EEVblog

Fundamentals Friday.
Designing and measuring basic and precision opamp peak detector circuits.
All EEVblog Opamp related videos are here:
• Opamps - Tutorials & P...
Forum: www.eevblog.com...
EEVblog Main Web Site: www.eevblog.com
EEVblog Amazon Store:
astore.amazon.c...
Donations:
www.eevblog.com...
Projects:
www.eevblog.com...
Electronics Info Wiki:
www.eevblog.com...

[@MylesShannon]

I love how he starts off with a simple circuit and talks about how great it is, and then starts talking about how it sucks in some way, and then he gives us a new circuit that fixes that problem, and then it starts all over again. I love this guy!

[@gwav1a]

Exactly-

[@user-pc5sc7zi9j]

I love the fact you start of so rock basic people understand the workings instantly to then build up the complexity for everyone to follow along rather than just slaping down the state of the art from the get go and tearing it down into understandable sections later down the line (or let alone just going through it front to back).

[@martinda7446]

That was meat and drink to a hungry engineer. Thanks again Dave, hope you are not bored of the complements, you have really inspired me and I cant get enough at the mo.

[@alexreeve]

Says the NPN to the PNP: "oh what a nice complement" ;) scnr

[@jagr1460]

My EE anxiety eases as soon as I hear this man's voice. Dave has saved my college life, the man is a God 🙌🤣

[@gregben]

Seven years later and still a GREAT video. Thank you.

[@lisakingscott7729]

I came across this as I was looking for a peak detector to detect the time when the peak occurs rather than the value of the voltage. I've got a microcontroller with A-D doing it, but I wanted a simple circuit to generate a pulse, whenever the voltage starts to reduce. Just to make things more difficult it needs to be immune from noise, but that really just means a LPF on the input. It's for triggering a TRIAC on the AC mains peak for switching inductive loads. It should be pretty simple with a comparator some resistors and capacitor, so that in effect one comparator input lags the other and the output changes state whenever the input changes from increasing to reducing and vice versa, or something like that. In fact a Google search finds such a concept was even patented in 1969.

[@logitech446]

A scope painted a thousand words on trade-offs. Excellent presentation, Dave.

[@davidfof13]

This is the perfect way of teaching. You go through the conceptual design , the different implementations then you implement the circuit and test it.

[@lleaves]

Thank you for the video. I have designed peak detectors for AM waveform demodulation using op-amps to buffer the input and output similarly to what you have done here. In my applications I have needed the capacitor to discharge in order to follow the envelope of the AM waveform. I sometimes refer to the peak detector circuits I designed as envelope detectors for this reason. I did learn a few things from your video that I was not aware of. I may try to incorporate some of the techniques you presented here in my future designs. Thanks again!

[@SirMo]

This is one of your best videos Dave. Thank you!

[@somperohd7718]

so true

[@tonydengineering]

I start designing a nonlinear filter. Cut-off frequency must depend on the average amplitude of a signal. Your explanation is very helpful to me!

[@andymouse]

Awesome to be able to come back 10 years and get this level of tuition.....cheers.

[@EEVblog]

You can, but as I said, the diode drop changes with current and temperature. You can do a reasonable approximation tweek though, which should be suitable for something fairly crude like a VU meter.

[@hanedoggy]

Man I wish I had these videos 13 or 14 years ago... Would have made college soooo much easier. Sometimes breaking down complex concepts into basic functions is all a bad teacher needs to do to be a good one. Thanks Dave!

[@DjokovicWu]

Just discovered you channel and love it. I am an EE undergrad at Caltech, and this is just perfect for practical knowledge that we don't get in classes.

[@neilalbaugh4793]

Nice video, Dave. You've done a good job of explaining the peak detector. In practice, your circuit will work nicely with a storage capacitor of 1nF to 10nF if you put a 10 to 50 ohm resistor in series with that capacitor to ground. This cures the instant instability of the op amp caused by the capacitive load on its output. The loop stability is complicated by the diode's varying ON resistance as the current charging the capacitor varies. Regards, Neil ex-Burr-Brown Corporation

[@topquark22]

I built this circuit. It was unstable, and I spent hours trying to guess what could be wrong. Your suggestion cured the problem! Thank you so much! One other thing I did: I connected the output to a 4098 monostable multivibrator, driving a LED. The output also goes to a 2n7000 MOSFET that drains the capacitor, so that the cycle can repeat.

[@neilalbaugh4793]

@@topquark22 I'm glad it fixed your problem. Your feedback is much appreciated.

[@NickAlexeev]

Note that the diodes reduce the dynamic range. For the first circuit with a single diode, the max input voltage is the OpAmp's max input, minus one diode drop. For the second circuit with two diodes in series, the max input is the OpAmp's max input, minus two diode drops. These are useful circuits, and this is a good video.

[@bwack]

Search around for peek detection and here you are again. Awesome. I'm making a audio amp automatic switch box for my dads vintage amplifiers such that he can use one set of speakers without switching cables manually. The detector will tickle a state machine made out of 4000 series logic ic's.

[@daveb5041]

i gorgot how much I like this channel. The stupid algorithm didnt pop up any eevblog for over a month and I forgot about it. Why did you tube stop putting eev videos in my que?

[@adrianaldea5869]

Awesome!!! We are using peak detector for detecting peak pressure with the pusher centrifuge!!!

[@EEVblog]

And the final circuit I gave sucks in a way too! So it can start all over again in part 2 if there is enough interest...

[@markszlazak]

What is the video called for part 2? I do not see something like a "Peak Detector Curcuit - Part 2" among your videos.

[@PIXscotland]

Nice one. I've just implemented the second circuit in my Arduino power supply: 1. As a peak current detector using your I>V from your power supply. 2. To hold the output of a "PWM/CR filter" voltage output for the regulator setting. Works very well for both parts of the design.

[@Stelios.Posantzis]

Great tutorial! Highlights the gremlins that may lurk in an otherwise seemingly simple circuit!

[@EEVblog]

You don't need to get into physics, but yes, there is a lot more to this when you start playing with opamps, capacitor loads, open loop recovery etc. This simple circuit could spawn several branch topics to explain it.

[@HighestRank]

EEVblog too late, tho.

[@freon500]

Tour d' force. Awesome! There is enough here to keep me playing this video for quite a while!

[@AbhijithCRX]

I was searching for a peak detector circuit in google last week.. Today I got a video suggestion email from youtube with this link. Very useful and right on time for me. Thank you! (I am using a simple diode + cap circuit and adding the diode drop in software)

[@BryanByTheSea]

Wish you would bring back Fundamental Fridays. Love these.

[@jmi5969]

One thing to note is that if the op amp must NOT have input protection diodes - these would short the capacitor to the input node when the latter goes negative. CMOS opamps are usually fine (no protection diodes), as for bipolars it varies, e.g. the 5532 does have protection diodes, the 358 does not.

[@EEVblog]

Oh, right. Unfortunate similarity in the names, didn't think of that. They are of course different applications.

[@sagem116]

Dave u are an amazing teacher, i learned so much from this video and actually showing us the waveforms on the scope was very very helpful. thanks

[@fdk7014]

Great video! It would have been interesting to see the last circuit too, the "ultra precision" one

[@maxwellstrange4572]

Nothing like the aussie accent to put you in the mood to learn about circuits :D

[@DuanvantSlot]

If I may add an improvement suggested by Art of Electronics 3rd edition: Add another resistor and diode to the output of op amp 1. This will clamp the voltage to one diode drop below the ouput voltage, which dramatically increases the charging time of the capacitor. (The slew rate of the op amp is fixed, so keeping it at a diode drop below the output is much better than the output going all the way to Vee).

[@GV757]

page 255 fig.4.58

[@bendavis1215]

loving all the vids electronic theory you've done so far. I'd be really excited to see your spin on opto couplers/zero crossing/Traics... you are very easy to understand and I enjoy the way you expand problems and given problems in a way the audience can follow. Thank you

[@jonathanknight8251]

14:15 slew rate (the time it takes for the op amp to swings its output back up to the point where the diode begins to conduct again) and the value of the holding cap are unrelated. The size of the cap only matters when the diode is conducting, and in that case, the charge time is determined by the maximum output current capability of the op amp. They are both important but separate issues.

[@HighestRank]

Clearly we can see the output of the first op amp isn’t connected to the diode anyway.

[@RexxSchneider]

That's exactly right, and the issue of recovery from being in saturation is a different issue again. This is the issue faced by comparators and some opamps are much better than others when changing from the mode where you have a large differential between the inputs and returning to linear operation.

[@maqsoodu23]

good one EEV Blog.. i like this video but key point to remember...never drive a cap directly with a op amp. Reason: There is output capacitance built in to the op amp from the manufacturer,..so when you drive a cap with directly with an op amp, you change the pole frequency and thats changes the phase margin and creates instability. You can find this data in the datasheet of the op amp...*look at the open loop frequency response curve..i.e bode plot of the op amp

[@Nate_dog01]

This is by far my favorite series on KZbin.....nice job...

[@Sevalecan]

EDIT: Forgot about the opamp section, that it wasn't just a basic buffer after the fact. Should help a lot. I'm so impulsive, posting before I finished the video. Been a while since I last watched. Seems to me that the diode drop could also suffer at high frequency, right? Your diode has an ohmic region after the knee, as well as the capacitance and leakage you mentioned, but assuming we capture and buffer the value or store it digitally I can ignore the leakage. But we'd be creating an RC filter across the cap, or if we made the cap a very small value, then it leakage becomes much more significant..... I think I need to play with some circuit simulations. I was thinking of creating a peak inrush current detector. Of course, I haven't googled known solutions to the problem yet, where's the fun in that? Maybe I'll do that now. :)

[@RexxSchneider]

Around 5:30 the "droop" from leakage is almost never becoming greater as shown. Particularly for droop caused by resistive loads, the fastest rate of discharge is at the beginning, and it slows down over time. The shape of the voltage line is concave, not convex.

[@tocsa120ls]

Excellent video, nice demonstation! I'd like to add that it's really easy to get out of a common diode's comfort zone, for example, a 1N4001 will start 'zenering' at about -40V, so you really should keep the voltages above that. Also, sub-1V measurements are also 'out'. Too much distortion.

[@EngineeringAllAround]

It's possible to avoid very high negative oversaturation with just one additional diode and resistor (for the schema with two op amps). This reduces overshoot and switch on time.

[@CristiNeagu]

17:23 Smooth edit there...

[@nouveriantechno]

Gotta love the mid sentence jump cuts. If you look away you won't notice.

[@pepe6666]

haha brilliant. deserves an award for that one

[@YO6TJJ]

Thank you Dave for learning us electronics here! I just want to thank you for every video you uploaded and dedication!

[@JohnnyNoBueno]

In the ideal diode and cap circuit, can't you just put a buffer between the diode and storage cap to reduce the slew noise down to that first blip when the input goes negative? It looks like the cap impedance is causing most of the distortion.

[@viksohal]

Excellent tutorial. I do wish you had time to show the scope view of the final configuration with the second diode. Homework I guess :-)

[@jessstuart7495]

You can keep the op-amp from saturating at the negative rail by modifying the circuit a little bit. 1. Break the feedback connection from the cap to the 1st opamp's inverting input. 2. Use a 2nd opamp to buffer the cap voltage. 3. Add a diode from the 1st opamp's inverting input to the 1st opamp's output. The anode should be connected to inverting input, the cathode to the opamp's output. 4. Place a feedback resistor (10k?) from the buffer output (2nd opamp) to the 1st opamp's inverting input. This will allow the first opamp's output to effectively track the input signal, staying 1-diode drop below the input voltage when the input goes below it's peak value. The 1st opamp's output doesn't saturate at the negative rail that way, and the slew rate limitation is fixed. You can also use shottkey diodes which have a faster switching time. But you will probably need to use the 0-volts across the second diode trick, as shottkey diodes have higher leakage than regular PN diodes. If you need a fast response, you might have better luck using a fast comparator to compare the input with the buffered cap voltage (fast fet input opamp) and generate a digital signal. The digital signal could be NORed with a reset signal, or the digital signal could be gated by the reset signal using a series p-fet. The gatedl signal would control an n-fet connected to Vdd that would charge the storage capacitor connected to its source (source-follower). A second discharge n-fet would be controlled by the reset signal directly. If overshoot was a problem, you could add a small resistor in series with the charging n-fet to decrease the charging rate. Be sure to use adequate bypass capacitors as the charging fet will draw large spikes of current from the supply. This would make a peak-detector that would track very fast pulses probably up to a few MHz bandwidth.

[@davidmiles]

Thank you very much for this! Very well explained and the example/demo circuit helps tremendously for those of us who aren't engineers but like to tinker.

[@spiros1994]

there are mathematical equations for all that values you need do place in order to make it "perfect"?

[@aduedc]

The second opamp is in positive feedback loop and could osculate or with the diode could drive the output to positive rail. This could happen when there is a new peak at the input.

[@douggale5962]

Shouldn't you put another diode in the feedback loop of the second opamp to compensate for the diode before the capacitor? EDIT: ah, I tried it, the high input impedence causes the negative feedback voltage to be trapped and drives the output of the buffer low.

[@HAPPYChaU123]

i think the capacitor discharge curve on 4:20 and 5:13 should have positive second derivative which means its slope is increasing with time

[@arcadianelectronics]

My favorite part is when you go "wrrrrr!" and wave the marker around about a processor with fast ADC brute force reading the signal instead of using a peak detector. Excellent and entertaining video!

[@carmelpule6954]

Good video. I wonder if it would help if one mentions that the voltage sources used are not perfect and have a finite output impedance and so they cannot provide the high current to charge the capacitor very fast.

[@paguerchanel]

You are an absolute genius! Thanks a lot for taking the time to make this videos

[@theemelex]

I jus luv all these segments youve come up with. Of all the videos youve done, the best series i thoroughly enjoyed was the power supply series & I really miss those really meaty electronics videos.

[@anthonyvolkman2338]

A peak detector is basically a half wave rectifier. The opamp (precision diode) circuit reminds me of a synchronous rectifier.

[@kevinbeckenham3872]

Thank you Dave for another good tutorial on peak detector

[@daniele1209]

Thanks for great video. you deserve every single dollar you make out of this KZbin channel.

[@Scrogan]

So, is there a way to stop the first op-amp from saturating so much? Like by adding a diode to ground as a part of its feedback loop, or something.

[@supercircuitsfun]

The video creator only presented the theoretical circuit. When the capacitor is a short circuit you are burning the output of the operational amplifier. A 1 kohm resistor needs to be connected in series with the diode to limit the maximum opamp output current.

[@fahadnadeem6130]

Thank you very much, sir. I wish you were my teacher in school.

[@Scrogan]

But a small-signal MOSFET’s leakage current is going to be higher than a 1N4148, to say nothing of a low-leakage diode. Also I’ll look into using a precision rectifier with that extra diode to prevent saturation. Will still need to worry about slew rate though. I’m thinking maybe to use a JFET op-amp buffer on the cap, and feed that to a higher current op-amp’s input for the actual capacitor driving.

[@theemelex]

Lookin 4ward to the new uSupply. Only this time share some of the software / firmware as well. Now I know youre primarily a hardware guy but still why dont you do a segment called software Saturdays Id really like to learn a little something on the software side of things from Daves unique perspective! Splendid job thus far mate!

[@user21XXL]

cap discharge curve is bent the wrong way

[@MarkMcDaniel]

Yeah, it's asymptotic decay towards zero.

[@ohmslaw6856]

@@MarkMcDaniel hey man it's 2021.. everything is bent the wrong way

[@MarkMcDaniel]

@@ohmslaw6856 -- Fair.

[@ohmslaw6856]

@@MarkMcDaniel haha best wishes

[@MarkMcDaniel]

@@jaxtyneddie131 -- No, because we are not the loser you appear to be.

[@douggale5962]

Wouldn't it droop quickly at first then flatten out as the capacitor voltage approaches zero? Your graph implies that it droops slowly at first then more rapidly.

[@voice4voicelessKrzysiek]

Yes, exactly. Wrong discharge curve.

[@jackcaam2109]

Why not add a small resistor in series with the diode on the op amp output? It’ll limit the bandwidth but will make the load less capacitative and reduce inductance/capacitance/current spikes.

[@douro20]

Matsushita used to make small glass-encapsulated LEDs which had leakage current in the femtoamp range. I have a couple of them which came out of a CD player and they look very much like a small silicon diode.

[@michaelsohnen6526]

2 questions: What happens when the difference between the voltage on the cap and the input signal have a difference greater than the reverse breakdown voltage of the diode? What happens when you attach a load to the circuit, won't the capacitor drain through that?

[@gwav1a]

Thanks very much. I gain a lot from your videos. Brilliant!!

[@m.alkaff9110]

Yooo sir thank you very much, plz don't stop sharing

[@fabi5783]

Guarding technique ... awesome!

[@deltapparo4451]

I don't own a scope and keep claiming I have no need for one. But watching these videos sure makes me want to own one anyway.

[@GratedArseCheeks]

thanks for the video, the only thing that puzzled me was around the 5:00 minute mark. How come the capacitor is discharging like that? I was expecting it to discharge following a normal capacitor discharge curve. Is it something to do with the diode or am I missing something? Cheers :)

[@GratedArseCheeks]

***** True, but if they behaved like standard exponential decay, wouldn't the peak detector be quite bad? It would lose it's value quite quickly..

[@mkschreder]

Do you have a video where you design a zero crossing detector using an opamp?

[@tubical71]

You just have demonstrated, why good designs aren´t that simple. My own design of a Full wave precision rectifier, for my latest "Tinker-Goal" ended up with 5 OpAmps. But as nealy always i´m not fully satisfied with it. I started with a standart "tech-teach" 2 Op toppic, as also could be found in the LM3915 datasheets. Thanx for this vid!!

[@mooseteets]

if you were using a vu meter to display the cap voltage it would be tweakable to add the voltage drop of the diode, i assume other display devices could be tweaked the same way.

[@GiveAcademy]

he is referencing #471 the overload/ peak detector circuit you did.

[@ADEpoch]

So mate.... :-) ... do you give personal circuit advice to people trying to build their own peak detector? I'm trying to build a precision peak detector for an ADC (MCP3008) reader to read the output of a current transformer. But no matter what I do with the op-amps all I get is a constant DC out, even when a signal isn't applied. I get what I expect if I just use the simple (no op-amp) circuit, but then... you know... diode voltage drop :-( Tried multiple opamp configs). Including your version. I think op-amps just don't like me.

[@fl4tcircl3]

Your didactic skill is superb. Chapeau.

[@WhiskeyRichard.]

Oh baby, sample and hold me

[@mrjohhhnnnyyy5797]

Very nice! Now I know how to build an ideal diode, thanks!

[@famossfla]

Wonderful Dave Always learn a lot from your video's. Thanks for your time and energy.

[@kitanowitsch]

Thank you Dave. These tutorials are really great and helpful. Solving my engineering problems and going forward :)

[@metalmolisher666]

Im wondering why the green line is belowe the purple line. Shouldnt the Votlage directly behind the OPA be higher than behind the diode since it has to compensate the diode drop?

[@555propow]

Hi Dave. Its off topic, but could you give some advice or do a video how "gang" power transistors together for both a voltage and current output. eg you have a PWM and say want to drive a larger motor so you make a circuit running the smaller PWM into a large "power" circuit to drive the larger motor?? Thanks

[@darlingtonpear940]

Fundamentals Friday is awesome! Keep them coming!

[@网汉]

question please. Why there are multiple peaks shown at 29:30 instead of just following the input voltage smoothly??

[@ksrele]

This is very good video. I never imagine that simpla peak detector can be so complicated. Can you PLEASE make videos about zener and varicap diodes? And I want to know more about DIAC and TRIAC and I know that when you explain it I will understand it then :) Thanks you for nice videos. Greatings from Serbia.

[@TomMinnick]

I love Fundamentals Friday. Thanks Dave, as always a big thumbs up!

[@pohsinhee3076]

I learn a lot from your CIRCUT explanation .GOOD TO LEARN FROM U THANK.

[@paganodamiano]

Hi, the video is very interesting, congratulations. Maybe in the comments you have already answered, but I wanted to ask you how you generate the periodic signal you use in the example from the minute 20:40. Thank you!!!

[@gwc1410]

That periodic signal is a standard waveform generated by the function generator inside the scope.

[@jnkettenring]

You are a very good teacher. I learned a lot from this video. Thank you.

[@josephlunderville3195]

That oscillation and jankiness in the ideal diode op amp - would some of that not be due to the capacitive coupling through the diode itself, due to its parasitic capacitance?

[@aregjan]

You're an amazing teacher!

[@gtag174]

Hmmm i guess you just touch part of zero detection. Would it possible for you to explain the zero detection circuit. the circuit and working.

[@nouveriantechno]

Another drawback of the crude circuit is that the active discharge will also clamp the signal source to ground through the diode.

[@emmettpalaima9526]

Hey great video! What significance does the choice of op-amp / capacitor have in the behavior of circuit you use for the oscilloscope portion of the video? When I tried to recreate the behavior you're getting adding any resistance in parallel with the charge cap other than my scope probe caused the circuit not to work correctly (behaved more like a comparator with one end of the waveform clipping). I'm wondering if this is because I'm using a non-cmos LM358 and ceramic capacitors. I'd really appreciate if you could clarify. Thanks!

[@CagnPolat]

Most of the time I forget to press the like button. Sorry about that. You gave me lots of lots of useful informations. So thank you a lot.

[@BaeHat]

Just love these FF videos. Really educational, and simple explanations for "complex" circuitry, love it. If you anyday need some inspiration for a FF subject, i would love to hear something about diode reverse recovery time, and capacitance in the N-P junction on a diode.

[@ahmadhasan8607]

very useful, and interesting lecture. what if I want to detect, and measure a minimum voltage. what kind of op amp circuit should I use?

[@technobabble_]

I fucking love this series, Dave. Seriously. Every week I burn up my F5 waiting for it.