Here's a sketch showing a momentary snapshot of a voice-coil moving through a magnetic field. Where the field is getting stronger the induced voltage is positive, and it's negative where the field weakens. It has to be remembered that the voice coil can't be treated like a resistance followed in series by an inductance, because both properties are spread across the whole length of the coil.
On the graph, the green line represents the voltages along the coil in an open circuit. The voltage is induced by cone movement, and current "wants" to flow from right to left.
The black line shows what would happen if an ordinary amplifier was connected to the speaker. The amplifier controls the voltage at the two ends, but some current flows (from right to left) because a voltage is still induced near the middle of the coil. This means that it's still possible to push the vc., but not as easily as with an open circuit.
The red line shows what would happen if a working NOI amplifier is in control of the voltages at the ends of the vc. It's much harder for something else to move the coil, but not impossible. It's only stable if the system is "imperfect", ie: a small amount of cone movement is permitted so that induced current can be measured and amplified. Induced current flows from right to left, and extra current from the amplifier mostly opposes the unwanted cone movement. Notice that the induced voltage "humps" are now much smaller (meaning that the coil is hardly moving at all), but the total voltage difference from end to end is bigger so there is more current flowing than with an ordinary amp.
CM
On the graph, the green line represents the voltages along the coil in an open circuit. The voltage is induced by cone movement, and current "wants" to flow from right to left.
The black line shows what would happen if an ordinary amplifier was connected to the speaker. The amplifier controls the voltage at the two ends, but some current flows (from right to left) because a voltage is still induced near the middle of the coil. This means that it's still possible to push the vc., but not as easily as with an open circuit.
The red line shows what would happen if a working NOI amplifier is in control of the voltages at the ends of the vc. It's much harder for something else to move the coil, but not impossible. It's only stable if the system is "imperfect", ie: a small amount of cone movement is permitted so that induced current can be measured and amplified. Induced current flows from right to left, and extra current from the amplifier mostly opposes the unwanted cone movement. Notice that the induced voltage "humps" are now much smaller (meaning that the coil is hardly moving at all), but the total voltage difference from end to end is bigger so there is more current flowing than with an ordinary amp.
CM
Attachments
p_a, I'm not sure if I should bother with tweaking the Qtc. I'd rather just bring it very close to 0, producing a nearly perfect 6db/oct differentiator, and fix it by putting an integrator in front. I could then tweak the room gain by adjusting the shelving frequency of the integrator. Blocking DC or other HP filtering would still need to be separate though.
Now the real brain-teaser will be blending the properties of NOI with regular voltage control at higher frequencies. Or, perhaps NOI could still be stable and useful (or at least benign) where there are cone break-ups?
CM
Now the real brain-teaser will be blending the properties of NOI with regular voltage control at higher frequencies. Or, perhaps NOI could still be stable and useful (or at least benign) where there are cone break-ups?
CM
I don't see any reason to go to that extreme. A Qtc of 0.5 would be O.K. for me. It doesn't make sense to "forbid" the system to behave like a 2nd 0rder highpass (at least not to me).
Furthermore, if you overdo the compensation of the driver's Re you will get a very temperature-dependant Qtc.
Regards
Charles
Furthermore, if you overdo the compensation of the driver's Re you will get a very temperature-dependant Qtc.
Regards
Charles
CeramicMan
I want to make an amp with a Negative Output Resistance, but HOW??
very simple, see the attached file
amp_negative_out_resistance.zip
best regards
Petr
I want to make an amp with a Negative Output Resistance, but HOW??
very simple, see the attached file
amp_negative_out_resistance.zip
best regards
Petr
Old-school guitar tube amps use(d) negative feedback of the current signal, to actually raise output impedance. Lows and highs are enhanced due to the speaker impedance curve. Also, the speaker "resonance" is increased. I think this is one of the main reasons tube amps (for the guitar) sound so good. I tried it with one of those National Semi 50W power amp chips. It worked great!
The same project also used a pair of CD "column" speakers, with something like 3-inch speakers in a 60-hz bass reflex cab. The resonant frequency was about 180Hz. I "borrowed" a Yamaha patent from ca. 1991, to use positive feedback of the current signal to get negative output impedance. Equivalent circuit of the speaker looks like a LC tank circuit in series with an 8-Ohm resistor. If you can get the amp output impedance down to near -8 Ohms, the speaker can actually be controlled right through it's resonant peak. Just one issue - too much feedback and it will oscillate! In the end, a small low-Q notch filter smoothed out the residual resonant peak.
This was for Tascam, about 15 years ago. They totally changed it around, basically ruined it, then released the product. It was a real flop in the market. LOL
The same project also used a pair of CD "column" speakers, with something like 3-inch speakers in a 60-hz bass reflex cab. The resonant frequency was about 180Hz. I "borrowed" a Yamaha patent from ca. 1991, to use positive feedback of the current signal to get negative output impedance. Equivalent circuit of the speaker looks like a LC tank circuit in series with an 8-Ohm resistor. If you can get the amp output impedance down to near -8 Ohms, the speaker can actually be controlled right through it's resonant peak. Just one issue - too much feedback and it will oscillate! In the end, a small low-Q notch filter smoothed out the residual resonant peak.
This was for Tascam, about 15 years ago. They totally changed it around, basically ruined it, then released the product. It was a real flop in the market. LOL
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.