Hi everyone.
I would like to discuss the following topic with you.
if on the same support portion of the loudspeaker where the coil in which the signal flows there was another electrically separate coil going to the amp or to the preamp or to the active analog line crossover (in differential mode) it could be obtained negative feedback and get equalization? .
I would like to discuss the following topic with you.
if on the same support portion of the loudspeaker where the coil in which the signal flows there was another electrically separate coil going to the amp or to the preamp or to the active analog line crossover (in differential mode) it could be obtained negative feedback and get equalization? .
I tried this 40 years ago the main effect would be induction from the driver coil to the secondary coil I modded mine by putting rf through the secondary coil and then picked this up with a static sense coil in front of the cone
Phillips used a piezo sensor It worked to an extent
Having said that this does nothing for cone resonance etc
Trev
Phillips used a piezo sensor It worked to an extent
Having said that this does nothing for cone resonance etc
Trev
Looks like you're trying to implement some sort of LVDT to sense cone excursion from the loudspeaker. You might get some electromagnetic induction with poor coupling factor, but I guess that would still carry the information you want, with some degree of non-linearity. However, I don't think that would provide any equalisation though.
Adding to Trev's post: In their Motional Feedback Loudspeakers, Philips used a piezo-electric accelerometer to provide information about the speed and direction of the bass cone to the amplifier. https://hifipig.com/philips-motional-feedback-speakers/
A long time ago, there was an AES article about such a set-up. The first coil was current driven for minimal distortion. The voltage from the second coil was corrected for the effect of magnetic coupling and then used to give a bit of motional feedback, just enough to get the fundamental resonance properly damped despite the current drive.
This is not largely different from the impedance (including phase) response of a single voice coil and hence the importance of the classic buffering factor. Some success has been achieved with negative amplifier output impedance that cancels some of the voice coil resistance in order to better sense the cone response. The problem with this and other speaker feedback plans is that the movement sensed can be nodal and local to the sensor location during cone break-up. So, the frequencies that benefit from feedback are limited to the bass where the cone movement is as a single piston without break-up. Even in multi-way systems with electronic crossovers, an inductor in series with the wolfer is a good idea to deliberately avoid voltage buffering at higher frequencies that cause the wolfer to sound "bright", and a shunt capacitor is avoided for the same reason. Some people advocate "current drive", perhaps misguided for similar reasons. The base cone resonance is analogous to parallel resonance and therefore best controlled be a voltage drive vs high frequency resonances that are analogous to series resonance and therefore best current driven. Often the best choice is a ~matching impedance that is a good compromise for all frequency modes. As other have pointed out, it is somewhat pointless to attempt to solve ever response anomaly in detail when the sound will create nodes due to the dimensions of the room and objects in the listening room, unless, you have an anechoic chamber for your sound system.
Those who advocate current drive usually do so because it reduces nonlinear effects such as harmonic distortion, intermodulation distortion and response variations due to changes in voice coil temperature (compression). Once you choose current drive, you need to find some alternative way to damp the fundamental resonance, because the electrical damping is gone. A bit of motional feedback via a second voice coil can do just that, although other types of sensor may be better.
Negative resistance drive aggravates everything that current drive improves.
Negative resistance drive aggravates everything that current drive improves.
Russel Breden described this concept in 1997 Wireless World article: (a) Roaring Subwoofer.
Google for WW+Russel Breden+Motional Feedback. WW is available online from 1950-2000.
Google for WW+Russel Breden+Motional Feedback. WW is available online from 1950-2000.
The Piezo did not work as well as hoped, at least they did not sound as nice as Phillips hoped. Second coil did not prove very effective on subs. Others have tried strain gauges. I came up with an optical system but processing was not fast enough and at the time, I could not get sufficient resolution. ( shutter over a tapered slot) I had it sort of working in analog, but the former did not move as uniformly as I needed to prevent the components touching. Maybe some focused lenses or laser would have worked, but that was not possible then. Today, a laser rangefinder work, but the cost might be a problem. Someone should try.
Easiest on a sub, but we are least sensitive in the bass. We need it in the 1 to 5K.
Easiest on a sub, but we are least sensitive in the bass. We need it in the 1 to 5K.
I've been thinking along similar lines: a small field coil together with a ferrite (or similar) motor that provides saturation. The field coil might use a CCS bias, so the power requirements don't seem like a big issue.
How could this be useful? I'm not sure yet. I'm not sold on negative output impedance, or high damping in general, but it seems feasible to use it in some kind of shunt feedback.
Regarding the different types of drive, the enduring popularity of tube amps suggests that applying little to no braking force on the voice coil may be easier on the ears than applying too much. (Harmonic distortion seems like a red herring to me, as any amount of HD must also produce IMD, so whatever crimes tubes are guilty of, it could be outweighed by issues that high-damping solid state amps introduce that are only measurable with real speaker loads).
One thing I've found on simulations is that even a rudimentary LCR network for the speaker tends to shift the phase of global feedback (or at least voltage feedback that includes the output stage), which can increase THD measurements x10. Now imagine what happens in reality, when it's not just something that can be corrected by flattening the speaker impedance with a few passive components, but reverb with huge random phase shifts from the spider, cone and air are all picked up, and the amplifier attempts to sense it for "negative" feedback to provide a strong correcting current. And the correction current can then inter-modulate with the audio signal. It already seems standard that many amplifiers have reduced stability when dealing with cone break-up modes. Not only is the feedback phase shifted, but the speaker tends to have maximum sensitivity in that range so it produces the highest voltages. By intuition, negative output impedance would seem to make this problem worse.
Too high output impedance might have its own issues like causing weird 'break-in' effects. If a paper cone woofer tends to flex in its upper registers, there's a question of what happens when the speaker is driven at higher voltages at precisely those frequencies where the cone (or spider etc) is already resonant? Excluding other effects like resistance going up from over-heating, does regulating the current equate to driving a speaker 'harder' where it already has natural resonances? And how is that likely to affect Q over time?
Apart from all that, linearising the BL seems like it could be a very worthwhile effort, and doesn't have to be related to the output resistance.
How could this be useful? I'm not sure yet. I'm not sold on negative output impedance, or high damping in general, but it seems feasible to use it in some kind of shunt feedback.
Regarding the different types of drive, the enduring popularity of tube amps suggests that applying little to no braking force on the voice coil may be easier on the ears than applying too much. (Harmonic distortion seems like a red herring to me, as any amount of HD must also produce IMD, so whatever crimes tubes are guilty of, it could be outweighed by issues that high-damping solid state amps introduce that are only measurable with real speaker loads).
One thing I've found on simulations is that even a rudimentary LCR network for the speaker tends to shift the phase of global feedback (or at least voltage feedback that includes the output stage), which can increase THD measurements x10. Now imagine what happens in reality, when it's not just something that can be corrected by flattening the speaker impedance with a few passive components, but reverb with huge random phase shifts from the spider, cone and air are all picked up, and the amplifier attempts to sense it for "negative" feedback to provide a strong correcting current. And the correction current can then inter-modulate with the audio signal. It already seems standard that many amplifiers have reduced stability when dealing with cone break-up modes. Not only is the feedback phase shifted, but the speaker tends to have maximum sensitivity in that range so it produces the highest voltages. By intuition, negative output impedance would seem to make this problem worse.
Too high output impedance might have its own issues like causing weird 'break-in' effects. If a paper cone woofer tends to flex in its upper registers, there's a question of what happens when the speaker is driven at higher voltages at precisely those frequencies where the cone (or spider etc) is already resonant? Excluding other effects like resistance going up from over-heating, does regulating the current equate to driving a speaker 'harder' where it already has natural resonances? And how is that likely to affect Q over time?
Apart from all that, linearising the BL seems like it could be a very worthwhile effort, and doesn't have to be related to the output resistance.
"One thing I've found on simulations is that even a rudimentary LCR network for the speaker tends to shift the phase of global feedback (or at least voltage feedback that includes the output stage), which can increase THD measurements x10. "
If more people would just understand that! If speaker companies would just understand that. Just do a LTSpice model and swap around various speaker models with a resistive load. Not just distortion, but stability. I only know about voltage feedback. Current feedback I am still trying to understand.
A lot of people do not understand Beta droop either and why on earth the shift to 4 Ohm speakers. I would rather see 16. Transistors amplify voltage a lot better than current.
Two servo theories: A stable system you nudge into lock, or a highly unstable you drag into lock. Not easy. Harder when it is a moving garget. All servo will have ripple introduced. How much? Is it self defeating?
If more people would just understand that! If speaker companies would just understand that. Just do a LTSpice model and swap around various speaker models with a resistive load. Not just distortion, but stability. I only know about voltage feedback. Current feedback I am still trying to understand.
A lot of people do not understand Beta droop either and why on earth the shift to 4 Ohm speakers. I would rather see 16. Transistors amplify voltage a lot better than current.
Two servo theories: A stable system you nudge into lock, or a highly unstable you drag into lock. Not easy. Harder when it is a moving garget. All servo will have ripple introduced. How much? Is it self defeating?
Rythmik Subs uses a secondary pickup coil and feedback in the amplifier to control the woofer motion. They call it "Direct Servo" the pickup coil is a much finer winding, as it does not need to carry current. They allow for DIY, they sell the drivers and amplifiers separately so you can roll your own systems.
https://www.rythmikaudio.com/technology.html
https://www.rythmikaudio.com/technology.html
I suspect Arivel speaks some Latin language, where the phrase is written that way.
Remove "one" or replace it by "a" for Germanic language structure (such as English):
at least flat SPL smoothing at all frequencies can be achieved?
or
at least a flat SPL smoothing at all frequencies can be achieved?
better now?
Attached are two articles by Dutch researchers on the topic of Motional Feedback. The first one was written in 1960.
Yes Sir, thank you.
I think sound pressure is proportional to Xmax and the square of the frequency and therefore the Xmax needed to obtain the same SPL at different frequencies could be different. Also, the above may not apply once pistonic motion ceases to exist (breakup).