Regarding linear and nonlinear operation of devices

[kenambo]

Hi,

i have a doubt regarding linear and non-linear operation of devices.

For example i am having a 2 stage amplifier. and i am a giving small signal input of say 10mV and my first stage converts it into say 100mV.So My first stage acts as a linear amplifier because of the small signal input.

Now the 100mV output is the input of my second stage. so my second stage should act as a non linear amplifier. Isn't it?
and any effects because of this.?

And the circuit simulation tool response is different for these two amplifiers?

thanks

 

[johnjoe]

Just check the transfer function of your second amp: in the case of linear operation: Vout=Av*Vin, Av small signal gain. If Vout differs from your calculated Vout, you're in the nonlinear region. The results of nonlinear region are for example, nonlinear transfer function , increased harmonics and so on. Maybe tell more about your both amplifiers and your simulation results.

 

[kenambo]

Hi..

Yeah we can . But due to supply voltage limitation the output will be saturated and we cant find the exact Av* Vin and the ac analysis in cadence also takes small signal gain into account.

Is it possible to analyse the harmonics of the circuit in cadence virtuoso?

thanks

 

[johnjoe]

If your output signal of the second amp already reaches supply voltage limitation you're already in nonlinear region. But you have to post some specs of your amps for further help.

To simulate large signal behavior, you should run some harmonic balance simulation, also to plot harmonics.
https://www.cadence.com/Community/b...03/18/setting-up-harmonic-balance-part-1.aspx

 

[kenambo]

Thanks for the reply..

i will check and tell you the specs soon.

 

[yuvan]

Kenambo,

First are you checking an comparator or opamp. If opamp, I wonder why you would want to check it in open loop. And if you are checking for a small signal operation, considering the second stage is NMOS CS (for eg), then the gate voltage of NMOS CS cannot change lineraily, since the current in the second CS is fixed. If so, then you cannot actually split into linear and nonlinear model as split above.

 

[kenambo]

hi Yuvan,

The circuit i proposed is neither a comparator nor an op-amp. And i am just checking the concept of linear and non-linear with two single stage amplifiers which are cascaded.

And if the definition for Linear and non-linear holds, then, what i am asking is correct.. isn't it?

If so, how good the circuit simulators like cadence virtuoso, takes this effect into account for the output calculation.

And i am analyzing the answer of jhonjoe mentioned above.

So as far as my design is concerned we can split it into linear and non-linear.. am i right?

thanks

 

[LvW]

So as far as my design is concerned we can split it into linear and non-linear.. am i right?

No - you are not.
In reality, there are no "linear" circuits. All parts and, thus, all circuits have non-linear input-output relationships.
The main difference is if the degree of non-linearity - for a particular application - can be neglected or not.
Now - as far as circuit simulation is concerned, it is important if the model of the particular part contains some non-linearities.
If this is the case, you have to discriminate between two fundamental different analyses:
* Analyses in the time domain (Transient analyses) take these non-linearities into consideration (example: limitations due to power supply rails).
This applies even in the case you are not able to observe these non-lineraities on the screen.
* Analyses in the frequency domain (AC analyses) are small-signal analyses, which means that the program linearizes the circuit around the corresponding DC bias point.
As a consequence, it does not matter at all if the input signal is 1mV or 1kV. The input-output relationship is independent on the actual signal levels. It is just the frequency which matters.

 

[yuvan]

Yes, I agree with LuW. You cannot actually split the operation into linear and non-linear one. And if you are checking your harmonic behavior of your open loop amplifier then I would prefer both the analysis to be carried out. (and the FFT analysis is a must for these open loop amplifier choose your frequency such that 3rd harmonics fall inband)

 

[kenambo]

If this is the case, you have to discriminate between two fundamental different analyses:
* Analyses in the time domain (Transient analyses) take these non-linearities into consideration (example: limitations due to power supply rails).
This applies even in the case you are not able to observe these non-lineraities on the screen.
* Analyses in the frequency domain (AC analyses) are small-signal analyses, which means that the program linearizes the circuit around the corresponding DC bias point.
As a consequence, it does not matter at all if the input signal is 1mV or 1kV. The input-output relationship is independent on the actual signal levels. It is just the frequency which matters.

Thanks for the clarification..

i just came to know the linearisation about a biasing point.

so the degree of non-linearity depends on the frequency not on the amplitude of the signal?
if so.. the mos models small-signal model and large signal model also use the same concept?

thanks

- - - Updated - - -

Yes, I agree with LuW. You cannot actually split the operation into linear and non-linear one. And if you are checking your harmonic behavior of your open loop amplifier then I would prefer both the analysis to be carried out. (and the FFT analysis is a must for these open loop amplifier choose your frequency such that 3rd harmonics fall inband)

could you please elaborate the FFt analysis and the frequency of 3rd harmonics fall inband?

what is inband here ?

thanks

 

[LvW]

so the degree of non-linearity depends on the frequency not on the amplitude of the signal?
if so.. the mos models small-signal model and large signal model also use the same concept?

Who said this?
I repeat: The AC analysis is a LINEAR analysis. That means: It is NOT sensitive to amplitudes. Its only purpose is to reveal the frequency dependence of the circuit.

 

[kenambo]

thanks for the reply..

Is it necessary to do a harmonic analysis while designing amplifiers or op-amps?

 

[johnjoe]

Is it necessary to do a harmonic analysis while designing amplifiers or op-amps?

If you want to know the "large signal" or nonlinear behavior, sure! ac-analysis is a small-signal analysis.