If he means that the amp will take care of the back emf because it senses it thru the feedback and corrects for it, because thats what feedback does Then yes. Otherwise he doesnt understand feedback either.
Many statements are based on the basis of looking at an amplifier as an isolated device. Regardless of that, an amp is to be seen in real life, in it's surroundings. These are of given by the devices in that chain before and after it. E.g. many speakers sold are way too small to deliver on what is to be perceived as a good copy of the real thing.
A euphonic amp might very well deliver more "weight" to the music and hence might be considered a better amplifier.
Once you take an amp out of it's environment and put it on a test bench, these things are out of scope. I guess that's one of the main reasons why we're getting nowhere on these discussions.
"When the oversimplification of the thinking of some technicians replaces the curiosity of the one who presides over scientific research, and the religious faith in what they have been taught replaces in their way of thinking the doubt which is the basis fundamental to the scientific process, we see every day the level of stupidity and self-satisfaction in which humanity is gradually sinking."
Clear cut and content-free. Funny how this can go hand in hand.
Jan
No surprise that the author of these words felt particularly targeted by those he quoted.
No surprise either that our self-proclaimed vending machine of academic good points doesn't find the content that he doesn't perceive.
Let us consider a piece of piano music. The music consists of waveforms of different frequencies being switched on and off with the condition that more than two waveforms may be emitted at the same time and the T_on period is variable to suit the music piece.
Mathematical analysis of these switching piano notes is the mathematics of pulses which, according to a post in this forum, requires and infinite bandwitdh. This means, the music seems to require an infinite bandwidth for perfect reproduction. However, we have to keep in mind, neither human hearing can boast of an infinite bandwidth, yet, it deals with lone piano notes without the least of issues.
Can this requirement of lone pulses for an infinite bandwidth explain why different amplifiers may give the impression of different sound reproduction?
I have always had some difficulty understanding how the transition from a pulse to a pulse train from the perspective of Fourier Analysis. Strictly speaking, every wave train/pulse in the real world is a pulse, albeit many are complex when we allow for repetitive waveforms.
Mathematical analysis of these switching piano notes is the mathematics of pulses which, according to a post in this forum, requires and infinite bandwitdh. This means, the music seems to require an infinite bandwidth for perfect reproduction. However, we have to keep in mind, neither human hearing can boast of an infinite bandwidth, yet, it deals with lone piano notes without the least of issues.
Can this requirement of lone pulses for an infinite bandwidth explain why different amplifiers may give the impression of different sound reproduction?
I have always had some difficulty understanding how the transition from a pulse to a pulse train from the perspective of Fourier Analysis. Strictly speaking, every wave train/pulse in the real world is a pulse, albeit many are complex when we allow for repetitive waveforms.
The mistake is that musical notes don't switch on and off. They grow and fade out. So there is no infinite bandwidth involved.
Jan
Jan
I don't mind where people's beliefs are greater than their basic understandings. Just the few who don't want to learn...
Is it over simplistic to remind people that an amplifier doesn't know anything about waveforms, however complex they may be, but only a single voltage at any one moment?
What I dont understand is how some people dont think error correction circuits dont correct errors. Do the math, your guesses are wrong.
And all it has to do is multiply that single input voltage and output one single voltage, and this all happens almost instantly (with insignificant delay) including the feedback. There is no delay in the feedback, this is what people need to be reminded of
Let us do a thought experiment and imagine we are an amplifier with an inverting input and a non inverting input, both having a very large gain. Let us say, the gain is 100,000. Now, let us assume at our non-inverting input we have a signal of 1mV. Without considering the other input, the output would be 0.001x100,000 = 100V. Let us assume we have a fraction of the output arriving at our inverting input. Assuming this fraction is 1/100 of the output, we will have to use a signal of 1V + 0.001V at the non-inverting input to compensate for the 1V at the inverting input and get the same output as before. With the fraction of voltage arriving at the inverting input from the output, the gain is now:
100,000*(1.001 - 1)/1.001
= 100/1.001 ~= 99.9
As an exercise let us assume the raw gain drops to 10,000 at 10kHz.
Using the same calculation as above, also for an output of 100V, we need a non-inverting signal of 1V + 10mV = 1.01V.
Let us use the same expression above to get the gain for this reduced open loop gain at 10kHz.
10,000*(1.01 - 1)/1.01
= 100/1.01 ~= 99.0
So, a drastic drop of 90% of the open loop gain, results only in 1% loss of gain with negative feedback.
As you can see, this is almost 1/negative_feedback_fraction = 1/0.01 = 100
The above calculations illustrate why negative feedback is so often put to use by electronic engineers and technicians.
100,000*(1.001 - 1)/1.001
= 100/1.001 ~= 99.9
As an exercise let us assume the raw gain drops to 10,000 at 10kHz.
Using the same calculation as above, also for an output of 100V, we need a non-inverting signal of 1V + 10mV = 1.01V.
Let us use the same expression above to get the gain for this reduced open loop gain at 10kHz.
10,000*(1.01 - 1)/1.01
= 100/1.01 ~= 99.0
So, a drastic drop of 90% of the open loop gain, results only in 1% loss of gain with negative feedback.
As you can see, this is almost 1/negative_feedback_fraction = 1/0.01 = 100
The above calculations illustrate why negative feedback is so often put to use by electronic engineers and technicians.
Last edited:
Yawn. Anyone who's ever heard of NFB knows the basics. Why repeat them? Are they conducive to good sound?
Given the exceedingly limited number of recordings that have no GNFB if you believe feeback is bad you are pretty stuffed for listening material.
This goes along with the 100's of recordings mixed through several dozen 5534's. Seems like the badness wears off with time.