REALLY struggling to grasp op amp theory!

[eti]

So all these "tutorials" we find online about op amps seem to omit entirely or just skim over, WHY op amps always try to make the inputs equal.

I've tried Dave's video, and whilst very thorough, nope, it's not sinking in. Tried another bloke "w2aew", nope, same applies - he appears to dive right into the video, omitting the WHY of op amps wanting to do this. So so confused, maybe it's me having a learning barrier (I think it is).

Funny how in life, 99/100 people explain something using the same kind of thought patterns and process, and you don't get it despite watching the video 10x... then one day you'll find that ONE person who happened to approach it from a more intuitive, non-assuming angle, and it "just clicks"...

... well, if you can find me that one person, I'd hugely appreciate it since this is driving me barmy. Most tutorials say they do this, but not WHY, and a lot of stuff is skipped.


Please help me? Many thanks.

[ebclr]

A seesaw is your answer

"WHY op amps always try to make the inputs equal."

Point 1 ) This is not true, This only happens when you have a negative feedback,  One Ideal op amp have infinite gain, and infinite input impedance

Point 2 ) when you connect with a RESISTOR NEGATIVE FEEDBACK They will adjust the  input voltage until v+ and v- are equal, Because othwersizw the output will / increase / decrease, who will give even more force to make then equal. exactly like a seesaw with someone ( the negative feedback) changing weights to one side from another

 

[eti]

A seesaw is your answer

"WHY op amps always try to make the inputs equal."

Point 1 ) This is not true, This only happens when you have a negative feedback,  One Ideal op amp have infinite gain, and infinite input impedance

Point 2 ) when you connect with a RESISTOR NEGATIVE FEEDBACK They will adjust the  input voltage until v+ and v- are igual, Because othwersizw the output will / increase / decrease, who will give even more force to make then igual. exactly like a seesaw with someone ( the negative feedback) changing weights to one side from another

Thanks, but that's another thing - what IS negative feedback and why? That's something that's also assumed we know.

[ebclr]

https://www.allaboutcircuits.com/textbook/semiconductors/chpt-8/negative-feedback/

[eti]

You're being very kind and patient, I really am very grateful. Thank you

[newbrain]

Maybe this could help visualizing what goes on:

https://youtu.be/HbMnQdRzD8A

[eti]

Maybe this could help visualizing what goes on:

https://youtu.be/HbMnQdRzD8A

"!CLICK!"

Okay, that's the very loud sound of the lightbulb going off in my brain, and as the light illuminates the bigger picture, suddenly ALL is clear! It seems you found the 1% who isn't afraid to explain in a unique and fun way, instead of a dry, dull whiteboard only mumbling.

The man in your video needs to receive a teaching award. Thank you!

[Zucca]

I hate when they just tell V+ must be equal to V- and do not explain why.

[Benta]

This is why mathematics and Kirchhof equations are the basis for analog E-engineering. Not Youtube videos.

[HobGoblyn]

So all these "tutorials" we find online about op amps seem to omit entirely or just skim over, WHY op amps always try to make the inputs equal.

I've tried Dave's video, and whilst very thorough, nope, it's not sinking in. Tried another bloke "w2aew", nope, same applies - he appears to dive right into the video, omitting the WHY of op amps wanting to do this. So so confused, maybe it's me having a learning barrier (I think it is).

Funny how in life, 99/100 people explain something using the same kind of thought patterns and process, and you don't get it despite watching the video 10x... then one day you'll find that ONE person who happened to approach it from a more intuitive, non-assuming angle, and it "just clicks"...

... well, if you can find me that one person, I'd hugely appreciate it since this is driving me barmy. Most tutorials say they do this, but not WHY, and a lot of stuff is skipped.


Please help me? Many thanks.

I had the same experience learning the basics.

I saw many youtube vids and read many books/web pages that explained volts and amps using the first attached pic,  the Amp guy stuck in a pipe, which I'm led to believe works for most people. 

While I sort of got it, it wasn't until I saw a youtube vid which used the second attached pic, that the light bulb went on and I instantly understood.

[chris_leyson]

Control loop theory is why. Simply this https://en.wikipedia.org/wiki/Closed-loop_transfer_function
I struggled with control loop theory because of the pages and pages of complex maths, like not being able to see the wood or forest for the trees. But when you get a light bulb moment it all starts to make sense.
Consider your op amp as a gain block with an open loop gain, G, of say 100dB at DC. A 741 for example. That's a voltage gain of 100,000 at DC. Now, assume you are just using your op amp as a non inverting buffer by feeding the output voltage back into the negative input terminal. In this case H = 1 and the closed loop transfer function G/(1+GH) = 0.9999900... almost 1.0 but there is an error of 0.00001. The op amps input terminals act as a summing node subtracting Vin- from Vin+ and then amplifying the difference or error voltage by the open loop gain G, 100,000 in this example.
To assume that an op amps inputs are equal is a false assumption, it's just a mathematical trick that assumes the op amps gain is infinite. The closed loop transfer function just reduces to 1/H because G is assumed to be infinite. It's good enough for a ball park estimation of closed loop gain but ignores the error voltage that must be present at the summing node. Op amp inputs are close to equal but can never be equal.

[TimFox]

My simple explanation is to start from the requirement that the amplifier output be “finite”, defined here as being within the useful range of output voltage for the device, perhaps between +12V and -12V for a device operating with +15V and -15V supplies.  At DC, assuming a gain of 100,000 V/V, the change in differential input voltage required to swing the output from -12 to +12V would be 24/10⁵ = 0.24 mV.  However, the differential voltage required to force the output to zero is the “input offset voltage”, which might be specified as between +/- 5mV, over the range of parts that meet the spec.  Therefore, that 0.24 mV input range is centered around that offset.
If the feedback works properly, the differential input voltage will meet that requirement.  At higher frequencies, the gain falls off as shown in the data sheet, and the swing in differential input voltage required for a given output swing increases, but it is still centered on the static offset voltage.
The linear circuit analysis is based on that, along with the details of the feedback network.  Non-linear effects past that include current limiting on the output (into the load) and slew-rate limiting on how fast the output voltage can change for high frequencies or fast transitions.

[VooDust]

The trouble stems from the fact that explanations like "opamp wants to do X" are already a condensed generalization. Much like a conclusion, but without the facts leading to it. Like a cookbook. You can choose to simply accept it, and that's fine. In fact, the more I study electronics, the more I'm willing to just accept things. Because else it always leads you down a rabbit hole.

The second problem is, we think the abundance of youtube videos and blog posts about every conceivable bit of knowledge will lead us to understanding eventually. But this is wrong. To really learn something takes a solid base, with a dedicated path laid out, during which you dive into the subject. If only somebody took the time to assemble a complete curriculum which you could work through, helping you out when you got stuck.

Oh wait they did that, it's called a university degree.

On a more practical note, the op amp section in the book "The art of electronics" really furthered my understanding of opamps. To be fair, I didn't have any before!

What I can recommend to you is: Build an opamp yourself on a breadboard with discrete components. You start out simple with two transistors. Hook it up to an oscilloscope and examine their behavior closely. Realize what a puny device this is yet.

Then you improve it with things like the long tail resistor, current source loads, current source emitters, gain stages, etc. For each step, you can verify and visualize the effect the component has on the circuit.

I guarantee you that then you will better understand and appreciate concepts like common mode voltage, stability, transconductance etc. it's a really useful learning exercise.

[eti]

Also, this video is extremely good:

[TimNJ]

I made this video a while back. After watching my own videos many times, it’s hard for me to tell if it’s really any good but people seem to like it. I try to explain the mechanism of feedback in a somewhat intuitive way, though i do tend to babble a little. You can skip to the section on feedback via table of contents.

[Brumby]

.... but that's another thing - what IS negative feedback

It's not actually that difficult

Feedback => taking a signal from the output and feeding it back into the input.  For an Op-amp, there are typically two input connections.

  Negative feedback is where a positive change in the feedback signal causes a negative change in the signal being amplified - and where a negative change in the feedback signal causes a positive change in the signal being amplified.  For an Op-amp, this is typically achieved by applying the feedback signal to the inverting (-) input.
  Positive feedback is where a positive change in the feedback signal causes a positive change in the signal being amplified - and where a negative change in the feedback signal causes a negative change in the signal being amplified.  For an Op-amp, this is typically achieved by applying the feedback signal to the non-inverting (+) input.

and why?

Now THAT is a bigger question.  The reasons are a discussion that can range from straightforward to rather challenging.

[Simon]

I tend to just assume they are black boxes and do this, you can then write whatever math is required to describe the feedback network.

[Doctorandus_P]

Even much more silly is the common phraze of "equal but opposite". Especially when talking about vectors such as forces. It just does not make any sense to me.

For opamps it's simple:
A regular opamp has 2 inputs. The voltage difference between those two inputs is multiplied a gazillion times (that's a lot), and then goes to the output.
That's all there is. The rest (inclusive the exact number of gazillion) are imperfections imposed by the real world.

An opamp never tries to make "it's inputs equal". Whether it seems to do so is entirely dependent on the surrounding circuit.

When analyzing a circuit in which an opamp is used, make use of another imperfection of an opamp called "slew rate". Slew rate means there is a maximum speed at which the voltage of the output of an opamp can change. If the Non-Inverting "+" input has a voltage higher then the inverting input, the output voltage goes up, if the "-" input is higher, then the output voltage goes down.

For more about Opamps, go fetch yourself a copy of "SLOD006B", a.k.a: "Opamps for everyone". For example this copy:
http://web.mit.edu/6.101/www/reference/op_amps_everyone.pdf

Go on, fetch.
It's free.

[rstofer]

For opamps it's simple:
A regular opamp has 2 inputs. The voltage difference between those two inputs is multiplied a gazillion times (that's a lot), and then goes to the output.
That's all there is. The rest (inclusive the exact number of gazillion) are imperfections imposed by the real world.

And that is exactly the reason why we use negative feedback.  We need to redefine 'gazillion' as something practical because a gazillion times the voltage difference will result in an output voltage that exceeds the power supply - among other things.

Maybe we just want a gain of 1 or 10 or 100.  All we need to do is select the proper input and feedback resistors and we can redefine gazillion to suit our needs.

There are MANY tutorials on the Internet and the "Op Amps For Everyone" book, linked above, is excellent.  Pay particular attention to chapter 4 where they deal with combined offset and scaling.  Like if you wanted to spread a 12V battery voltage which might vary from 10 to 14 volts over the 0-5V range of a uC ADC.  This comes up all the time and there are formal methods for dealing with it.  Chapter 4!

[rstofer]

You can read about op amps all you want but you will actually 'learn' op amps on a breadboard.

Here is an op amp playlist for w2aew's videos:
https://www.youtube.com/playlist?list=PLBCjWUUpRpOeAFKEPvkys15YvOt86qGnU

#79 might be helpful in choosing an op amp to play with.  Single supply, dual supply, 5V, +-15V, +-5V, etc.

Today, the emphasis is on single supply low voltage devices.  Your choice!

The easiest to play with are the older devices with a + and - 15V supply.  That's why the 741 is still around.

Here's a simple power supply:

https://www.jameco.com/z/PD-2515-MEAN-WELL-Power-Supply-Dual-Output-Open-Frame-15V-1A-Negative-15V-1A-24W_2100857.html

You could probably use a couple of 12V batteries or even a couple of 9V batteries for some devices but this obviously reduces the output swing.

Advanced experiment:  Take a dual rail op amp and use a 1M ohm input resistor and a 1 ufd feedback capacitor.  Experiment with  input signals and you will see where this circuit is an integrator for very low frequency signals (change the resistor and capacitor to change the integration time period).  Tie a few integrators together and you can control just about anything.  The integrator is used in analog computers and many types of control systems.  My favorite circuit.  While you're at it, create a multiple input summer and a unity gain inverter.  Voila', you have an analog computer.

http://www.analogmuseum.org/english/homebrew/vogel/

I built that computer a few years back:

[eti]

I spent an hour in Photoshop etc making an animated GIF to illustrate the concept so simply conveyed in this superb video:




... so I hope I have clarified this, and got it right, for those who want a quick, visual representation:

[Simon]

I would have been neat had he put holes in the ruler or some other stick and put it on a pivot to demonstrate the gain effect which would work in the same way as a lever. As you mover the pivot point you get a different gain.

[vk6zgo]

And, after they learn all this, someone will design a "radio" with LC networks at input & output, & wonder
why it oscillates!

[gcewing]

I would have been neat had he put holes in the ruler or some other stick and put it on a pivot to demonstrate the gain effect which would work in the same way as a lever.

But one of the points he's making is that there *isn't* anything equivalent to a pivot at that point in the circuit. The only reason it stays put is that the output is being actively adjusted.

[Simon]

I would have been neat had he put holes in the ruler or some other stick and put it on a pivot to demonstrate the gain effect which would work in the same way as a lever.

But one of the points he's making is that there *isn't* anything equivalent to a pivot at that point in the circuit. The only reason it stays put is that the output is being actively adjusted.

It's a graphical repesentation.