Not at all and in fact it can be quite the opposite.
More feedback possible? Yes. At least from a Mathematical point of view.
Inherently more stability? no.
Gain is one thing, phase margin and stability a completely different one.
I have the feeling that current capability is somehow more important for good bass than Df.
A big Krell i used to have boasted an unspectacular Df of 60 into 8ohms and despite of this had probably the best bass of any amp i have owned.
A relatively recent experiment with an amp which allowed for variable Df showed that nothing much improves in the bass after Df 30-40 is reached and other frequency ranges suffer.
A big Krell i used to have boasted an unspectacular Df of 60 into 8ohms and despite of this had probably the best bass of any amp i have owned.
A relatively recent experiment with an amp which allowed for variable Df showed that nothing much improves in the bass after Df 30-40 is reached and other frequency ranges suffer.
I believe analog_sa is right. For example, have a look at the Stereophile review of the Parasound jc1. If you look at the circuit, this amp wil have definitely not very high feedback, probably even less than the average ClassAb.
However, in the review they praise the enormous authority in the bass. This amp has enormous current capability in combination with a very stiff PSU.
And I can confirm similar experience in my recent build. 7pair 15A output transistors with a 1kVA Toroidal and quite some capacitance behind it.... WOW.. very good bass, not floppy of what so ever, just fast, controlled..
However, in the review they praise the enormous authority in the bass. This amp has enormous current capability in combination with a very stiff PSU.
And I can confirm similar experience in my recent build. 7pair 15A output transistors with a 1kVA Toroidal and quite some capacitance behind it.... WOW.. very good bass, not floppy of what so ever, just fast, controlled..
And yet every amplifier has enormous amounts of excess gain in the bass (see attachment to post #11), thus enormous amounts of feedback, thus enormous distortion reduction in the bass. If the supply sags, no problem, there's enormous amounts of feedback; the feedback adjusts the amplifier output so it is faithful copy of the input, no matter what the supply is doing.
In my opinion: get a bigger better speaker.
An Eight will not fill a room with big bass authoritatively. Yes, vent (radiator) helps a little. Nearly any modern amp will damp fine (though 4 Ohms on that chip-amp is asking for garble, it runs out of current).
Bass is Air Motion. When moving air, there is no substitute for a "Big Paddle".
After years of Tens and a decade with a Fifteen, I am now back to Eights. My listening tastes have changed. If I was back into Big Bass, I'd throw out some furniture and use a Fifteen or an Eighteen.
Regardless of the speaker it's connected to an amplifiers damping factor is swamped even by the cables that are used. Well that is a well designed amplifier with solid state power devices. I can't imagine an amplifier with high output impedance having good bass control.
An amplifiers ability to deliver current to the load is obviously necessary to control the speaker its driving. But if it only needs 5A to do this and the amplifier can supply that 5A then having the ability to provide 10A, or 100A makes no difference.
What does matter is for the amplifier to be correctly designed and implemented so that it can deliver those 5amps, at the relevant voltage swing, whilst maintaining linearity. Many amplifiers start to fall apart as they deliver more power. You can see this in stereophile reviews. Instead of THD+N decreasing, or remaining constant, with increasing output power, it goes up.
An amplifiers ability to deliver current to the load is obviously necessary to control the speaker its driving. But if it only needs 5A to do this and the amplifier can supply that 5A then having the ability to provide 10A, or 100A makes no difference.
What does matter is for the amplifier to be correctly designed and implemented so that it can deliver those 5amps, at the relevant voltage swing, whilst maintaining linearity. Many amplifiers start to fall apart as they deliver more power. You can see this in stereophile reviews. Instead of THD+N decreasing, or remaining constant, with increasing output power, it goes up.
Wouldn't an amplifier that's been designed for as much gain as possible out to high frequencies, inherently have good phase margin?
Are you thinking of a guitar amp or one used to reproduce signals accurately?
And you probably were meaning only global feedback, local feedback is very commonly employed as the only way to get anything remotely linear out of most active devices when large output swings are involved.
Global negative feedback is an immensely powerful and capable technique to get repeatable high performance amplification.
Thinking about audio, HiFi, amps.
Sound better subjectively, on specification they are worse than high NFB amplifiers (as I wrote).
Global NFB.
As I know, guitar amps are built to have more distortions, probably you too know that.
And I'm not a NO-NFB type of guy, agree with you that NFB is a great tool to get wanted performance of amplifiers and preamplifiers.
Sound better subjectively, on specification they are worse than high NFB amplifiers (as I wrote).
Global NFB.
As I know, guitar amps are built to have more distortions, probably you too know that.
And I'm not a NO-NFB type of guy, agree with you that NFB is a great tool to get wanted performance of amplifiers and preamplifiers.
Last edited:
60/1 is a fairly large damping factor. It means the output impedance is 0.13 ohms, that's going to damp a speaker well.
Remember that a shorted speaker is the most damped it can be, and for an 8 ohm speaker that's 8 ohms of damping. With 0.13 of additional resistance that's 8.13 ohms for damping - not very different.
If the amp has 3 ohms output impedance, then the damping resistance becomes 11 ohms instead of 8, and that will make a noticable difference to the cone behaviour typically. That's a damping factor of 2.7:1
Anything higher than 100:1 is probably unnecessary as speaker leads will contribute more resistance, anything around 30:1 or better is going to be substantially better than any under-damped setup. Negative feedback increases damping factor automatically, its a side-effect, so very large damping factors are commonly seen, but you don't need that much, it comes for free with the increased accuracy of reproduction.
You can short a bass driver's terminals with various values of resistance and hear the response change as the driver is tapped with a soft object, if you want to figure out how much damping is relevant. Electrical damping is aided by mechanical damping (for instance foam inserts in the port of a ported enclosure can provide damping control around the port frequency).
[ most modern bass drivers are designed in the assumption of electrical damping ]
Last edited:
NFB increases accuracy of reproduction only into a dummy load. When real speaker is connected to output, it sends all speaker distortion, coloration, and sound decay into amplifier' input via feedback loop, which muddies the sound.
I agree partially and disagree partially.
Using a resistive dummy load, measuring across the load, the distortion and phase will measure better than measuring distortion and phase across a real speaker. The driver is by far the greater source of distortion. I am not sure why people focus so much on vanishing levels of amplifier distortion when in the listening room the speaker generates several percent distortions.
Many people miss the fact that the speaker connects to the output terminals and the output terminals connect to the amplifier feedback loop. The feedback loop connects back to an inverting stage of the amplifier. So in fact feedback reduces speaker distortion and resonance (damping).
Thanks DT
Wow thats a new one. Feedback sends the "sound decay" back into the amp. No limit to the audiophool imagination.
Speaker is a two-way device, it also acts as a microphone. One may consider a simple gedanken experiment with speaker connected to amplifier with deep global NFB, say 40-60 dB. If we give speaker's cone a mechanical pulse, it will excite multiple resonances that will take some time to decay. Across the spectrum, this decay is represented by waterfall plot. The complex decay signal will feed into amplifier input by the NFB loop, and then will appear amplified at the output terminals. One might say that this feedback-amplified signal will be out of phase with the mechanical decay and will cancel it, but that will be true only for the main speaker resonance. Higher order resonances will not be exactly in opposite phase with the feedback signal due to mechanical delays in the cone/voice coil system. So, NFB is only helpful at main resonance, but it is in fact harmful at higher frequencies, where it acts to amplify speaker distortion.
It depend how we use the compensation. Two pole compensation can have flat open loop gain almost 20kHz and have good phase margin. Many member designed that amplifier: Dadod, LKA, myself, etc.
Negative feedback compares the output to input so anthing that shows up at the out put that isnt at the input is reduced by the loop gain. It dosnt matter where this difference (distortion)comes from, transistors in the amp or this so called speaker distortion. Learn some control theory and read The f-word.
https://linearaudio.net/sites/linearaudio.net/files/volume1bp.pdf
The problem can be at RF but this is easily resolved. This distortion you talk about can be easily measured.
Last edited:
Are we talking about voltage feedback or load current? Combined coupling allows you to get a negative output impedance.
The speaker is an electromechanical system with complex impedance (Z) and counter-emfs (E).
At low frequencies, true feedback is electromechanical. With diffuser acceleration sensor (piezoelectric, optical or microphone).
Velodyne for example.
The speaker is an electromechanical system with complex impedance (Z) and counter-emfs (E).
At low frequencies, true feedback is electromechanical. With diffuser acceleration sensor (piezoelectric, optical or microphone).
Velodyne for example.
No, not RF. With speaker size comparable to acoustic wavelength, mechanical delays and phase shifts are very significant.