First cycle distortion - Graham, what is that?

[sreten]

Originally posted by Nelson Pass
I was rather under the impression (and I could be wrong) that
Graham was making reference to distortion created by the
active character of the loudspeaker, not intrinsic distortion of
an amplifier using feedback.

Feedback in power amplifiers is far too fast to create first
cycle distortion due to feedback delay, at least in the context
of audio frequencies. The notion that it takes a cycle or two
for the amp to "get it" is erroneous.

The loudspeaker on the other hand may not exhibit the same
impedance on the first cycle as subsequently due to the back
emf generated by a moving voice coil, and so can draw more
current on the first cycle. It seemed to me that this or
something similar was the basis of Graham's argument.

GM :

In another thread I mentioned distortion of the leading edges of first cycles, and called it 'FCD' = first cycle distortion.

I would not build this amplifier because R1+C1, C3, R10+Q8/C9 and R11+Q9 are a series of first cycle distortion generators. Values and turnovers not known, the delays might be low, but I cannot assume that they are. Stable - yes; accurate - no.

Well you tell me, it all seems smoke and mirrors to me.

A cohesive account of the phenomena is lacking.

The dynamic impedance effects of loudspeakers has been
covered by D.Self using waveforms with discontinuities.

The question is : is FCD a genuine phenomena, or an
alternative viewpoint of known existing phemomena.


sreten.

[PMA]

Per,

have a look here: http://w3.mit.edu/cheever/www/cheever_thesis.pdf

Pavel

[subwo1]

But I'm left wondering the same thing I was left wondering in the previous thread; what's the fundamental difference between the back EMF of the loudspeaker and the back EMF of any other RLC resonant circiut?

When the moving mass of the cone is coupled to the voice coil, the result is an AC generator. How much this active power input to the amplifier's output terminal differs from the nominal inductive phase shift of the voice coil depends on the speaker.

[peranders]

Hm, all the experts are their heads. Has Graham discovered something new?

Originally posted by PMA
Per,
have a look here: http://w3.mit.edu/cheever/www/cheever_thesis.pdf

Heavy report I'm afraid I haven't got the time to read right now.

[sam9]

A-----------------
"Actually, I believe Graham was talking about the amplifier itself, possibly including loading effects. He simulated the transient response to a sine wave in SPICE, then did an FFT on the first cycle to determine the harmonic distortion of just the first cycle."

I would want some assurance that what is being seen isn't just a
computational artifact from SPICE. I have doneof SPICE transient analyses and specified that recording of data does not start until after 1ms and found that the FFT looks better sometimes. I've assumed it had a lot to do with the math and next to nothing to do with circuit performance under normal conditions. Of course, I may be wrong which is nothing new.

B---------------------
"Music is comprised of transients. There is almost - exept when a tone is prolonged on purpose - no steady-state in music. Hit a string on a guitar and it will start swinging as soon as the plectrum or your finger lets go of the string. But the amplitude of the swinging wil not be constant, it will decrease untill it stops. Concluding; No steady state. As you may or may not see/know, the first swing from start point to start point is the one with the biggest amplitude, as it will fade gradually as already sayd.

Here comes the question: What can cause for a signal to be dampened in it first swing - or cycle (as Graham appropiately calls it)?"

This seems to imply that aqmplifier performance during the current cycle is dependant on it's state during the immediately preceeding cycle. "First cycle distortion" would just be a special case of this proposition. If you assume, for the sake of argument' that the slew rate is adequate for the peak voltages implied by the rails, then it seems to me you are arguing for some form of semi-conductor memory. I'm not familiar enough with the relavent literature to know whether this has been investigated or not, but it seems to me this is what the proposition comes down to if you eliminate inadequate slew rate as a mechanism.

[jneutron]

You cannot get accurate spectral analysis results with an FFT of one cycle....That is a horrible window.

Best you can do is record it, and duplicate it in time, making it a pseudo sine wave.

Or, record it, and apply nulling techniques with sine waveforms.


John

[SY]

I'm afraid I haven't got the time to read right now.

Don't bother, it was a terrible piece of work. Amazing he got that thesis passed.

[PMA]

Originally posted by subwo1


When the moving mass of the cone is coupled to the voice coil, the result is an AC generator. How much this active power input to the amplifier's output terminal differs from the nominal inductive phase shift of the voice coil depends on the speaker.

The explanation hereabove is often used in the DIY community. In fact the electrodynamic speaker can be described by the attached schematics. R2 a L2 are the resistance and the inductance of the voice coil in the "braked" state (not moving). R1, L1 and C1 are the components calculated from mechanical side of the speaker to the electrical one. They describe the resonance effect of the speaker. The component values will differ according to the real speaker.

[andy_c]

Originally posted by sam9
I would want some assurance that what is being seen isn't just a
computational artifact from SPICE. I have doneof SPICE transient analyses and specified that recording of data does not start until after 1ms and found that the FFT looks better sometimes. I've assumed it had a lot to do with the math and next to nothing to do with circuit performance under normal conditions.(...)

Actually it's intimately related to the transient response of the circuit. What you're seeing is the need to wait until the transient response settles out until the signal you're doing the FFT of becomes truly periodic. You can show analytically (using Laplace transforms) that for a first order low-pass filter, the response to a pulsed sine wave is an undistorted pulsed sine wave with shifted phase, plus a decaying exponential. Since the FFT of a single cycle is just the spectrum of the periodic extension of that single cycle (provided the sample interval is chosen right), the spectrum is that of the periodic extension of the sum of two signals:

1) The undistorted pulsed, phase shifted sine
2) The periodic extension of the decaying exponential (from the transient response)

This is for a single pole low-pass filter only. The presence of (2) above causes the distortion to appear, since the periodic extension of a decaying exponential looks like a high-pass filtered square wave with every other half-cycle negated. It can be shown analytically that the amplitude of the component (2) above goes to zero when the phase shift of the sine goes to zero. Thus the "first cycle distortion" of the output goes to zero as the bandwidth goes to infinity.

[PMA]

So how about the transient response of the digital filter of the CD player, for example? It has rise time (10% - 90%) no shorter than 17us and limited initial dv/dt, far below slew rate of the contemporary amplifiers. And how about analysis of the transients of the musical instruments itself?