What a listener feels is a quite separate issue. A sound is physical and therefore mathematical; whether it is also musical is a matter of human taste and culture. Fourier deals with its physicality: it is a set of tones, however fondly or fervently some people may imagine that it is not.45 said:
Back EMF is not inherently nonlinear. You are now changing the subject, and on the verge of admitting that the nonlinearity is built-in to actual speakers rather than magically arising from the combination of two linear forces as you originally claimed.
Can we agree that:
1. any number of linear forces add to create a linear force
2. in the real world, including loudspeakers, many forces are somewhat nonlinear
3. putting a waveform though a linear filter can give the superficial impression that non-linear distortion has taken place when in fact it has not
So you are now saying that I am right!
Again. I didn't write that!
I never said that comes from the combination of linear forces. I said that separately they are linear but then linearity is lost when they are simultaneously present. Is this a difficult thing to understand?
To achieve a good degree of linearity you MUST control the current, not the voltage.
This comes before you take any step in the routine loudspeaker design.
I can understand people are confused but it is not my problem. It is not because I call something with its historical name that creates confusion because if one knew the subject (not just reading here and there) would know it.....
... or maybe we should concentrate on this exciting new develoment:
Transcendental resistors simplify precision design | EDN
Jan
Transcendental resistors simplify precision design | EDN
Jan
I was hoping to lighten up this thread, and gambling that most are not as smart as you Andrew
Jan
Well one could argue that when you try to communicate and people get confused by what you try to tell them, is is your problem. You are wasting your time.
But let's try to cut through the chase. When you say 'transient' what exactly do you refer to? What is this transient? A transient can cause non-linear distortion, i.e. cause additional frequencies to appear that were not part of the original signal.
Jan
I would like to sift through this book, although I would not expect too much.
A long time ago, a guy named Cherry showed that there are exactly three ways to reduce an amplifier distortion: a) increasing the bias, so that active devices are in a more linear region (that is, dynamic variations around the bias points are smaller, Class A principle) , b) use distortion cancellation (that is, use circuit symmetry wherever possible, to cancel the even harmonics) and c) use negative feedback (including local feedback).
That's all we have, anything else is an illusion.
Sorry for being unclear and you thinking this was argumentative. I should have written my anecdote better. The punchline was that the group turned to me and said "But it's linear!" and thus implying the signal could be easily recovered.
It's a form of positive feedback, not the negative feedback you mentioned.
Actually it could fit in your cat. a). class A, if you adapt the description slightly from 'keeping device current constant' to 'keeping device voltages and currents constant'.
But I am sceptical that Cherry or anybody else can determine the limit to the what the different possibilities are. I don't think you can 'prove' that there is not another concept that is waiting to be found. Sort of 'you cannot prove a negative'.
Jan
Last edited:
No problems, but that is just my point. An NRZ date stream can look totally trashed, no clock recovery, no data, and a simple equalizer can fix it because the system is linear.
Yes, of course there are new frequencies appearing that are not part of the original signal.
For example, just to make it not too complicated, if I have an input signal made of two components that have the same shape in the time domain but different amplitudes and send it thorugh a non-linear system I will get new spectral components, absent in input signal.
However my discussion with DF96 is a little bit different in the sense that the degree of linearity or non-linearity of the system (driver) not only is intrisic but can also depend on the way it is driven (i.e. by controlling the current or the voltage)!
Hmmm... If you accept that all complex non linear transfer functions are complex analytic, then a Taylor series converging around any point exists, and it is unique. You may then want to retain a convenient number of terms, starting with 0 (DC). Then the concept of bias adjustment (increasing the DC term contribution to decrease the overall nonlinearity) and distortion cancellation (by making series terms cancelling themselves, using symmetry) are pretty obvious, and I don't think there's another first order approach than these two.
Feedback linearizes, that's the third way and that's about it.
Yeah, I also can't wait his second edition book and see if his take on TMC in the first edition has changed, and if so, if there's any credit granted. I won't hold my breath 😀.
No. I am saying you are confusing two different things. It is silly to expect that Fourier analysis says something about human appreciation of music, just as it is silly to expect Beethoven to have a view on orthonormal basis sets for an infinite vector space (i.e. Fourier theory).45 said:
Yes, it is an impossibly difficult thing to understand for anyone who knows any mathematics or physics. Someone who is suitably ignorant may find it easier to understand. However, thanks for repeating the statement of your fundamental misunderstanding so that onlookers don't need to refer back to previous posts to check if you really said what you appear to have said.
OK. I made three simple statements of fact and you imply that I am the one who is confused?
No, that is not what we are discussing. It is obvious that the degree of nonlinearity in a system could depend on how it is driven. What we are discussing is your extraordinary claim that two linear forces can produce nonlinearity when both are present. Your argument for this appears to be analogous to saying that a linear filter can produce non-linear distortion of a transient, which Jan is tackling.45 said:
Incidentally, I remember having a brief argument in the letters section of Wireless World about this many years ago. Someone who should have known better said that an input low pass filter on a power amp introduced distortion. At first I thought I must have misunderstood him, but it turned out that despite being a minor UK audio guru he had got this elementary point completely wrong. I suppose this means that other people getting this wrong may (wrongly) consider themselves to be in good company.
- Status
- Not open for further replies.