Would love to know the harmonic distortion on the output. Did you build a beautiful sine wave or a not quite perfect one? Interesting how the final input? to the transistor looks nothing like a sine wave, more like something half-rectified, but other end of the amp does look just like one.
@@kensmith5694 Hewlett's Wein bridge oscillator with tungsten incandescent lamp in the feedback loop... that comes to mind. (Hewlett's design had very low distortion for the time.)
@@kensmith5694 RG-59A/U PE (Belden 8241) veolcity factor is 0.66 0.66 * c = 198000000 m/s where c is the free space speed of light. for a 10 KHz oscillator 1 wavelength in the cable is 198000000 m/s / 10000 1/s = 19800 m need 180 degrees phase shift (180 degrees / 360 degees) * 19800 m = 9900 m or 9.9 Km. If you raise the frequency by a thousand times 10 MHz, and a better transistor then, the cable has to be 9.9 meters. For 100 MHz (definitely need a much better transistor) then just about a meter of cable. (But... might need to take attenuation into account.)
A very nice demonstration of the oscillator with the phase shift per stage. With three identical stages each one advances the phase by 60 degrees to get a sum of 180 degrees. In the practical example it is not really ideal and also a bit nonlinear at the transistor. In the comparison of the signals on the scope you see that the phase per stage is not delayed but advanced. A delay would go with low pass elements, in this example we have high pass elements.
@@nigelrhodes4330 Never seen that. Usually you need 3 filter stages. The oscillator is only stable if there is more than 360 total phase shift at higher frequencies. That's not possible with just 2 stages... or is it? (Usually R and C chosen so that each RC stage has more than 60 degrees phase shift at a frequency above the desired oscillation.)
It amuses me when people talk about phase changes as one can only see phase change when one sees the past and the present on an oscilloscope!!!!!! In real time, " The Present" all the voltages appear at the same time but they have various magnitudes, due to the memory abilities of the " reactive elements.". When the output of the oscillator is zero voltage, there is one or more capacitor which has memorised a voltage and this is what keeps the circuit alive and working.................THE MEMORY OF THE CAPACITOR AT THE PRESENT TIME, It all has to do with the RATE OF CHANGE OF THE INPUT WITH RESPECT TO THE OUTPUT. All the action takes place at the present time and one should not use a two dimensional space to see a sinewave The voltage at the output of an oscillator is between two terminals all occurring at the same time, The present is such a short duration of time and one needs memory , rather than " phase".
I put a 8ohm speaker coming from the emitter and can hear a pretty little sine wave. I would like to amplify so i can hear it with out the speaker directly against my ear but when i try the signal disappears. Im assuming it has something to do with the high output impedance but im knew to these things. If anyone has advice on how to amplify the signal for a speaker it would be greatly appreciated. Thanks Imsai Guy for all these videos that inspire thinking and creativity!!
With 9V supply, the circuit as shown has about 10% distortion. If you add a 50R in series with the 10uF, you will get about 3%. With 62 Ohms you would get it down to about 1% but there is a danger of it not oscillating at all.
How do you know this? Have you somehow calculated it or tried and measured? As far I know electronics the resistor you added to the emitter capacitor adds some negative feedback to signal flow of the amplifier and maybe thus it improves linearity of it. Am I right? I'll try it.
@@erykbaradziej3639 I modeled it in spice to confirm my guesstimate. I estimated the amount the gain was greater than what was needed to sustain oscillations. I then assumed that the gain was decreased to the amount that just makes a constant amplitude by clipping. I got that it would be something near 10% as in something more than just a few and less than 30%. I went with spice rather than trying to work out the math in detail.
It's interesting how clean the sine shape of the output is, even though the intermediate phase-shift filter stages show increasing amounts of distortion in the shape. Perhaps some nonlinearity of the output amplifier cleans it up again?
You see these oscillators all the time In old electric organs for their vibrato. They make a good vibrato, very smooth, better than a triangle wave that synths have for their lfo.
Many thanks for sharing this with us. I have a quick question: What is the function of the resistors between the capacitors? Are they being used as filters? What happens if the value of the resistors are changed? Does it affect the oscillation frequency? 👍
Each [series C + parallel R] is its own little RC filter, in this configuration a high pass filter. I'm sure there's an IMSAI Guy video on the opposite config which forms a low pass filter (or maybe both). Each one only lets signals of certain bandwidth or above pass through. In the end, there's one bandwidth which 1. is let through by the high-pass filters, and 2. is phase shifted by each filter stage so that it ends up 180 degrees out-of-phase at the input to the transistor amp again; finally the transistor amp does another 180 degree phase shift (otherwise known as inverting) to give a full 360 degrees positive feedback loop to that one frequency. Changing the resistor(s) will change the nature of the RC filter(s), and yes therefore change its phase-shift response, and so change the optimal frequency which receives an amplification by the circuit.
@@ivolol I think you need better than 180 degrees phase shift... so greater than 60 degrees per stage (if you use the same capacitor value and resistor value... which is usually done.) The idea of the low pass filter is it limits the oscillation to the desired frequency. (That might be fiddly.)
The C-R network as shown does not cause delay. In fact it advances the phase! (As actually shown by the scope traces). This in principle is prone to producing harmonic distortion (think high pass filter characteristics). So, you depend on transistor slight non-linearity for stabilizing the amplitude and producing decent sine wave. . In this respect - can you make an oscillator without the phase shift? As it happens, a book by James G Holbrook, Laplace Transforms for Electronic Engineers, (Library of Congress Catalog Card 59-12607) outlines some circuits that can work without the inverting gain stage. I believe he patented it as a tube circuit, cathode follower. The book was first published in 1959, so the patent must be expired years ago. By the way, he also analyzed the R-C circuit turned into low pass, and found it required a gain of -29. On the other hand, if you have a basic Twin-T circuit bridged with an additional longitudinal capacitor, you can get at resonance 0 phase shift and a few % gain (say 1.06 or similar). That could be just enough to cover the miscellaneous losses in in the components and wiring to function as a suitable feedback element in an emitter follower (or cathode follower - or source follower FET) buffer to form a sine wave oscillator. See, one way to make an RC oscillator, even without a voltage amplifier...
@@InssiAjaton Any oscillation involved a phase shifted signal that gets imposed over the original to create the feedback at the right delay. In a simple LC tuned circuit the phase delay comes from the signal reflecting back and forth between the capacitor and the inductor.
Here we go with the here's again... Thanks shango066, attention span is now compromised. Great visual explanation man, but it would be neat to add ELI the ICE man...or phase shifts that happen because of the RC/LCnetworks.
@@IMSAIGuy a physicist is all about mathematics. I read a book of practical electronics by Paul Scherz who is also a physicist. It is an excellent book. I recommend you check it sir it presents material on paper just the way you do on screen sir
@@awaismushtaq5719 yes, I have a degree in Physics and a degree in Mathematics. I avoid the math here on this channel to reach more people and impart an understanding at a more basic level.
@@IMSAIGuy I am an engineer not an electrical engineer though. Every engineer studies electronics as it is deeply integrated into every engineering discipline. I learned electronics by reading and messing with bread boards and components. Burnt loads of them eventually cutting my losses when I got the grip. First successful thing I built was an AM transmitter with about 50 yards of range. When I was testing it a police vehicle was parked right outside our house. They were there for a different matter yet they had scanners. They picked up my transmission which I was playing from my computer cursing a local politician. They knocked at my door and confiscated all of it. Warned me not to do it again or I'll end up in jail. Electronics is fun. Yet that didn't stop me from building an other one 10 watts of solid power. Still use it off and on but carefully. Your videos are fun, loads of fun. Primarily because you learnt electronics the way I did though you are far better than I'm. Good luck sir!
Nice and interesting! Thank you! Could it's freq be varied by varying the 3 10k simultaneusly? With some sort of VCR... FETs for instance? What about the distortion of the sine with this topology? It's purity
@@IMSAIGuy Thank you, I'll take a look! My idea is to keep it simple and analog. I usually use DIY vactrols (LED+LDR) for instance and will try soon FETs as VCR. But I will study the MCP4131 for sure, thanks!
Thanks, nice demo. I however don't understand why is this called a _positive_ feedback. Isn't the feedback, after shifting 180°, being an _inversion_ of the output? Calling this a negative feedback would make much more sense to me...
If I'm not mistaken, the signal is inverted 180 by the transistor, it is then routed back through the series of RC circuits, putting the signal back into phase. So the feedback path is adding to the signal, not subtracting.
I had a class that touched on the subject of phase shift oscillators. My hazy recollection is that there are two key elements. 1... there needs to be 360 degrees of phase shift in the feedback loop, and 2... the loop gain needs to be exactly 1. If the gain is less, the oscillation will die out, and if it is more, it starts bumping into the supply rail(s). As such, it's a bit incorrect to say the feedback is either positive or negative, but oscillators are generally considered to have positive feedback.
@@SkyhawkSteve That's the Barkhausen criteria... but he used to get mad if one didn't realize that's not enough for oscillator stability. (They said however... his bark was than his hausen... if that makes any sense... at all.)
@@SpentelectronsI also remember from college days in order for a filter with a feedback loop to oscillate, you have to have an odd and min 3rd order (3 caps) to start oscillating. I maybe wrong. It's been 30-40 years since I saw this circuit. The unity gain would be balanced out by the filters. I remembered I built a chebyshev or Butterworth filter and made the circuit to oscillate. Oh no .... 3rd order transfer function, Laplace transforms, all those things that I hated and never thought of until today!
as shown this has 10K impedance. for many applications this is just fine, or follow this with an emitter follower for low impedance. this circuit works into 100's of megahertz which op amps will have a hard time. there is no one size fits all in electronics.
![Cody Johnson - I'm Gonna Love You (with Carrie Underwood) [Official Music Video]](http://i.ytimg.com/vi/yy9PuYMU29g/mqdefault.jpg)
Would love to know the harmonic distortion on the output. Did you build a beautiful sine wave or a not quite perfect one?
Interesting how the final input? to the transistor looks nothing like a sine wave, more like something half-rectified, but other end of the amp does look just like one.
Capacitance.
The distortion as shown will be about 10%. Add a 50R in series with the 10uF to make it about 3%
There are ways to get lower distortion if needed.
@@kensmith5694 Hewlett's Wein bridge oscillator with tungsten incandescent lamp in the feedback loop... that comes to mind. (Hewlett's design had very low distortion for the time.)
coax makes a handy phase shifter with less loss, but it'll have to be pretty long to oscillate at 600 Hz 🙂
Yes, about the longest delay I ever did intentionally that way was 2uS. It was a roll of coax.
@@kensmith5694 RG-59A/U PE (Belden 8241) veolcity factor is 0.66
0.66 * c = 198000000 m/s where c is the free space speed of light.
for a 10 KHz oscillator 1 wavelength in the cable is
198000000 m/s / 10000 1/s = 19800 m
need 180 degrees phase shift
(180 degrees / 360 degees) * 19800 m = 9900 m or 9.9 Km.
If you raise the frequency by a thousand times 10 MHz, and a better transistor then,
the cable has to be 9.9 meters. For 100 MHz (definitely need a much better transistor) then just about a meter of cable. (But... might need to take attenuation into account.)
A very nice demonstration of the oscillator with the phase shift per stage. With three identical stages each one advances the phase by 60 degrees to get a sum of 180 degrees. In the practical example it is not really ideal and also a bit nonlinear at the transistor.
In the comparison of the signals on the scope you see that the phase per stage is not delayed but advanced. A delay would go with low pass elements, in this example we have high pass elements.
I notice they are using 3 tantalum capacitors could you not just have easily done this with two electrolytic as they move the phase by 90?
@@nigelrhodes4330 Never seen that. Usually you need 3 filter stages. The oscillator is only stable if there is more than 360 total phase shift at higher frequencies. That's not possible with just 2 stages... or is it? (Usually R and C chosen so that each RC stage has more than 60 degrees phase shift at a frequency above the desired oscillation.)
@@nigelrhodes4330 , Hi! No.
It amuses me when people talk about phase changes as one can only see phase change when one sees the past and the present on an oscilloscope!!!!!!
In real time, " The Present" all the voltages appear at the same time but they have various magnitudes, due to the memory abilities of the " reactive elements.". When the output of the oscillator is zero voltage, there is one or more capacitor which has memorised a voltage and this is what keeps the circuit alive and working.................THE MEMORY OF THE CAPACITOR AT THE PRESENT TIME,
It all has to do with the RATE OF CHANGE OF THE INPUT WITH RESPECT TO THE OUTPUT. All the action takes place at the present time and one should not use a two dimensional space to see a sinewave The voltage at the output of an oscillator is between two terminals all occurring at the same time, The present is such a short duration of time and one needs memory , rather than " phase".
Each RC shifts it about 60 degrees so I'm total 180 degrees and the rest done by amplifier. Congrats!
I put a 8ohm speaker coming from the emitter and can hear a pretty little sine wave. I would like to amplify so i can hear it with out the speaker directly against my ear but when i try the signal disappears. Im assuming it has something to do with the high output impedance but im knew to these things. If anyone has advice on how to amplify the signal for a speaker it would be greatly appreciated. Thanks Imsai Guy for all these videos that inspire thinking and creativity!!
With 9V supply, the circuit as shown has about 10% distortion. If you add a 50R in series with the 10uF, you will get about 3%. With 62 Ohms you would get it down to about 1% but there is a danger of it not oscillating at all.
How do you know this? Have you somehow calculated it or tried and measured? As far I know electronics the resistor you added to the emitter capacitor adds some negative feedback to signal flow of the amplifier and maybe thus it improves linearity of it. Am I right? I'll try it.
@@erykbaradziej3639 I modeled it in spice to confirm my guesstimate. I estimated the amount the gain was greater than what was needed to sustain oscillations. I then assumed that the gain was decreased to the amount that just makes a constant amplitude by clipping. I got that it would be something near 10% as in something more than just a few and less than 30%. I went with spice rather than trying to work out the math in detail.
Now I understand, thank you for the explanation.@@kensmith5694
It's interesting how clean the sine shape of the output is, even though the intermediate phase-shift filter stages show increasing amounts of distortion in the shape. Perhaps some nonlinearity of the output amplifier cleans it up again?
Love the chair creaking. The sound needs examining on a ‘scope and a SA😊
You see these oscillators all the time In old electric organs for their vibrato. They make a good vibrato, very smooth, better than a triangle wave that synths have for their lfo.
love this kind of thing.
Many thanks for sharing this with us. I have a quick question: What is the function of the resistors between the capacitors? Are they being used as filters? What happens if the value of the resistors are changed? Does it affect the oscillation frequency? 👍
Each [series C + parallel R] is its own little RC filter, in this configuration a high pass filter. I'm sure there's an IMSAI Guy video on the opposite config which forms a low pass filter (or maybe both). Each one only lets signals of certain bandwidth or above pass through. In the end, there's one bandwidth which 1. is let through by the high-pass filters, and 2. is phase shifted by each filter stage so that it ends up 180 degrees out-of-phase at the input to the transistor amp again; finally the transistor amp does another 180 degree phase shift (otherwise known as inverting) to give a full 360 degrees positive feedback loop to that one frequency.
Changing the resistor(s) will change the nature of the RC filter(s), and yes therefore change its phase-shift response, and so change the optimal frequency which receives an amplification by the circuit.
@@ivolol I think you need better than 180 degrees phase shift... so greater than 60 degrees per stage (if you use the same capacitor value and resistor value... which is usually done.) The idea of the low pass filter is it limits the oscillation to the desired frequency. (That might be fiddly.)
The C-R network as shown does not cause delay. In fact it advances the phase! (As actually shown by the scope traces). This in principle is prone to producing harmonic distortion (think high pass filter characteristics). So, you depend on transistor slight non-linearity for stabilizing the amplitude and producing decent sine wave.
.
In this respect - can you make an oscillator without the phase shift? As it happens, a book by James G Holbrook, Laplace Transforms for Electronic Engineers, (Library of Congress Catalog Card 59-12607) outlines some circuits that can work without the inverting gain stage. I believe he patented it as a tube circuit, cathode follower. The book was first published in 1959, so the patent must be expired years ago. By the way, he also analyzed the R-C circuit turned into low pass, and found it required a gain of -29. On the other hand, if you have a basic Twin-T circuit bridged with an additional longitudinal capacitor, you can get at resonance 0 phase shift and a few % gain (say 1.06 or similar). That could be just enough to cover the miscellaneous losses in in the components and wiring to function as a suitable feedback element in an emitter follower (or cathode follower - or source follower FET) buffer to form a sine wave oscillator. See, one way to make an RC oscillator, even without a voltage amplifier...
@@InssiAjaton Any oscillation involved a phase shifted signal that gets imposed over the original to create the feedback at the right delay. In a simple LC tuned circuit the phase delay comes from the signal reflecting back and forth between the capacitor and the inductor.
Wow just one transistor, HP sure did waste a lot of parts! 🙂
Here we go with the here's again... Thanks shango066, attention span is now compromised. Great visual explanation man, but it would be neat to add ELI the ICE man...or phase shifts that happen because of the RC/LCnetworks.
ELI likes this video: ruclips.net/video/34G4pzWr9kA/видео.htmlsi=AWkcR0PcSpMyD2lz
Fun and interesting. Thanks.
You are not into nerdy mathematics that much, are you sir?
I avoid it on this channel. Lots of other places on RUclips if that is of interest.
@@IMSAIGuy a physicist is all about mathematics. I read a book of practical electronics by Paul Scherz who is also a physicist. It is an excellent book. I recommend you check it sir it presents material on paper just the way you do on screen sir
@@awaismushtaq5719 yes, I have a degree in Physics and a degree in Mathematics. I avoid the math here on this channel to reach more people and impart an understanding at a more basic level.
@@IMSAIGuy cool I like it
@@IMSAIGuy I am an engineer not an electrical engineer though. Every engineer studies electronics as it is deeply integrated into every engineering discipline. I learned electronics by reading and messing with bread boards and components. Burnt loads of them eventually cutting my losses when I got the grip. First successful thing I built was an AM transmitter with about 50 yards of range. When I was testing it a police vehicle was parked right outside our house. They were there for a different matter yet they had scanners. They picked up my transmission which I was playing from my computer cursing a local politician. They knocked at my door and confiscated all of it. Warned me not to do it again or I'll end up in jail. Electronics is fun. Yet that didn't stop me from building an other one 10 watts of solid power. Still use it off and on but carefully. Your videos are fun, loads of fun. Primarily because you learnt electronics the way I did though you are far better than I'm. Good luck sir!
How would you go about measuring spin of the electron?
paint a stripe on it.
oh sure you can do this too: Stern-Gerlach experiment
Nice and interesting! Thank you! Could it's freq be varied by varying the 3 10k simultaneusly? With some sort of VCR... FETs for instance? What about the distortion of the sine with this topology? It's purity
you could try these: ruclips.net/video/9SzruMuWF8E/видео.htmlsi=xGmGTg8UY5943kih
@@IMSAIGuy Thank you, I'll take a look! My idea is to keep it simple and analog. I usually use DIY vactrols (LED+LDR) for instance and will try soon FETs as VCR. But I will study the MCP4131 for sure, thanks!
I don't know about the resistors, but my 10nF caps are tied up in another project, so I used 6.8nF ones. My frequency was 836Hz.
Thanks, nice demo.
I however don't understand why is this called a _positive_ feedback.
Isn't the feedback, after shifting 180°, being an _inversion_ of the output?
Calling this a negative feedback would make much more sense to me...
If I'm not mistaken, the signal is inverted 180 by the transistor, it is then routed back through the series of RC circuits, putting the signal back into phase. So the feedback path is adding to the signal, not subtracting.
I had a class that touched on the subject of phase shift oscillators. My hazy recollection is that there are two key elements. 1... there needs to be 360 degrees of phase shift in the feedback loop, and 2... the loop gain needs to be exactly 1. If the gain is less, the oscillation will die out, and if it is more, it starts bumping into the supply rail(s). As such, it's a bit incorrect to say the feedback is either positive or negative, but oscillators are generally considered to have positive feedback.
@@SkyhawkSteve That's the Barkhausen criteria... but he used to get mad if one didn't realize that's not enough for oscillator stability. (They said however... his bark was than his hausen... if that makes any sense... at all.)
@@SpentelectronsI also remember from college days in order for a filter with a feedback loop to oscillate, you have to have an odd and min 3rd order (3 caps) to start oscillating. I maybe wrong. It's been 30-40 years since I saw this circuit. The unity gain would be balanced out by the filters. I remembered I built a chebyshev or Butterworth filter and made the circuit to oscillate. Oh no .... 3rd order transfer function, Laplace transforms, all those things that I hated and never thought of until today!
Output impedance is too high. Op-amp oscillator is better.
as shown this has 10K impedance. for many applications this is just fine, or follow this with an emitter follower for low impedance. this circuit works into 100's of megahertz which op amps will have a hard time. there is no one size fits all in electronics.
If you want a low impedance, put a 50R in series with the 10uF and take the signal from there.
It will be small but low impedance.
@@IMSAIGuy maybe a 2N2222A for a better ft... if you're going to 300 MHz?
@@kensmith5694 r_eb is about 100 ohm. You'll cut down the gain a lot. A minimum gain of 29 is needed for the amplifier.