I would pay for an EE Fundamentals course hosted by you that covers things like this. This is excellent -- it goes just beyond what I learned in college, where all the traps for young players are conveniently swept under the rug.
Thank you, Dave. I'm a good ways into my electronics engineering degree, and tomorrow I'll be examinated in either MOSFETs/SMPS/Op-amp errors, and this video provided a great refresher as well as a few golden nuggets! :)
I know this is an old video, but I came across something interesting that could help if this issue ever comes up in the future. Art of Electronics talks about making an integrator as a good way to measure offset voltage and bias current. You ground the input to measure the combined total, then float the input to measure just the bias current. Subtract to find the offset voltage.
Dave, I am very, very grateful for what you are doing. I can find here the easiest explanations of the problems which otherwise I would be searching for days studying books. Thank you very much!
Really wonderful job man. I have been thinking a lot about building a high precission multipurpose dso and your videos are just pure gold. I ll definitely send you a prototype to demonstrate in one of your videos. Thanks again.
Dave, I watched one of your earlier videos and then this one, and I have to say that you have improved alot on your teaching technique! Not saying anything bad about your earlier stuff. Have you considered teaching at a university? You've got the touch!
Hi Dave! Thank you for these videos, helps me reinforce what I'm learning at school. Would love to see one of those in-depth videos regarding the internal circuitry of op-amps if you could find the time! Greetings from (Northern) NY!
i fcking love this guy, he's entertaining and he gives me good grades. I don't wanna go to school anymore i just wanna learn on the internet from people like you
At 11:04, I don't see how RB tacked on to the inverted input then ground should be set to R1. I definitely don't see how it should be RF//R1. I don't see it in this example. Maybe if you change the non-inverted input to have some current coming through R1 instead of tying it to ground. Makes more sense for it to be RF as hashed out on the white board.
Another good one. Wish you had taught some of the analog classes at my college. This stuff might help me out at work some day, like a lot of your other vids!
1. There are part in your computer that are probably even more complicated than opamps - I would recommend to study processing of the signal coming out of your hard disk headers. 2. Opamps have more transistors in one chip which means they use current sources instead of resistors and differential and push - pull stages instead of simple common emitter circuits - this two design principles gives them huge advantage over anything you are able build from discrete components.
I'm pretty sure you messed up the calculation of the output voltage error caused by the input bias current. The current flows through your feedback resistor and causes a voltage of U=RI=3µV. But this voltage doesn't get amplified because it's on the output of the OpAmp. Compared to the error caused by the input offset voltage I would consider this effect negligible.
My remedy for this with my application is to use an NPN or PNP transistor with opamp, where the base is responsive to neg or pos signal otherwise you get zero output on transistor. My idea was to work with the unwanted offset swing. Done
great information, can you do an episode on negative impedance inverters to provide a -50OHM impedance out to at least 1GHz. device seletion, and a detailed operation. Your explanations are excellent. thank you
2:43 "It's a current mirror configuration" excuse me? It's differential stage, not a current mirror. Current mirrors are indeed heavily used in IC schematics however what you draw is not one of those.
Iam amazed that so many of my projects worked xD Well thanks dave you taught me some good things again. Ive allways chosen high resistors in my class-D amp input amplifier... I never thought that the offset of the inverted signal could come from this :D
Take a guess what's at the interface of your computer to the "analog" world? Audio in/out, VGA, HDD heads, DVD drive, mouse, gamepad, even the power supply and many more. They all use some kind of opamp (customized to fit the purpose). Opamps are the reason that you can interact and do something useful with your PC at all. You may see your PC as a "digital" machine, but in essence all signals are analog, even the bits.
Any advice on designing a very low-noise Transimpedance Amplifier to amplify a Photomultiplier Tube? The LTC6268 is very good, but I want to understand more of the theory. I know I want to use a FET input Op-Amp, since they have much lower noise current, but I've read that noise voltage is also important even in a TIA. I don't have specifications of the output impedance of the tube and I can't measure it. I can only hypothesize about the source characteristics. I do know that keeping the coaxial cable as short as possible will help keep my gain reasonable at higher frequencies.
19:45 This explanation with the 100x gain where 3uV turn into 300uV on the output eludes me. Since the 30pA go through the 100K resistor, and we started off by saying that V+ and V- are basically the same (so we didn't have to take in to account the 1K resistor), then that would mean that the output voltage is the 3uV and not the "3uV* 100 = 300uV". I sense that somehow V- is 3uV, but how, where does the voltage drop occur? Could someone please clear up my confusion. Thank you.
If you wanna measure the input bias currents, what voltage do you input to the op-amps? Do you ground both inputs before you measure the bias currents?
Yepp, my comment was aiming to "hint" that the usage of this term, without showing the whole opamp architecture is somewhat inappropriate. Besides, it is not necessarily that there would be a current mirror in an opamp (excluding the biasing circuitry), e.g. take a two stage miller opamp.
Sure, opamps are just a collection of transistors and whatnot. You can do what an opamp does with discrete transistors, but your problems will be even greater because those transistors aren't matched either. In fact they will be worse. Note that these "problems" only apply to "precision" circuits. Most circuits you don't have to worry much about it.
Understood, I usually use these for audio/ADC-antialiasing in a split power supply setup and compensate the OUT offset with a 9 turn trimmer, but that's just me, nothing critical... For other videos it would be *great* to know what the application is you are discussing, when you say "high precision", what do you mean? I agree you shouldn't explain Opamp basics, but sometimes explain HOW exactly are used. Those little things are helpful sometimes to move to the next level for people like me...
HI. Do you know where I can find a guide that explains the internal design of an op amp? Is it present in your channel by any chance? A thousand thanks!!
very interesting, and helpful tutorial. Most op amps have null offset leads by which we can zero adjust the output of the op amp. I'm wondering why we should worry about input offset voltage of op amps while we have this tool to zero adjust the op amp output. can you explain me?
did you clean the board after changing resistors? soldering may add extra leakage current, that depends on the type of the soldering tin. really like your blog!!!
You said 100 pA is a "very low input bias current op amp", referring to the one you have on board. No worries, mate! Just trying to point out to interested parties that they can go a lot lower!
Hmm, but why does adding a resistor from the opamp non-inverting to ground not work for a LM358? It makes it 2mV higher on the output. With 101x gain, it's about 102mV on the output.
Hi Dave, Im having exactly the same problem with the MAX4478, Im having 16.9mV output => 169uV Voltage offset when is typical 70uV Voltage offset. I tried to compensate the input bias current with the 1K resistor but it doesnt fix the issue. I tried also, lowering the power supply and about 2.7V of power supply i get near 80uV of Voltage offset. Can u give me some advice please? Regards, great videos!
I'm confused about the data sheet lingo. Why do you keep implying that a data sheet may list the "input offset current" as much greater than the "Input bias current"? Wouldn't the offset be, at MOST, only double the bias current? (ie. When the opamp's two inputs have equal bias currents but in opposite directions.).
Dave can I ask you a question? How in the Fluke 5020A calibrator, the instrument make different value of resistor based on a input and capacitance? Beszt regards, Alberto.
Very interesting video! For your application you're just gonna tweak the resistor according to your 3V lithium battery? Or by adding a stable voltage regulator?
I've made some op-amp headphone amplifiers with the NE5532 and I have to use a resistor from + to gnd to prevent heavy distortion, why is this? Also, how can current flow out of the base of a bjt (take a look at the schematic of the NE5532)? It's a PN-junction, I don't see how it's possible. Also, if the non-inverting input is connected to GND (and it's a single rail application), how can current flow one way or the other? If the input is a PN junction, isn't 0.6V required for any current to flo
Serious question, with a rail to rail op-amp there is no way you can be accurate rigt? Seems to me if one was to run a sinewave through it it would have little notches around the 0 line?
Hey Dave, If you are reading this comment, Please please answer the question which you asked at the end of previous video of this playlist. ( Question was asked at about 48:40 by you is that unusual behaviour of opamp due to change in resistors (10k, 100k) to (1k, 10k), and unexpected potential differences between inverting and non inverting inputs. ) I have asked my seniors and professors. And I am still in search of that answer.
Peration amplifiers are analog components. Digital systems don't have these problems with accuracy or offsets or what not. They have their own problems but as a digital systems they either work or don't.
I would pay for an EE Fundamentals course hosted by you that covers things like this. This is excellent -- it goes just beyond what I learned in college, where all the traps for young players are conveniently swept under the rug.
You explained in 30 minutes what took my professor 3 weeks... and I understand it better too. Apologies to Prof. Wormwood RIP.
I learnt more in this video than in the many years spent in the engineers school. Why ?
Because they were too busy teaching Math and not teaching Electronics.
so true lol
Because the teachers care more about their coffee breaks than teaching their students
@@k6eep593 same here in Turkey. Even at best technical university, the same.
Maybe, you did not realise what you missed and needed then,
it really happens a lot
Thank you, Professor Dave!
This is how classes at universities should be.
"Catch you next time!" - You bet!
Thank you, Dave. I'm a good ways into my electronics engineering degree, and tomorrow I'll be examinated in either MOSFETs/SMPS/Op-amp errors, and this video provided a great refresher as well as a few golden nuggets! :)
Yes, for a regular precision type bipolar opamp. The fA ones are pretty specialised, and I did mention fA is the lowest at the start of the video.
Fantastic video... I was struggling to understand some example circuits in op-amp datasheets, and they make perfect sense now. Rock on!
I´m so happy that I can understand Inglish..
I´m watching you from Argentina.
amazing video, breakdown the detail on input bias current and show how does it affect offset, it's very intuitive. Thx for sharing
Please more fundamental Friday videos! I especially love the practical part (the part when you explain how input bias current can screw you up).
This is perhaps the best "Fundamentals Friday™" to date!!!
I know this is an old video, but I came across something interesting that could help if this issue ever comes up in the future. Art of Electronics talks about making an integrator as a good way to measure offset voltage and bias current. You ground the input to measure the combined total, then float the input to measure just the bias current. Subtract to find the offset voltage.
Interesting. I'm compelled to give this a try.
Dave, I am very, very grateful for what you are doing. I can find here the easiest explanations of the problems which otherwise I would be searching for days studying books. Thank you very much!
I am so happy you are doing Fundamentals Friday section! Thanks!
You should get a job at my college. Everything was clear as you can possibly put it. Subscribed.
Awesome of you to put this information out here for free. This was a great review for me, even though it's not friday where I live.
That just blew my mind.
These tutorials on circuits and components you do are great.
Cheers
Really wonderful job man. I have been thinking a lot about building a high precission multipurpose dso and your videos are just pure gold. I ll definitely send you a prototype to demonstrate in one of your videos. Thanks again.
Excellent Explaination which are not taught in University. Thank you @dave
can't express how much I admire your fundamentals lessons!
Dave, I watched one of your earlier videos and then this one, and I have to say that you have improved alot on your teaching technique! Not saying anything bad about your earlier stuff. Have you considered teaching at a university? You've got the touch!
Hi Dave! Thank you for these videos, helps me reinforce what I'm learning at school. Would love to see one of those in-depth videos regarding the internal circuitry of op-amps if you could find the time!
Greetings from (Northern) NY!
I fully agree.
Current shunt resistors in the 10mOhm range are another example where you could need an opamp to make the voltage range fit your adc.
0:56 "But of course, in practice that's COMPLETE BULLSHIT" xD
i fcking love this guy, he's entertaining and he gives me good grades. I don't wanna go to school anymore i just wanna learn on the internet from people like you
At 11:04, I don't see how RB tacked on to the inverted input then ground should be set to R1. I definitely don't see how it should be RF//R1. I don't see it in this example. Maybe if you change the non-inverted input to have some current coming through R1 instead of tying it to ground. Makes more sense for it to be RF as hashed out on the white board.
I really love the Fundamentals Friday circuit explanations!
Great instructor, great lecture on op-amp input bias current.
You are the king at this.
On 8:55, before adding Rs2, is Ib+ equal to Ib-? I want to see more detail about how cancellation work.
I'm loving your fundamental fridays, it help me to remember many things here at argentina. Greets ;)
My GF (she doesn't do electronics) while listening to me watching this: "Is he talking about input bias carrots?"
Carrot can flow out of these pins and ruin your day. Imagine that!
Oi, look at all them cakes.
I got a good laugh out that 😂
great explanation of opamp basics. Thank you.
Dave, at 16:03 you say clic it down below where down below what is the number of the video?
Excellent Video!!! You are a very good teacher. I learnt a lot from this video :).
Another good one. Wish you had taught some of the analog classes at my college. This stuff might help me out at work some day, like a lot of your other vids!
1. There are part in your computer that are probably even more complicated than opamps - I would recommend to study processing of the signal coming out of your hard disk headers.
2. Opamps have more transistors in one chip which means they use current sources instead of resistors and differential and push - pull stages instead of simple common emitter circuits - this two design principles gives them huge advantage over anything you are able build from discrete components.
I'm pretty sure you messed up the calculation of the output voltage error caused by the input bias current. The current flows through your feedback resistor and causes a voltage of U=RI=3µV. But this voltage doesn't get amplified because it's on the output of the OpAmp. Compared to the error caused by the input offset voltage I would consider this effect negligible.
This
I don't know why anyone would go thumbs down on this video!
My remedy for this with my application is to use an NPN or PNP transistor with opamp, where the base is responsive to neg or pos signal otherwise you get zero output on transistor. My idea was to work with the unwanted offset swing. Done
Analog will never die!!!!!!!
this one was hard to swallow i ll watch it again :)
great information, can you do an episode on negative impedance inverters to provide a -50OHM impedance out to at least 1GHz. device seletion, and a detailed operation. Your explanations are excellent. thank you
2:43 "It's a current mirror configuration"
excuse me? It's differential stage, not a current mirror. Current mirrors are indeed heavily used in IC schematics however what you draw is not one of those.
Iam amazed that so many of my projects worked xD
Well thanks dave you taught me some good things again. Ive allways chosen high resistors in my class-D amp input amplifier... I never thought that the offset of the inverted signal could come from this :D
I absolutely love when Dave cusses. So classy.
Take a guess what's at the interface of your computer to the "analog" world?
Audio in/out, VGA, HDD heads, DVD drive, mouse, gamepad, even the power supply and many more. They all use some kind of opamp (customized to fit the purpose).
Opamps are the reason that you can interact and do something useful with your PC at all. You may see your PC as a "digital" machine, but in essence all signals are analog, even the bits.
Really like Fundamental Fridays for young players like me.
Thank You Dave for this video!
if only you had been my op amps teacher back in college
Great stuff dave, thanks for putting in the work.
Hell Dave, I started shutting lights off here to see if the Peaks went down. :-)
Really good video, thanks Dave.
There are some opamps that have extra pins attached for the input bios option.
Any advice on designing a very low-noise Transimpedance Amplifier to amplify a Photomultiplier Tube? The LTC6268 is very good, but I want to understand more of the theory. I know I want to use a FET input Op-Amp, since they have much lower noise current, but I've read that noise voltage is also important even in a TIA.
I don't have specifications of the output impedance of the tube and I can't measure it. I can only hypothesize about the source characteristics.
I do know that keeping the coaxial cable as short as possible will help keep my gain reasonable at higher frequencies.
TH-cam: Dave uploaded a new Fundamentals Friday video.
Me: AWESOME! :D
19:45 This explanation with the 100x gain where 3uV turn into 300uV on the output eludes me. Since the 30pA go through the 100K resistor, and we started off by saying that V+ and V- are basically the same (so we didn't have to take in to account the 1K resistor), then that would mean that the output voltage is the 3uV and not the "3uV* 100 = 300uV". I sense that somehow V- is 3uV, but how, where does the voltage drop occur? Could someone please clear up my confusion.
Thank you.
Thank you for the lesson.
So awesome seeing this.. I literally ran into op amp calcs on the job and needed a refresher to finish my work.
If you wanna measure the input bias currents, what voltage do you input to the op-amps? Do you ground both inputs before you measure the bias currents?
Yepp, my comment was aiming to "hint" that the usage of this term, without showing the whole opamp architecture is somewhat inappropriate. Besides, it is not necessarily that there would be a current mirror in an opamp (excluding the biasing circuitry), e.g. take a two stage miller opamp.
Sure, opamps are just a collection of transistors and whatnot. You can do what an opamp does with discrete transistors, but your problems will be even greater because those transistors aren't matched either. In fact they will be worse. Note that these "problems" only apply to "precision" circuits. Most circuits you don't have to worry much about it.
Understood, I usually use these for audio/ADC-antialiasing in a split power supply setup and compensate the OUT offset with a 9 turn trimmer, but that's just me, nothing critical... For other videos it would be *great* to know what the application is you are discussing, when you say "high precision", what do you mean? I agree you shouldn't explain Opamp basics, but sometimes explain HOW exactly are used. Those little things are helpful sometimes to move to the next level for people like me...
HI. Do you know where I can find a guide that explains the internal design of an op amp? Is it present in your channel by any chance? A thousand thanks!!
very interesting, and helpful tutorial. Most op amps have null offset leads by which we can zero adjust the output of the op amp. I'm wondering why we should worry about input offset voltage of op amps while we have this tool to zero adjust the op amp output. can you explain me?
i love when i come home to a video
Vos on that part is huge. 100 times the AD part.
did you clean the board after changing resistors? soldering may add extra leakage current, that depends on the type of the soldering tin.
really like your blog!!!
That's just gold, thanks!
You said 100 pA is a "very low input bias current op amp", referring to the one you have on board. No worries, mate! Just trying to point out to interested parties that they can go a lot lower!
Hmm, but why does adding a resistor from the opamp non-inverting to ground not work for a LM358? It makes it 2mV higher on the output. With 101x gain, it's about 102mV on the output.
I could spend hours looking on the internet but never find out that op amp can be made out of 2 transistors
How am I only finding this channel now... I am already on my 3rd semester of electronics engineering
Great video! Very practical! Now just to get a job that would allow me to utilize these concepts...
Awesome. Thank you! do you think you can explain CMRR?
Hi Dave,
Im having exactly the same problem with the MAX4478, Im having 16.9mV output => 169uV Voltage offset when is typical 70uV Voltage offset. I tried to compensate the input bias current with the 1K resistor but it doesnt fix the issue. I tried also, lowering the power supply and about 2.7V of power supply i get near 80uV of Voltage offset. Can u give me some advice please?
Regards, great videos!
+Hernan Olave any luck
+Yueeiuyoo Nope =\
very detailed teaching , thank you.
I'm confused about the data sheet lingo. Why do you keep implying that a data sheet may list the "input offset current" as much greater than the "Input bias current"? Wouldn't the offset be, at MOST, only double the bias current? (ie. When the opamp's two inputs have equal bias currents but in opposite directions.).
Great tutorial Dave, thanks!
Excelente video, esta información no se encuentra en español
Why not add a capacitor to the source impedance to block any DC offset signal.
Dave can I ask you a question?
How in the Fluke 5020A calibrator, the instrument make different value of resistor based on a input and capacitance?
Beszt regards, Alberto.
Very interesting video! For your application you're just gonna tweak the resistor according to your 3V lithium battery? Or by adding a stable voltage regulator?
Thanks for the Video Dude!!!!
I've made some op-amp headphone amplifiers with the NE5532 and I have to use a resistor from + to gnd to prevent heavy distortion, why is this? Also, how can current flow out of the base of a bjt (take a look at the schematic of the NE5532)? It's a PN-junction, I don't see how it's possible. Also, if the non-inverting input is connected to GND (and it's a single rail application), how can current flow one way or the other? If the input is a PN junction, isn't 0.6V required for any current to flo
Awesome Dave! More op amp material next Friday?
Adding a 100 ohm resistor at the input to ground will make the input impedance just as low, won't it?
Serious question, with a rail to rail op-amp there is no way you can be accurate rigt?
Seems to me if one was to run a sinewave through it it would have little notches around the 0 line?
could AD8603 be more suitable for your design?
~$1 1ku, RRIO, 1.8~5V supply, 50uV Vos max., 1pA Iin max.
2:35 - A current mirror configuration?
Sagan has reserved Dave for tomorrow.
He said you "could" have a Darlington configuration there if you wanted. He didn't actually draw that
I said that around 2:00, but may have goofed elsewhere because my mind is never in sync to the my mouth.
Hey Dave,
If you are reading this comment,
Please please answer the question which you asked at the end of previous video of this playlist.
( Question was asked at about 48:40 by you is that unusual behaviour of opamp due to change in resistors (10k, 100k) to (1k, 10k), and unexpected potential differences between inverting and non inverting inputs. )
I have asked my seniors and professors.
And I am still in search of that answer.
Hey Dave! Long time lurker and commenting on your video for the first time! Will you be making more Fundamentals Friday videos in the near future?
There currently isn't much competition!
Peration amplifiers are analog components. Digital systems don't have these problems with accuracy or offsets or what not. They have their own problems but as a digital systems they either work or don't.
"One of the rules for an ideal op-amp is that no current flows into input pins. But of course, in practice that's complete Bullshit. "
thanks for this video its helps me ........ for my project ..... thanks
no easy way out? what about an instrumentation amplifier?
I love the negative feedback t shirt :D
are you still going to have another one tomorrow, or is this it?