices of resistors the amp will produce an output voltage such that, Vmotional = constant x Vin - ie the velocity of the cone is proportional to the input signal. Note that this will be low in bass in a sealed box, because the air load is a almost all massive at low frequencies and hence the pressure generated is the proportional to the derivative of cone velocity. The problem is easily fixed by integrating the signal post crossover.
Errors introduced by heating of the voice coil at high power, and voice coil inductance etc are easily modelled. However, non-linearities in the speaker drive are not, so the method only works well when excursions are in the liner region - but is that really a problem for quality sound?
I think the best way to think of this circuit is that it nulls out the DC resistance on all or part of the voice coil/cabling. This effectively reduces the Q whole thing. Again, it is fairly easy to model the effect of a reduced VC DCR on the Q of speaker-in-a-box and if that is done it soon becomes clear that even partial reduction in the apparent DCR is very effective at reducing cone mass/enclosure+suspension resonances. This allows the use of a small box for a big speaker - which is a good thing for many of us.
I would't write it off, but then, like everything, it's got to be done right...
I can post the analysis if anyone's interested.
Do you mean this one ?
http://www.diyaudio.com/forums/subw...l-parameters-electrical-means.html#post724460
Cheers ,
Rens
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.