Operational Amplifier

Papabravo

Joined Feb 24, 2006
16,107
Operational amplifiers in and of themselves don't have to satisfy any conditions to be linear. If the application allows it, they will be linear. If it does not allow them to be linear, they won't be. What are some examples:
  1. If the amplifier is in an open loop configuration the output will likely be as close to the rails as it can possibly get due to the high gain and the internal construction.
  2. If the amplifier is configured to perform a non linear task such as being a limiter. This is a non-linear transfer function.
  3. If the amplifier is configured as a multiplier. This is also a non-linear function.
  4. If the amplifier is configured as a low pass filter. This is a linear function.
  5. If the amplifier is configured as a difference amplifier, this is also a linear function.
Does this convey some feel for the situation?
 

Thread Starter

nick0210

Joined May 9, 2021
12
Operational amplifiers in and of themselves don't have to satisfy any conditions to be linear. If the application allows it, they will be linear. If it does not allow them to be linear, they won't be. What are some examples:
  1. If the amplifier is in an open loop configuration the output will likely be as close to the rails as it can possibly get due to the high gain and the internal construction.
  2. If the amplifier is configured to perform a non linear task such as being a limiter. This is a non-linear transfer function.
  3. If the amplifier is configured as a multiplier. This is also a non-linear function.
  4. If the amplifier is configured as a low pass filter. This is a linear function.
  5. If the amplifier is configured as a difference amplifier, this is also a linear function.
Does this convey some feel for the situation?
Yes this information helps. I will do some more homework. Thank you :)
 

crutschow

Joined Mar 14, 2008
27,190
There are some op amps that have an internal configuration designed to minimize audio distortion (such a crossover at the output) e.g. NE5532, but generally an op amp's circuit linearity is determined by the external circuit, as PB stated.
 

MrAl

Joined Jun 17, 2014
8,241
An LM358 dial or LM324 quad opamp are not linear because they are designed for low power supply idle current therefore they have crossover distortion, noise and a poor high frequency response.
Hi,
Yeah but if we are going to start knocking op amps because they are "not linear" then i think we can say that there is no op amp that is linear because it will always have some distortion and some noise and a frequency limitation. In fact if we are going that far then we can say that nothing physical on earth is linear.
Also keep in mind that the crossover distortion goes away when the op amp is used in a way that avoids cross over distortion. All op amps have some frequency limitation we dont use a 10 dollar op amp on a 2 dollar project, and noise is not always an issue again depending on application. It is an industry standard part made by many manufacturers. It also avoids some of the problems with using the ua741.
We could make a list of the pros and cons.

Many op amps have their particular niche they fit into well and those are no exceptions, but i do know how much you hate those op amps :)

I think the only way these devices will go away forever is when nano technology and quantum technology finally take over everything and that is inevitable.
 
Last edited:

MrAl

Joined Jun 17, 2014
8,241
What are the conditions an operational amplifier must satisfy to be linear?
Do you mean the op amp itself or the circuit the op amp is put into?
The circuit would include other things like resistors and possibly capacitors and inductors, maybe even diodes, and of course a power supply.

This is most likely about the circuit but i could also see it being about the op amp and it's internal operation because we can in theory construct an op amp out of devices that are all linear themselves.
 

Audioguru again

Joined Oct 21, 2019
3,173
There are hundreds or maybe thousands of opamps available. The LM358 and LM324 opamps are the only ones I know that have severe crossover distortion and lots of noise. A few other opamps have their limited frequency response.

I have replaced LM358 dual opamps with an MC33172 and replaced LM324 quad opamps with an MC33174. They have NO crossover distortion, much wider 30kHz power bandwidth, same low idle current, same input common mode to the negative supply or ground, same minimum supply of 3V and a typical noise spec that seems lower than the no noise specs for the LM358 and LM324 opamps. The MC33172 is made in only surface-mount packages today.
 

MrAl

Joined Jun 17, 2014
8,241
There are hundreds or maybe thousands of opamps available. The LM358 and LM324 opamps are the only ones I know that have severe crossover distortion and lots of noise. A few other opamps have their limited frequency response.

I have replaced LM358 dual opamps with an MC33172 and replaced LM324 quad opamps with an MC33174. They have NO crossover distortion, much wider 30kHz power bandwidth, same low idle current, same input common mode to the negative supply or ground, same minimum supply of 3V and a typical noise spec that seems lower than the no noise specs for the LM358 and LM324 opamps. The MC33172 is made in only surface-mount packages today.
As far as the -72 model, that's not a bad choice. A little bit more expensive in 1 piece quantities, but about 4 times more in large quantities that could be a problem but maybe you can find cheaper i didnt want to look everywhere for pricing.
I wonder if this part is going to go obsolete though i see a lot of the packages obsolete already even though they are SMD.

There are some other candidates too by TI.

I've used the LM358 in so many projects both as a hobby and professionally that anyone that says that they are no good has to be ignoring many industrial and military applications that dont care about noise or even crossover distortion or even bandwidth. If the part fits the job use it, especially if it is a tried and proven part over years of being available to manufacturers and hobbyists alike.
I like to talk to you about it though because i know you hate it and it is funny to hear what you have to say about it each time it is brought up in a forum :) But maybe you are more geared toward audio exclusively.

There was a project that came up where i designed for the LM358 but decided later i wanted much faster response. It was a DC power supply. I wanted very quick reaction to current and voltage and load changes. I went to a more expensive part about 10 times the price but it was pin for pin and voltage supply and current compatible and 10 times faster. Cant remember the part number now (National Semi at the time) though but i loved that part just wish it was cheaper.
 

RBR1317

Joined Nov 13, 2010
631
An operational amplifier will be linear if the output has a linear relationship to its input. This statement is simple, exact, and not actually very useful. Op-amps are normally used in circuit configurations where the high open-loop gain is used to minimize non-linear characteristics. Properties of an op-amp that minimize circuit non-linearity are high open-loop gain, low input offset voltage, high input impedance, low output impedance, and high frequency response for low dynamic distortion.
 

Papabravo

Joined Feb 24, 2006
16,107
An operational amplifier will be linear if the output has a linear relationship to its input. This statement is simple, exact, and not actually very useful. Op-amps are normally used in circuit configurations where the high open-loop gain is used to minimize non-linear characteristics. Properties of an op-amp that minimize circuit non-linearity are high open-loop gain, low input offset voltage, high input impedance, low output impedance, and high frequency response for low dynamic distortion.
The high gain and the feedback make operational amplifiers a great deal more useful than they might be otherwise.
 

Audioguru again

Joined Oct 21, 2019
3,173
The crossover distortion is shown when the gain of an LM358 or LM324 is only one so it has as much negative feedback as is possible. The crossover distortion must be much more at higher frequencies and at higher gain.
 

MrAl

Joined Jun 17, 2014
8,241
An operational amplifier will be linear if the output has a linear relationship to its input. This statement is simple, exact, and not actually very useful. Op-amps are normally used in circuit configurations where the high open-loop gain is used to minimize non-linear characteristics. Properties of an op-amp that minimize circuit non-linearity are high open-loop gain, low input offset voltage, high input impedance, low output impedance, and high frequency response for low dynamic distortion.
The main required feature is "negative feedback" that's the key point for linear operation.
 

MrAl

Joined Jun 17, 2014
8,241
Negative feedback is not a feature of the op-amp, but rather of the circuit in which the op-amp is used.
Yes, and that is why i said negative feedback is a requirement for linear operation. You cant operate an op amp in linear mode without negative feedback.
 

RBR1317

Joined Nov 13, 2010
631
Yes, and that is why i said negative feedback is a requirement for linear operation.
Of course you are correct. However, the original question was: "What are the conditions an operational amplifier must satisfy to be linear?" Whereas negative feedback is a condition the external circuit must satisfy. Maybe that is splitting hairs, but it's what I do.
 

MrAl

Joined Jun 17, 2014
8,241
Of course you are correct. However, the original question was: "What are the conditions an operational amplifier must satisfy to be linear?" Whereas negative feedback is a condition the external circuit must satisfy. Maybe that is splitting hairs, but it's what I do.
Hi again,

Well i guess i thought it was strange since for one thing i already mentioned that previously in this thread.

But you know what else is strange, or rather perhaps just misspoken, is that there is no op amp that is linear in it's own right, therefore we might read "op amp" as "op amp circuit". Op amps are inherently non linear only the op amp circuit can be linear. The op amp can be used in linear mode, but that requires negative feedback.

This all leads me to believe that the right answer is "negative feedback".
Note that in saying that we dont even consider the effects of the power supply, such as when the input goes too high for the circuit to handle and so the very output of the op amp (and thus the circuit) gets pinned to one of the power supply rail voltages (like +15v or -15v). That again puts it in the non linear region, but we dont mention that in passing either because if we are going to explain all of that then we have to explain every possible operating conditions for the circuit. We also dont mention temperature effects.

A rationale for thinking in this way is that the question was so simple it probably is trying to invoke the most simple answer which i believe is "negative feedback".

Note that it is harder to tell things like this for sure on the internet because we dont have anyone we can immediately question to get more information from if the OP does not respond well. This will lead to a host of different viewpoints for some questions. Even simple back and forth conversation over pure text gets mixed up sometimes like this. All we can do is wait for more information to come in and until then we are left with our own viewpoints as to how it should be answered and what the best answer should be.
 
Top