Side note: Using a light bulb to create a stable sine generator was the idea behind Hewlett Packard's first product and they continued manufacturing them in basically the same form into the 70s. I have a 201c from the early 70s on my shelf that still works well.
Did something like this some years ago. Rather than a Wien Bridge, I used an all-pass filter with a two-gang AM tuning capacitor. Unlike the Wien Bridge, both parts of the capacitor can be grounded. With appropriate phasing, the output from the juction of a small tungsten lamp and a fixed resistor was fed to the input giving positive feed-back. The lamp stabilisation took a moment to settle but for its simplicity worked well enough. One drawback with the AM capacitor is that the corresponding resistor is in the megohm region for low audio frequencies and the lamp will not stabilise at sub-audio frequencies. Anyway, it is much easier to get the two parts of a tuning capacitor to track rather than a 2-gang potentiometer which is harder to source. Finally, in view of the high impedance levels especially at low frequencies, good screening is needed to keep out 50/60Hz from the mains supply.
Learned all this in my bachelor's in Electrical Engineering, which was mostly from a theoretical a numerical point of view, but after watching the applications of it and the simple explanation gave a clear understanding of it. Thanks for making this video, keep up the good work!
Isn't he just great. He's young enough to be my grandson and I've been on and off in EE for most of my life, but these descriptions are SIMPLE and effective. Just what you need to know without the BS!
GREAT VIDEO! I once made a "variable" sine wave oscillator with a switched-capacitor filter IC. With this filter chip (the mf-10) the output center frequency is determined by a square-wave input. I just varied the square-wave input (10x the output filtered frequency), and got a completly variable sine out, just by varing the input of a 555 timer. This allowed the the output sine frequency and amplitude to be independently varied with an op-amp at the output. thanks again for the video!!
My method is not in your video. I use a relaxation oscillator - usually based on a 74XX14 Schmitt trigger - then a wave shaper based on 6 diodes, one capacitor and a resistor. You can generate a wide frequency by varying the resistor (potentiometer) on the feedback RC filter of the Schmitt trigger.
A common way to generate a sine wave used in analog function generators is to generate a triangular wave by switching a current source on a capacitor and then distort the triangular wave to a sine wave with a ladder of diodes and resistors. The THD is not that good (typ. 0.5%), but it is easy to vary the frequency.
Beat me to it. Used this many times on test equipment back in the 80s when I needed a good-enough sine wave in the low teens of Hz. A simple pair of back-to-back diodes around an inverting amp plus a couple of resistors did the job.
40 years ago I used this method in a function generator that was my senior design project in college. It is easy and works pretty well, but a bit of distortion is indeed there.
Please make a video about “negative resistance devices”. You can theoretically make a sine oscillator with just a LC tank circuit and a tunnel diode. Another device that can work is called a “lambda diode”. It’s basically either two JFETs or a JFET/BJT pair that has a I/V curve looking like a capital lambda symbol. Speaking of other sine wave generator designs: 1.) Triple inverter hysteresis oscillator making a triangle wave over a capacitor followed by an integrator opamp active filter. 2.) Emitter-Coupled LC Oscillator, which is harder to explain but it’s basically an attenuated LC tank hooked to an “anti-current mirror” differential amp setup made from two NPN BJTs. Since the common emitter pointing in a constant current sink thing makes the opposite BJT do the opposite thing compared to the BJT in which we feed the LC signal, it acts to increase the oscillation with each swing. However, due to the current limiting element and due to dampening, the oscillation stays a clean sine wave. Furthermore, this oscillator is more exact than the Colpitts or the Hartley variants because it only uses one inductive element and one capacitive element. With like 3 resistors, 2 BJTs, an L and a C, this is one of the simplest oscillator designs besides the tunel diode oscillators. 3.) Double active integrator filter inverter loop. Very similar to phase locked loops except you have an active non-inverting proportional-integrator-derivative filter in series with an active integrator filter. It has 3 resistors, 3 identical capacitors and 3 opamps. One of the resistors on the derivative side of the pid needs to be a few ohms lower than the rest (for stability and to avoid saturation). 4.) Emitter-Coupled Crystal oscillator (yep, there are those too). Same as the one above but the LC + R is replaced with a crystal.
Very informative but fast forwarded. I'll sure pause it after every part and try to understand and test it in my small home lab until I learn it all. Thanks!
Double integrating a square wave can be used the generate a sine directly. I built a accurate tone generator for a communication system. Xtal oscillator followed by PLL driving 2 op-amp integrators. Output to 600R line driver. Same principle but different implementation, without the frequency drift 🙂
Very nice video! I vaguely remember that it's possible to create sinusoidal signals by feeding two or four integrators into each other, since sine and cosine are derivatives of another.
Yes but remember that the two integrator one generally doesn't work if you copy it straight out of the text book. All capacitors have a little bit of series resistance so the oscillation generally rings down. Just a little bit of positive feedback can fix this. It is a very nice oscillator to build because you get sin() and cos() outputs. It is also possible to make one where a single 10 turn pot varies the frequency over about a 11:1 range. The trick is to ground the wiper on the pot and make two vary different amplifiers where the gain is controlled by a grounded resistor. The CW and CCW ends of a 10K pot with 1K added resistors comes into it.
I've done every one of these in past when I was a designer. There's quite a few things can do with each circuit that makes the sine purer and output level more stable, but most real applications don't need them. The JFET agc is temperature sensitive output level, can get two orders more stability by using a bandgap reference plus long tail pair
Other than a tiny tungsten lamp to stabilise the gain of a Wien bridge oscillator it was common to use a vacuum-encapsulated self-heating thermistor designed for that kind of thing - R53 for example - but they're no longer made.
I once made a mathematical model of something, i hoped would become a sine generator. I started with a rectangular wave and fed it into a constant current source. That source is fed into a capacitor. The resulting wave will be a triangle. Then, feed that triangle into another constant current source and the output into another capacitor. The resulting signal is a sine wave. Well - i never built that. But i wrote a program, that calculates the currents and voltages. It then compared that to a real sine wave. The difference was about 1%, which surprised me a little.
A little note about the original pronounciation of Mr. Wien’s name. It was not “wine”. Maybe closer to English spelling would result from “ween”. Probably useless to try making any change, considering that the name of the city - Wien - is in English world also modified to Vienna… Just my today’s rattle!
All of my professors at an American University pronounced it more phonetically, less Germanic. I think in Engineering, it’s more important to communicate effectively given the wide diversity and general poor spelling ability of my colleagues.
I agree clarity of communication is paramount hence it should be pronounced CORRECTLY since this is the WORLD wide web after all! Nobody should pronounce anyone's name incorrectly, especially when their name is given to such a famous design. Besides pronouncing it as "Wine" when even in English an I before the E is not pronounced that way (Fiend for example). It also is phonetic when correctly pronounced! Unlike English MOST other languages ARE phonetic ... The list is endless.
Great pace and consice experimental & theoretical explanations. This is a great recap of popular topologies. I am a fan of crystal oscillators personally
To be honest I've always found very hard to visualize the functioning of oscillator circuits. I love this video and it will be of such great help to those who are intrigued by the inner workings of oscillator circuits. However, there is one circuit not shown in the video that I find the most didactic and easiest to understand how it oscillates: the Meißner-Armstrong oscillator. It can be somewhat tricky to build though, since it makes use of a transformer (magnetically coupled) to "inject" energy into a LC-tank.
You left out the Armstrong oscillator. This is one where the inductor is used as a transformer. The input of your amplifying device comes from one winding and the output of the amplifier drives the other.
Interesting channel, I like that you experiment with discreet elements. It has its charm, when I was your age (seventies) there was only such electronics for hobbyists, and this teaches the basics. :)
Where is the twin Tee notch ? It provides very low distortion, stability, and simplicity using an OPAMP. Very good presentation. Check out Analog Devices DDS (direct digital synthesis) for the ultimate in range. Programmed with SPI or I2C. Follow it up with a multiplying DAC for amplitude control.
It's not just theoretically; many commercial analog function generators were essential square/triangle generators with an added diode shaper for making the sine wave. One place where this idea is still used is in the modular synthesizer world. Also the XR2206 chip that is the active component on many cheap function generator kits has a diode shaper internally.
8:09 My home made sine wave generator uses the Wein bridge configuration with a full wave rectifier to control the JFET... it can generate Sine waves from 1.5 Hz up to about 130KHz and has a ridiculously low distortion (less than 0.05% between 1.5 and 20KHz and about 0.1% above that. So, it definitely is possible to make this work.
Good presentation! If you have time and the opportunity it would be interesting to analyze and see the quality of DDS sine waves generated from chips like the AD9833.
I wish you had compared phase noise and total harmonic distortion between all of them , because is the most important factors if you need a sine wave oscillator
Operating any amplifier in its unstable region generates a sine wave. Kicking an LC tank generates a sine wave. Sine waves are the most natural waveforms possible. The problem is that it is not a perfect sine wave; but, your square wave is not a perfect square wave either -- it has limited rise/fall times, ringing after transitions, and jitter. It is only difficult to generate perfect waveforms.
Years ago I built a sine wave generator using a fast op-amp, can't for the life of me remember which one, and a low current incandescent lamp, I think it had to be less than 60mA. It had several switched ranges, covered frequencies from a few Hertz to about 150kHz and the output was VERY pure, < 1% TDH. You can use an R53 bead thermistor, but those are rarer than chickens teeth now and most likely VERY expensive.
Great video ! I'm a little sad you didn't use a bulb, but it is a bit of hassle however I think the results are worth it ! maybe some day on the 'sine Lab' as it did launch the company HP and I don't mean the English sauce ! good stuff and its 'Veen'.....cheers :)
Excellent. I have done the transistor phase shift osc and the R/C filter, my target was 1K, it made my day seeing the results. Your Op-Amp suggestions look really cool and challenging enough, so I'll be doing that. Anybody can do a square wave, but class A is real class.
There used to be some chips available for generating sine waves like the ICL8038, and the XR-2206, but they're not made any more and are obsolete, just try finding any genuine unused ones now.
What I'd like to see is a circuit that can generate a two antiphase sinewaves at 50 hertz which can also be frequency stabilised, then you can build a pure sinewave power inverter that has the same crest factor as the mains, as the designs I've seen for home constructors always use CMOS chips to generate the signals to be amplified, which can easily be stabilised with a crystal but they only generate square waves, no good for inverters.
Wow, you pulled together several knowledge points, for me. Thanks. Kudos. If I recall correctly, it is rather difficult to create a true sine wave and that a lot of the circuits that we see actually only produce a good facsimile of the wave. A digital sine wave would be an example of this. If I am remembering correctly, analog circuit that starts with a square wave and reform it into a sine wave are also examples. So, do your five methods produce a true sine wave. Maybe you could make a video, on this distinction...
Very interesting video. It seems like most of the circuits presented in the video create sine waves with a bit of distortion. I’m wondering how analog function generators can create such perfect distortion-free sine waves? How do they manage to get the sine waves so clean, and also at the same time have the frequency changeable via a pot?
He mentionned it in the beginning - more filters (or more complex or more steep, simply as high Q as possible) = less distortion. But hey there's no such thing as "distortion free sinewave", even by probing it by oscilloscope you introduce another of many distortions. But yeah you can definitelly make it even less distorted by using more advanced filtering.
You can also use 2 antiparallel diodes in the Wien bridge oscillator, between feedback ressistor and op amp's output. This is how I built ḿy censoring button circuit.
It works better if you add a resistor in series with the clipping diodes. With the diodes off, the gain is about 3.3. With the diodes on, the gain is about 2.7. This leads to less distortion.
The name Wien bridge comes from Max Wien, a German physicist who worked in Jena. It's also a German name of Vienna (city in Austria). I would think it should be pronounced like in wiener (hot dog, also called Vienna sausage). It has nothing to do with wine. About everyone on RUclips pronounces it differently.
The question one would want answered when deciding which sine wave generator to pick is how close the resulting signal actually comes to being a sine wave. Unlike square waves, this is not easy to judge by looking at the scope trace. A spectrum analyzer would provide this info, but are there any theoretical considerations to decide which oscillator comes closest, produces the purest sine wave?
Yes, LC tanks with ideal components. Or phase shift oscillators with ideal components (ideally linearized amplifier, ideal capacitors and resistors) with perfectly matched phase (180°). If you're curious about the topic I recommend to try to simulate it and probe it via FFT if your simulator has it. You can try LTspice which is free or falstad simulator which is even available online. You'll see most generators are giving pure sinewaves or squares, the problems begin when you use real components.
@@pheymee88 "with ideal components", so all we need now is for you to tell us where we can purchase these ideal components. ;) Obviously, the point of the question was which design is best with real components. Perhaps we need to look at the design of high-end signal generators for real-world answers.
At 12:18 Crystals have two resonant frequencies, series resonance and parallel resonance. That may be why you are getting 6.5 MHz instead of 4.0. You may also want to look at IMSAI Guy's video #1807 Twin T Oscillator using an OP amp.
Good stuff. Definitely something to play with... though, isn't there also an option to use an op-amp as an integrator, and integrate a square wave?? (Or, were you doing that somewhere, and I just misunderstood it?) Also, another way to spell "2𝜋" is "𝜏", which can then be pronounced "tau", instead of "two pi". Just sayin'. ;) (Not sure what a 𝜏 would look like in that font, though! 🤔)
@@SineLab oops, yes, I was missing a step. Integrating the _triangle_ wave, in turn, gives you something that... well, approximates a sine wave, though isn't one in a strictly mathematical sense. :)
Tau is a really bad letter choice to designate 2*pi, especially in the context of electronics. It's already taken to designate the time constant of first order systems. I don't really know why some math folks are so fond of "tau", I've always been used to "2*pi" everywhere since my studies and I'm perfectly fine with that lol.
Tried using low pass filters to get sine out of a 555 square pwm, none worked. Spike sines top and bottom. Even repeating the circuit in falstad, same results. Power supply from a battery.
@@SineLab there's 101 channels that do basic electronics (another 'what is a diode' video anyone?), so great to see your intermediate level content. Delivery style is great too
@@SineLab Anything you're planning to do with that correction? Or are you just content with leaving misinformation uploaded as long as it brings you clicks? Not even a pinned comment with correction?
I should note a few issues with this otherwise informative video: - "Wien" is pronounced "veen", NOT "wine" or "vine". - One needs to be careful with so-called "sine wave oscillators", as a lot depends on how the resulting waveform signal will be used. In many applications, it is critical to get a low distortion sine wave, and the majority of circuits shown in this video, while 'Sine Lab' proclaimed them to be 'clean sine waves', were actually quite visibly distorted, and a distortion meter would have shown this numerically. It is a bit of an art form to visually recognize the presence of sine distortion, made much more difficult for people who have little time viewing truly 'clean' sine waves (but the visual method, even at its best, can miss significant amounts of distortion). Some things even a beginner can look for: symmetry issues (does the down curve shape match the up curve shape), how pointy are the peaks and troughs, do the sides of the waveform look straight rather than constantly curving, etc. - In a Wien bridge oscillator, some sort of dynamic amplitude control is needed, in order to minimize clipping and other distortion, and this video mentions this and that the incandescent light bulb (not just any light bulb) is traditionally used for this. And this video correctly points out that an FET may be pressed into service as an alternate to the incandescent bulb. But two LEDs, two diodes, etc; may also be used in place of the bulb, with varying degrees of distortion. However, in the incandescent bulb version, the bulb never gets very warm, and certainly does not get hot enough to incandesce. - In addition to the five ways of generating a sine wave, most modern professional grade waveform generators are doing so digitally; the precise sine waveform shape is stored numerically in a look-up table, and played back at various speeds; this works very well for very low distortion sine wave generation, but its complexity makes the generation of higher frequencies significantly more expensive.
You need to generate a PWM signal that represents the sine wave and feed a converter you choose. With a self oscillating UcD you can get as low THD as 0.0001%, with a buck/boost converter this is about 1-3%, with a full bridge converter and some modulating techniques (like a phase shift or a complementary detector) this is about 0.01-0.1%. Then make a choice according to your generator specs.
Hey, did you know you can also generate sine mechanically by rotating a round sine-shaped magnet and scanning its shape by coils? That's what Hammond organ is doing.
Which is then usually sent through a spinning speaker to create a phase shift effect. You can do amazing stuff with mechanical devices but they tend to be very big and gobble power :)
A filter reduces a signal based on its frequency. For example, a "low-pass" filter passes low frequencies and dampens high frequencies. An example where you'd see one is the subwoofer output on a stereo - it filters out all the highs and passes the lows. The "cut-off" frequency of a filter is (usually) the frequency where the signal is diminished by 3dB. So for a low-pass filter like what he's using here, you'd want the cut-off frequency to be a bit higher than the frequency of the sine wave you want to create. Square waves are really just an infinite number of sine waves added together, so by filtering out the frequencies higher than the one you're looking for you can turn a square wave into a sine wave. Filters are a pretty big topic and they can get pretty complex, but the basics aren't difficult. You can find lots of introductory articles on them around the 'net.
You forgot one of the most common ways to generate a sine wave that is used on all the arbitrary waveform generators... a microcontroller coupled with a digital to analog converter. 😉
So if a sine wave is, well a sine wave, then just out of curiosity, is there such thing as a cosine wave for electrical circuits? Since it uses trigonometry, then its got something to do with circles or spheres, is there any significance to that? Like, are signals actually 3D and we are just looking at it in 2D for simplicity? This probably sounds dumb af, but im curious.
The waves are just in two dimensions: amplitude (y) and time (x). A cosine wave would look identical to a sine wave except that it would have a 90° phase shift to the left.
4:15 Please don't forget to bypass R7. 10uF would be good, 100uF would be better, but even 1uF would be okay-ish and better than nothing. Wien is German and, as such, as closely as can be approximated with American phonemes, is pronounced "wean." (Actually it vould be closer to "veen," but ve can't be too choosy.) If it were pronounced "wine" or "vine," it would be spelled wein.
Educated people often mispronounce names or words because they've only read them. At 45 I have a huge list of things I know how to pronounce "correctly" and how it has lived in my brain for 30 years. I pronounce with with a German V "Vienbridge!" when I'm feeling sassy.
@@justovision No. "Educated people" worth the label actually verify what they're talking about, in particular when foreign languages are involved, and don't just parrot everything without a second guess. Because that's not how science works.
I wouldnt call any of these sine waves.If anyone needs a real sine wave with controllable frequency and low thd. Know that It's extremely hard and takes a box of parts. Luckily modern tech has DDS chips that spit out a magical sine wave with low thd (or use a dac if your freq isnt too high)
Side note: Using a light bulb to create a stable sine generator was the idea behind Hewlett Packard's first product and they continued manufacturing them in basically the same form into the 70s. I have a 201c from the early 70s on my shelf that still works well.
Awesome, I'm jealous! I have designed my own bulb stabilised Wien oscillator and that happened after reading about about the work done by Mr Hewlett !
wow cheap !@@justovision
Me too I will be home this weekend I would like to show you my design :) @@justovision
Jeri Ellsworth has a good video too. ruclips.net/video/vBsSASge7ls/видео.htmlruclips.net/video/vBsSASge7ls/видео.html
Did something like this some years ago.
Rather than a Wien Bridge, I used an all-pass filter with a two-gang AM tuning capacitor. Unlike the Wien Bridge, both parts of the capacitor can be grounded. With appropriate phasing, the output from the juction of a small tungsten lamp and a fixed resistor was fed to the input giving positive feed-back. The lamp stabilisation took a moment to settle but for its simplicity worked well enough. One drawback with the AM capacitor is that the corresponding resistor is in the megohm region for low audio frequencies and the lamp will not stabilise at sub-audio frequencies. Anyway, it is much easier to get the two parts of a tuning capacitor to track rather than a 2-gang potentiometer which is harder to source. Finally, in view of the high impedance levels especially at low frequencies, good screening is needed to keep out 50/60Hz from the mains supply.
Learned all this in my bachelor's in Electrical Engineering, which was mostly from a theoretical a numerical point of view, but after watching the applications of it and the simple explanation gave a clear understanding of it. Thanks for making this video, keep up the good work!
Isn't he just great. He's young enough to be my grandson and I've been on and off in EE for most of my life, but these descriptions are SIMPLE and effective. Just what you need to know without the BS!
GREAT VIDEO!
I once made a "variable" sine wave oscillator with a switched-capacitor filter IC. With this filter chip (the mf-10) the output center frequency is determined by a square-wave input. I just varied the square-wave input (10x the output filtered frequency), and got a completly variable sine out, just by varing the input of a 555 timer. This allowed the the output sine frequency and amplitude to be independently varied with an op-amp at the output.
thanks again for the video!!
My method is not in your video. I use a relaxation oscillator - usually based on a 74XX14 Schmitt trigger - then a wave shaper based on 6 diodes, one capacitor and a resistor. You can generate a wide frequency by varying the resistor (potentiometer) on the feedback RC filter of the Schmitt trigger.
A common way to generate a sine wave used in analog function generators is to generate a triangular wave by switching a current source on a capacitor and then distort the triangular wave to a sine wave with a ladder of diodes and resistors. The THD is not that good (typ. 0.5%), but it is easy to vary the frequency.
Beat me to it. Used this many times on test equipment back in the 80s when I needed a good-enough sine wave in the low teens of Hz. A simple pair of back-to-back diodes around an inverting amp plus a couple of resistors did the job.
40 years ago I used this method in a function generator that was my senior design project in college. It is easy and works pretty well, but a bit of distortion is indeed there.
Does anyone know what that circuit would be called?
Please make a video about “negative resistance devices”.
You can theoretically make a sine oscillator with just a LC tank circuit and a tunnel diode.
Another device that can work is called a “lambda diode”. It’s basically either two JFETs or a JFET/BJT pair that has a I/V curve looking like a capital lambda symbol.
Speaking of other sine wave generator designs:
1.) Triple inverter hysteresis oscillator making a triangle wave over a capacitor followed by an integrator opamp active filter.
2.) Emitter-Coupled LC Oscillator, which is harder to explain but it’s basically an attenuated LC tank hooked to an “anti-current mirror” differential amp setup made from two NPN BJTs. Since the common emitter pointing in a constant current sink thing makes the opposite BJT do the opposite thing compared to the BJT in which we feed the LC signal, it acts to increase the oscillation with each swing. However, due to the current limiting element and due to dampening, the oscillation stays a clean sine wave. Furthermore, this oscillator is more exact than the Colpitts or the Hartley variants because it only uses one inductive element and one capacitive element. With like 3 resistors, 2 BJTs, an L and a C, this is one of the simplest oscillator designs besides the tunel diode oscillators.
3.) Double active integrator filter inverter loop. Very similar to phase locked loops except you have an active non-inverting proportional-integrator-derivative filter in series with an active integrator filter. It has 3 resistors, 3 identical capacitors and 3 opamps. One of the resistors on the derivative side of the pid needs to be a few ohms lower than the rest (for stability and to avoid saturation).
4.) Emitter-Coupled Crystal oscillator (yep, there are those too). Same as the one above but the LC + R is replaced with a crystal.
Very informative but fast forwarded. I'll sure pause it after every part and try to understand and test it in my small home lab until I learn it all. Thanks!
Double integrating a square wave can be used the generate a sine directly. I built a accurate tone generator for a communication system. Xtal oscillator followed by PLL driving 2 op-amp integrators. Output to 600R line driver. Same principle but different implementation, without the frequency drift 🙂
Technically a parabolic wave.
Very nice video!
I vaguely remember that it's possible to create sinusoidal signals by feeding two or four integrators into each other, since sine and cosine are derivatives of another.
Absolutely quite easy, and easy to make voltage controlled with OTAs or VCAs like the 2164.
Very pure too.
this sounds like wien bridge osc- integrators are lowpass and differentiators highpass
Yes but remember that the two integrator one generally doesn't work if you copy it straight out of the text book. All capacitors have a little bit of series resistance so the oscillation generally rings down. Just a little bit of positive feedback can fix this.
It is a very nice oscillator to build because you get sin() and cos() outputs.
It is also possible to make one where a single 10 turn pot varies the frequency over about a 11:1 range. The trick is to ground the wiper on the pot and make two vary different amplifiers where the gain is controlled by a grounded resistor. The CW and CCW ends of a 10K pot with 1K added resistors comes into it.
@@kensmith5694 ofc its like a megaphone yelling into itself. without gain it wont oscillate and screetch
@@superpie0000I am trying to work out the intended meaning of your comment. Please explain what you mean.
I've done every one of these in past when I was a designer. There's quite a few things can do with each circuit that makes the sine purer and output level more stable, but most real applications don't need them. The JFET agc is temperature sensitive output level, can get two orders more stability by using a bandgap reference plus long tail pair
Other than a tiny tungsten lamp to stabilise the gain of a Wien bridge oscillator it was common to use a vacuum-encapsulated self-heating thermistor designed for that kind of thing - R53 for example - but they're no longer made.
I once made a mathematical model of something, i hoped would become a sine generator.
I started with a rectangular wave and fed it into a constant current source. That source is fed into a capacitor.
The resulting wave will be a triangle.
Then, feed that triangle into another constant current source and the output into another capacitor.
The resulting signal is a sine wave.
Well - i never built that. But i wrote a program, that calculates the currents and voltages. It then compared that to a real sine wave. The difference was about 1%, which surprised me a little.
The double integral of a constant is a quadratic. Gluing different quadratics to each-other does not an exact sine-wave make.
A little note about the original pronounciation of Mr. Wien’s name. It was not “wine”. Maybe closer to English spelling would result from “ween”. Probably useless to try making any change, considering that the name of the city - Wien - is in English world also modified to Vienna… Just my today’s rattle!
I came to post the same rattle 😃 Nearer 'veen', I suppose.
We write it and pronounce it Viena(Vee-E-nah) here in Brazil.😊
@@SpisakRazlogaMe too! Veen not Wine 👍
All of my professors at an American University pronounced it more phonetically, less Germanic. I think in Engineering, it’s more important to communicate effectively given the wide diversity and general poor spelling ability of my colleagues.
I agree clarity of communication is paramount hence it should be pronounced CORRECTLY since this is the WORLD wide web after all! Nobody should pronounce anyone's name incorrectly, especially when their name is given to such a famous design.
Besides pronouncing it as "Wine" when even in English an I before the E is not pronounced that way (Fiend for example). It also is phonetic when correctly pronounced! Unlike English MOST other languages ARE phonetic ... The list is endless.
Great pace and consice experimental & theoretical explanations. This is a great recap of popular topologies. I am a fan of crystal oscillators personally
To be honest I've always found very hard to visualize the functioning of oscillator circuits. I love this video and it will be of such great help to those who are intrigued by the inner workings of oscillator circuits. However, there is one circuit not shown in the video that I find the most didactic and easiest to understand how it oscillates: the Meißner-Armstrong oscillator. It can be somewhat tricky to build though, since it makes use of a transformer (magnetically coupled) to "inject" energy into a LC-tank.
Thank you. I did all of those a few months ago. I did a joule thief oscillator too. Good fun.
Oh except for the Cristal oscillator. Gotta do that. Definitely 😁
You left out the Armstrong oscillator. This is one where the inductor is used as a transformer. The input of your amplifying device comes from one winding and the output of the amplifier drives the other.
A true classic and the most common configuration used in a regernative receiver.
This was a very good presentation. Very nice to have the formulas for each circuit.
Interesting channel, I like that you experiment with discreet elements. It has its charm, when I was your age (seventies) there was only such electronics for hobbyists, and this teaches the basics. :)
Retired EE. Love it. Ready for DDS and multiplying DAC.
Where is the twin Tee notch ? It provides very low distortion, stability, and simplicity using an OPAMP. Very good presentation.
Check out Analog Devices DDS (direct digital synthesis) for the ultimate in range. Programmed with SPI or I2C. Follow it up with a multiplying DAC for amplitude control.
Theoretically speaking you can use a diode shaper to turn a triangle signal into a sine vawe
It's not just theoretically; many commercial analog function generators were essential square/triangle generators with an added diode shaper for making the sine wave. One place where this idea is still used is in the modular synthesizer world. Also the XR2206 chip that is the active component on many cheap function generator kits has a diode shaper internally.
You can also use a pair of diodes in op-amp feedback loop for automatic gain control.
8:09 My home made sine wave generator uses the Wein bridge configuration with a full wave rectifier to control the JFET... it can generate Sine waves from 1.5 Hz up to about 130KHz and has a ridiculously low distortion (less than 0.05% between 1.5 and 20KHz and about 0.1% above that. So, it definitely is possible to make this work.
Keep cascading R-C pairs after a 555, with increasing Z.
The results are excellent and cheap to do!
OK for light loads, or amplify after.
Awesome explanations i loved it. Wish u were part of my universities curriculum.
Happy I know a lot of this but still a 1st class video. Thanks.
Good presentation! If you have time and the opportunity it would be interesting to analyze and see the quality of DDS sine waves generated from chips like the AD9833.
Opamps are seriously so kickass.
I wish you had compared phase noise and total harmonic distortion between all of them , because is the most important factors if you need a sine wave oscillator
Operating any amplifier in its unstable region generates a sine wave. Kicking an LC tank generates a sine wave. Sine waves are the most natural waveforms possible. The problem is that it is not a perfect sine wave; but, your square wave is not a perfect square wave either -- it has limited rise/fall times, ringing after transitions, and jitter. It is only difficult to generate perfect waveforms.
No, many amplifiers don't generate a sinewave when in the unstable region.
Years ago I built a sine wave generator using a fast op-amp, can't for the life of me remember which one, and a low current incandescent lamp, I think it had to be less than 60mA.
It had several switched ranges, covered frequencies from a few Hertz to about 150kHz and the output was VERY pure, < 1% TDH.
You can use an R53 bead thermistor, but those are rarer than chickens teeth now and most likely VERY expensive.
Great video ! I'm a little sad you didn't use a bulb, but it is a bit of hassle however I think the results are worth it ! maybe some day on the 'sine Lab' as it did launch the company HP and I don't mean the English sauce ! good stuff and its 'Veen'.....cheers :)
I'll look around for a suitable bulb. :)
:)@@SineLab
Excellent.
I have done the transistor phase shift osc and the R/C filter, my target was 1K, it made my day seeing the results. Your Op-Amp suggestions look really cool and challenging enough, so I'll be doing that. Anybody can do a square wave, but class A is real class.
нет
@@rainy2063 yes
You have such great knowledge. Thanks for sharing. Looks like you have a great channel here.
Nice tutorial. For another elementary intro to oscillators I suggest looking up the video "Lab 04 Oscillator design" by Mazhari.
There used to be some chips available for generating sine waves like the ICL8038, and the XR-2206, but they're not made any more and are obsolete, just try finding any genuine unused ones now.
This video is freking great! Keep it up!!
Very well explained
Point to note: it's not a Wein (wine) Bridge, but a Wien (veen) Bridge as in Vienna (Wien) in German.
In english it should sound something like "Vine".
Very informative and nicely made tutorial . Thanks
Thanks for watching!
What I'd like to see is a circuit that can generate a two antiphase sinewaves at 50 hertz which can also be frequency stabilised, then you can build a pure sinewave power inverter that has the same crest factor as the mains, as the designs I've seen for home constructors always use CMOS chips to generate the signals to be amplified, which can easily be stabilised with a crystal but they only generate square waves, no good for inverters.
I think there is a slight mistake at 6:18 Vo/Vin should be Z2 / (Z1+Z2), but since R1=R2, C1=C2 then Z1=Z2 in that example and it doesn't matter.
Wow, you pulled together several knowledge points, for me. Thanks. Kudos. If I recall correctly, it is rather difficult to create a true sine wave and that a lot of the circuits that we see actually only produce a good facsimile of the wave. A digital sine wave would be an example of this. If I am remembering correctly, analog circuit that starts with a square wave and reform it into a sine wave are also examples. So, do your five methods produce a true sine wave. Maybe you could make a video, on this distinction...
Yes, they do. Just look at the scppe.
These five do produce a true sine wave. A video about the topic of sine wave approximations, digital or analog, sounds interesting though.
Very interesting video. It seems like most of the circuits presented in the video create sine waves with a bit of distortion. I’m wondering how analog function generators can create such perfect distortion-free sine waves? How do they manage to get the sine waves so clean, and also at the same time have the frequency changeable via a pot?
He mentionned it in the beginning - more filters (or more complex or more steep, simply as high Q as possible) = less distortion. But hey there's no such thing as "distortion free sinewave", even by probing it by oscilloscope you introduce another of many distortions. But yeah you can definitelly make it even less distorted by using more advanced filtering.
Good video. Very educative.❤ Thanks for sharing...
Please Sir, what's the title of the background music?
It's called fig leaf rag.
very excellent explanation
You can also use 2 antiparallel diodes in the Wien bridge oscillator, between feedback ressistor and op amp's output. This is how I built ḿy censoring button circuit.
It works better if you add a resistor in series with the clipping diodes. With the diodes off, the gain is about 3.3. With the diodes on, the gain is about 2.7. This leads to less distortion.
The name Wien bridge comes from Max Wien, a German physicist who worked in Jena. It's also a German name of Vienna (city in Austria). I would think it should be pronounced like in wiener (hot dog, also called Vienna sausage). It has nothing to do with wine. About everyone on RUclips pronounces it differently.
It's very easy to get a sine wave from a digital signal. Just run the digital signal through an ethernet cable. ;)
The question one would want answered when deciding which sine wave generator to pick is how close the resulting signal actually comes to being a sine wave. Unlike square waves, this is not easy to judge by looking at the scope trace. A spectrum analyzer would provide this info, but are there any theoretical considerations to decide which oscillator comes closest, produces the purest sine wave?
Yes, LC tanks with ideal components. Or phase shift oscillators with ideal components (ideally linearized amplifier, ideal capacitors and resistors) with perfectly matched phase (180°). If you're curious about the topic I recommend to try to simulate it and probe it via FFT if your simulator has it. You can try LTspice which is free or falstad simulator which is even available online. You'll see most generators are giving pure sinewaves or squares, the problems begin when you use real components.
@@pheymee88 "with ideal components", so all we need now is for you to tell us where we can purchase these ideal components. ;) Obviously, the point of the question was which design is best with real components. Perhaps we need to look at the design of high-end signal generators for real-world answers.
Thank you for your great content
At 12:18 Crystals have two resonant frequencies, series resonance and parallel resonance. That may be why you are getting 6.5 MHz instead of 4.0.
You may also want to look at IMSAI Guy's video #1807 Twin T Oscillator using an OP amp.
Another good one❤
Thanks for watching :)
Howdy.
The Hartley needs cap.s on both ends of the tank to block DC. Otherwise ok ...
Regards.
thank you for this!! I needed it :)
I needed a Sine wave to go threw me so I plugged my self into the wall
Excellent presentation!! 👏👏👏👌👌👌👍👍👍😊😊😊
Good stuff. Definitely something to play with... though, isn't there also an option to use an op-amp as an integrator, and integrate a square wave?? (Or, were you doing that somewhere, and I just misunderstood it?)
Also, another way to spell "2𝜋" is "𝜏", which can then be pronounced "tau", instead of "two pi". Just sayin'. ;) (Not sure what a 𝜏 would look like in that font, though! 🤔)
Integrating a square wave gives you a triangle wave. :)
@@SineLab oops, yes, I was missing a step. Integrating the _triangle_ wave, in turn, gives you something that... well, approximates a sine wave, though isn't one in a strictly mathematical sense. :)
Tau is a really bad letter choice to designate 2*pi, especially in the context of electronics. It's already taken to designate the time constant of first order systems. I don't really know why some math folks are so fond of "tau", I've always been used to "2*pi" everywhere since my studies and I'm perfectly fine with that lol.
Do you ever show any simulations using QSPICE ?
Is there any site wave oscillator that can replace inverters?!!
At 5:56 should the formula be Vo / Vin = Z2 / (Z1 + Z2) ?
Danke für das Video. Habe Dich abonniert. LG
How did you determine the 120 degrees phase shifts?
you have clipping in Wien oscillator ... why you do not have clipping in phase shift oscillator?
At 6:18 you said Vo/Vin= Z1/(Z1+Z2), I believe you wanted to say Vo/Vin= Z2/(Z1+Z2)
Tried using low pass filters to get sine out of a 555 square pwm, none worked. Spike sines top and bottom. Even repeating the circuit in falstad, same results. Power supply from a battery.
Very good.😊
Amazing vídeo 🎉
The "wine" bridge? 😂 Joking. Thanks for the video!
Thanks, I learnt a lot!
Glad it was helpful!
@@SineLab there's 101 channels that do basic electronics (another 'what is a diode' video anyone?), so great to see your intermediate level content. Delivery style is great too
At 5:58, Vout/Vin = Z2/(Z1 + Z2) not Z1/(Z1 + Z2). HTH
Thank you for pointing out that typo.
@@SineLab Anything you're planning to do with that correction? Or are you just content with leaving misinformation uploaded as long as it brings you clicks? Not even a pinned comment with correction?
I should note a few issues with this otherwise informative video:
- "Wien" is pronounced "veen", NOT "wine" or "vine".
- One needs to be careful with so-called "sine wave oscillators", as a lot depends on how the resulting waveform signal will be used. In many applications, it is critical to get a low distortion sine wave, and the majority of circuits shown in this video, while 'Sine Lab' proclaimed them to be 'clean sine waves', were actually quite visibly distorted, and a distortion meter would have shown this numerically. It is a bit of an art form to visually recognize the presence of sine distortion, made much more difficult for people who have little time viewing truly 'clean' sine waves (but the visual method, even at its best, can miss significant amounts of distortion). Some things even a beginner can look for: symmetry issues (does the down curve shape match the up curve shape), how pointy are the peaks and troughs, do the sides of the waveform look straight rather than constantly curving, etc.
- In a Wien bridge oscillator, some sort of dynamic amplitude control is needed, in order to minimize clipping and other distortion, and this video mentions this and that the incandescent light bulb (not just any light bulb) is traditionally used for this. And this video correctly points out that an FET may be pressed into service as an alternate to the incandescent bulb. But two LEDs, two diodes, etc; may also be used in place of the bulb, with varying degrees of distortion. However, in the incandescent bulb version, the bulb never gets very warm, and certainly does not get hot enough to incandesce.
- In addition to the five ways of generating a sine wave, most modern professional grade waveform generators are doing so digitally; the precise sine waveform shape is stored numerically in a look-up table, and played back at various speeds; this works very well for very low distortion sine wave generation, but its complexity makes the generation of higher frequencies significantly more expensive.
Thank you.
Thanks 👍
ممنون
What kind of oscillator can be used to create sine waves in the range of 100MHz - 200MHz.
Try playing with a PLL.
Use a DDS chip followed by a low pass filter.
@@gregorymccoy6797 Thank you for the suggestion!!!
@@geekthesteve6215 I'll check it, thank you foe the info
That is commonly done with an LC circuit. Colpitts common base works the best generally.
This gives low power Sine Wave what about sine wave for high power inverters.
just throw it in some linear amplifier? :D Sorry I'm laughing, I just think amplifying the signal is much easier than generating it.
You need to generate a PWM signal that represents the sine wave and feed a converter you choose. With a self oscillating UcD you can get as low THD as 0.0001%, with a buck/boost converter this is about 1-3%, with a full bridge converter and some modulating techniques (like a phase shift or a complementary detector) this is about 0.01-0.1%. Then make a choice according to your generator specs.
Hey, did you know you can also generate sine mechanically by rotating a round sine-shaped magnet and scanning its shape by coils? That's what Hammond organ is doing.
Which is then usually sent through a spinning speaker to create a phase shift effect. You can do amazing stuff with mechanical devices but they tend to be very big and gobble power :)
I thought you were great in "The Office" btw
what is R8 for at 4:12?
It adds some negative feedback to help stabilize the circuit.
Could you do one on triangle waves :D
Maybe sometime in the future...
@@SineLab cool! I am trying to build a class-d but I'm having a hard time making a clean high frequency triangle for sampling
at 06:28 Vo/Vin= Z2/(Z1+Z2) not Z1/(Z1+Z2)
Wine bridge!?
I was expecting something about DAC/DDS comparison. Usually You need to adjust frequency. And THD here I see really bad.
Can we convert DC to Pewor AC Sine wave
Yes, research inverter circuits.
What so you mean by 'cut off' ?
A filter reduces a signal based on its frequency. For example, a "low-pass" filter passes low frequencies and dampens high frequencies. An example where you'd see one is the subwoofer output on a stereo - it filters out all the highs and passes the lows.
The "cut-off" frequency of a filter is (usually) the frequency where the signal is diminished by 3dB. So for a low-pass filter like what he's using here, you'd want the cut-off frequency to be a bit higher than the frequency of the sine wave you want to create. Square waves are really just an infinite number of sine waves added together, so by filtering out the frequencies higher than the one you're looking for you can turn a square wave into a sine wave.
Filters are a pretty big topic and they can get pretty complex, but the basics aren't difficult. You can find lots of introductory articles on them around the 'net.
You forgot one of the most common ways to generate a sine wave that is used on all the arbitrary waveform generators... a microcontroller coupled with a digital to analog converter. 😉
Doesn't have that tube "warmth" brah.
So if a sine wave is, well a sine wave, then just out of curiosity, is there such thing as a cosine wave for electrical circuits? Since it uses trigonometry, then its got something to do with circles or spheres, is there any significance to that? Like, are signals actually 3D and we are just looking at it in 2D for simplicity? This probably sounds dumb af, but im curious.
The waves are just in two dimensions: amplitude (y) and time (x). A cosine wave would look identical to a sine wave except that it would have a 90° phase shift to the left.
💖💖💖💖
Is it ironic that the hardest component to find is a light bulb?🧐
The guy, Max Wien, who invented the Wien bridge is pronounced like "ween" in English.
Please stop whining. Thank you.
4:15 Please don't forget to bypass R7. 10uF would be good, 100uF would be better, but even 1uF would be okay-ish and better than nothing.
Wien is German and, as such, as closely as can be approximated with American phonemes, is pronounced "wean." (Actually it vould be closer to "veen," but ve can't be too choosy.) If it were pronounced "wine" or "vine," it would be spelled wein.
You missed one...
An electric motor, a linear potentiometer and a scotch yoke mechanism.
🙃
How to make a sine wave?
Step 1: Make an Inverter
Step 2:
*_hammers laptop keys_*
Iet ies pronounzed as 'weenbridge'
No wine or whine involved😊
Educated people often mispronounce names or words because they've only read them. At 45 I have a huge list of things I know how to pronounce "correctly" and how it has lived in my brain for 30 years. I pronounce with with a German V "Vienbridge!" when I'm feeling sassy.
@@justovision No. "Educated people" worth the label actually verify what they're talking about, in particular when foreign languages are involved, and don't just parrot everything without a second guess. Because that's not how science works.
@@Anvilshock Language is fluid. It's not science. There is no proof possible because it's all made up. What are you talking about?
I wouldnt call any of these sine waves.If anyone needs a real sine wave with controllable frequency and low thd. Know that It's extremely hard and takes a box of parts. Luckily modern tech has DDS chips that spit out a magical sine wave with low thd (or use a dac if your freq isnt too high)
This guy got 7mhz from a 555timer song wave
Where is walking ring counter dflip flop digital sinewave . where is spwm sine wave .
Don't want any of that digital rubbish !! LOL :)