Where do the harmonics come from in OpAmps?

[crazyamd]

Re: Harmonics

that's certainly not right. you can say "slewing" is a large signal phenomena, but not the harmonic distortion. low level signal just like low level injection in device physics is absolutely different from large signal. the point here is to make it clear that harmonic distortion is NOT just a large signal phenomena.

 

[terryssw]

Re: Harmonics

So, what is the results of slewing? This also leads to non-linearity, and thus causing output to have harmonic distortion. This is a great important research topic of linearity in Sigma-delta modulators, that one of the main nonlinearity in sigma-delta modultors is due to slew rate limiting. Slewing, same as harmonics distortion, is referred as large signal behavior since you consider the circuit nonlinearly.

I think you still don't get the correct sense what small signal in electronic engineering means. Actually as lastdance said, You are perfectly correct if you replace "large-signal" by "large-level signal", "small-signal" by "low-level signal" in all your previous few posts. Since in Electronics, small-signal has been already assigned (probably by some famous professor) a special meaning of "the signal level such that the assumption of linear systems is valid".

Again, harmonics distortion appears no matter you consider low-level or large-level signal, but it must be large-signal concept, since small-signal means "you don't consider nonlinearity". Make sure you understand this last sentence. Or, you can say, harmonics distortion is an "any-level" signal's behavior, but it must be a large-signal behavior.

 

[crazyamd]

Re: Harmonics

slewing is a large signal phenomena, which causes distortion too, but this doesn't mean if there's a distortion then it's caused by large signal blah blah...
large signal is large signal, small signal is small signal. they are certainly different although there may not be an absolute value used to distinguish them. no offense, but your problem is that you always tryied to prove your point by quotaing some "famous professors'" words or changing the point of discussion without considering the context. ;-)

remember we are argueing that if "distortion must be a large signal phenomena". but even yourself has agreed that at so called low level signal causes distortion too. in any event: low level signal != large signal.

also, small signal is defined in reference to op point, not in reference to if the circuit is linear. in ac analysis, this leads to the apporximation. this doesn't mean outside this context, all signasl are "large signal".

 

[terryssw]

Re: Harmonics

Yes, I agree nonlinearity will appear also in low-level signal, but I do not agree nonlinearity will appear in small-signal. Your concept and theory is correct, but you just using a "wrong definition" and "wrong word phase" to describe your theory. "small-signal" does not really means the signal is small.



In definition:
"small-signal" = "signal-level relative to the curvature at the operating point such that the assumption of linearity is valid"
"low-level-signal" = "relative low-amplitude signal, but still have nonlinearity since no circuit is perfectly linear"
"large-signal" = "without the assumption of linearity"
"large-level-signal"= "relative large-amplitude signal, but still have nonlinearity since no circuit is perfectly linear"

So nonlinearity is a large-signal behavior. Is it clear for you to understand?

The main reason people define small-signal is not just "small-signal level relative to op point" only, but furhter they need assumption of linearity for easier analysis. If not, why people need to distinguish small-signal and large signal? Why simulator call its AC analysis as small-signal analysis? They called this because AC analysis make ASSUMPTION of LINEARITY.

One more example. I think you have also clear understanding on what noise is. Why do you always say nonlinearity will caused harmonics distortion? Can you say that nonlinearity will cause noise?

In just the meaning of the words, noise just something you unwant. Why can't you call harmonics distortion as noise?

It is because in previous years someone force a definition on the noise, which is only related to random variation on the signal; while distortion is "defined" as signal-dependent. AND these defintion proposed by this "someguy" is widely accepted in the world of electrical engineering. Similar things happen between the word phase of "small-signal" and "low-level signal".

So, you cannot say "noise" = "distortion", and similariy "small-signal" = "low-level signal". Unless you can proposed some strong evidence to support your idea, and widely accepted in the world, then all of us also will "forced" to agree "small-signal" = "low-level signal".

It is not wise to just referred to someone famous, but before you say this, take a look on what those famous really said, understanding them, and give your reason why you argue them. They said these because they have reasons.

By the ways, in some time definition may be different in different people's sense. So don't always say "that's not true" unless you are very sure there is no other definition on the word "large-signal".

 

[Vamsi Mocherla]

Re: Harmonics

Firstly, non linearty is not due to the current equation though it is quadratic in manner. Anything which cannot be given by a single equation is a non-linearity. Distortion in amplifiers occurs due to the imperfections from the square law and the non-linear dependance of transconductance and the current on the square law. So, in ideal cases, there should be no distortion

 

[lastdance]

Re: Harmonics

I think i got my point across.
but the disagreement on large and small signal analysis is still unconcluded yet.
Small signal ANALYSIS is done by linearizing the circuit around its bias point
Large signal ANALYSIS is done by using the model equations, which are not linear.

Anyway, I think there's nothing much to say on this topic, if one insist on his small signal not being small signal analysis, let it be. u can't force him to accept what's generally practiced.

 

[crazyamd]

Re: Harmonics

Distortion can occur due to transient (output unable to catch up with input), output clipping, input signal purity, noise...

not meant to be picky, but i have to say you mixed different concepts together. usually when we talk about nonlinearity, we are talkinig about a property of the circuit. this has nothing to do with the input signal. for instance, a pure sinusoidal input to an gain amplifier, the output will contain higher order harmonics besides the fundamental - because the amp is not linear.

"output unable to catch up with input" is usually referred to "slewing". also, output clipping certainly a nonliear thing, but when we simulate/test linearity, we do not consider this case. you might wanna refer to the commerical datasheet for the measurement and definition of "harmonic distortion".

It's therefore not required that a high level signal drives the input to cause distortion.

What crazyamd mentioned should not be small signal, but rather low-level signal to avoid confusion with the classical literature definition.

you have some of my point. what i was trying to emphasis and argue is that "distortion is NOT just a large signal concept". as we know, large signal means the signal level is comparable or larger than the op point. we dont wanna make people feel that if your input signal is small then you don't need to worry about "distortion". distortion is alway sthere - consider it or not depends on your linearity requirement. again, for clarity, check the commerical datasheet or the measurement standard.

warm regards,

crazy

 

[Vamsi Mocherla]

Re: Harmonics

crazyamd is right.....One should not confuse concepts like slewing as the distortion due to harmonics.....

Importantly, small signal does not directly mean a very small voltage level. If one looks into the concepts of circuits, small signal is a signal which does not considerably effect the bias/DC levels of the circuit. So, when we do small signal analysis,we tie all the DC biases/Supplies to ground because, we assume that they are not affected by the variation in the input signal level

 

[lakeoffire]

Harmonics

Basically, there is no "small signal" in the circuits. All the transistors are model in the "large signal" domain, like the Id-Vgs, or Ic-Vbe. The reason why there is "small-signal" is just a kind of simplification of the analysis on the circuit. When the signal deviation from the operating point is small, the distortion will be quite small, and thus, the circuit can be approximated as a linear circuit, and the analysis can be greatly simplified. However, since the transistor performance is nonlinear, distortion is always there, no matter how small the signal is.

When we talk about small or large signal analysis, it deals with which equations we are going to use. When equation like av=gm*ro is used, it suggests that the linear model has been used, and thus small signal analysis is adopted. In this case, you cannot directly see distortion. That is why in AC simulations in EDA simulations, you can always put you input signal as 1V, and for a gain of 1000, you will get 1000V output, but without distortion being seen.

 

[okawa]

Harmonics

harmonic are mainly caused by the resistor capacitor, and the active elements in the circuit. if the small signal transfer function of the opamp is derived, it will be easy to find how the harmonic occurs. For transfer function h(s)=A*(s+0.001s^2), a signal x(s)=sin(100s) goes through this amp, at the output u will get a signal with frequency s^2. usually large quisient current is good for reducing harmonics, because it can reduce the Ron of MOSFET, then reduce time constant of RC filters, thus put the higher order component of the transfer function to higher frequency.

 

[blue2lucy]

Re: Harmonics

harmonic are mainly caused by the resistor capacitor, and the active elements in the circuit. if the small signal transfer function of the opamp is derived, it will be easy to find how the harmonic occurs. For transfer function h(s)=A*(s+0.001s^2), a signal x(s)=sin(100s) goes through this amp, at the output u will get a signal with frequency s^2. usually large quisient current is good for reducing harmonics, because it can reduce the Ron of MOSFET, then reduce time constant of RC filters, thus put the higher order component of the transfer function to higher frequency.

You really should go over some basic concepts before you post an "answer" for others.

 

[lastdance]

Re: Harmonics

Is harmonic distortion = distortion? YES
Is distortion only harmonic distortion? NO
Distortion could be due to many factors. Distortion is basically what you get compared to what you think you should get ideally. If you look at a mirror and you see your face has got something on it which shouldn't be there, you say the image is distorted, right? The distortion could be due to dirt, lighting, background, cracks....
Try not to be too narrow minded in discussions, criticize when you got the points, not when you think only you are right.
But one thing you are absolutely right though, this thread originated from someone who wanted to know about harmonic distortion. No point mentioning other form of distortion.

 

[layes2]

Re: Harmonics

everything
cap resistor transistor
and so on

 

[sergeyr77]

Re: Harmonics

OpAmp actually is not used alone, and it is function of the feedback network to be linear and linearize transfer function. OpAmp by definition is very nonliner device.
Transfer function of the system depends only on feedback until loop gain is big.
So distortions depend on DC gain, bandwidth (ldecreasing loop gain) and slew rate.