Hartley Oscillator: How Does It Produce a Sine Wave?

[Amith2006]

Why does Hartley Oscillator produce a sine wave?Is it just because Barkhausen condition is satisfied?

 

[waht]

Because the LC tank produces a sine wave. Positive feedback is coupled in the inductor in a Hartely oscillator.

 

[Amith2006]

But a phase shift oscillator produces a sine wave without a tank circuit.Is the RC network responsible for sinusoidal oscillations in this case?

 

[waht]

When the phase shift oscillator starts up, the whole spectrum of frequencies is produced by noise or random motion of electrons. A phase of 360 degrees of the RC loop will only occur at a particular frequency. Therefore, the positive feedback loop will reinforce only that one frequency of 360 degrees phase shift, and that depends of values of RC.


And similarly, with Hartley and Colpitts oscillators, the start up kick is produced by noise, but LC is naturally resonates at a sinusoidal frequency, so noise is more inclined to be at LC already. You can get a better selection and higher Q.


The Barkhausen condition says simply, if there is positive feedback with a phase 0 or 360 degrees, you will get oscillation. That doesn't nesessarly mean a sine wave, you have square wave, or pulse oscillators working because of this condition.

 

[antoker]

waht: are you sure about 360? ;) You're required to get a 180 degree phase shift in your feedback loop in order to be able to sustain oscillation at a particular frequency. Start up kick is not produced by a noise, but by an initial transition between zero volts and up to a Vcc level(resulting in step in inductor loop), which produces a transient in a feedback circuit allowing the circuit to start oscillation and continue to "chase its own tail"

P.S take a look at the png file attached, voltage is measured between the input/output of a feedback circuit in simple transistor circuit.

P.P.S Sometimes designers add a "tank circuit" to a already stable oscillator in order to use it as a Band-Pass filter, since oscillators tends to produce a lot of unnecessary harmonics.

 

[Averagesupernova]

waht: are you sure about 360? ;) You're required to get a 180 degree phase shift in your feedback loop in order to be able to sustain oscillation at a particular frequency. Start up kick is not produced by a noise, but by an initial transition between zero volts and up to a Vcc level(resulting in step in inductor loop), which produces a transient in a feedback circuit allowing the circuit to start oscillation and continue to "chase its own tail"

It is most certainly 360 degrees. The 180 you refer to applies due to the fact that the amplifier you are feeding back into is an inverting amplifier and inherently shifts the phase 180 degrees. Oscillators most certainly can be started by noise in the circuit. A transient is not necessarily required. It depends on the feedback fraction as well as a number of other things that I have probably forgotten.

P.S take a look at the png file attached, voltage is measured between the input/output of a feedback circuit in simple transistor circuit.

Yep, shows 180 degree phase shift like you said. Says nothing about the amplifier being an inverting amp.

P.P.S Sometimes designers add a "tank circuit" to a already stable oscillator in order to use it as a Band-Pass filter, since oscillators tends to produce a lot of unnecessary harmonics.

Some oscillators are more rich in harmonics than others. Some applications require cleaner outputs than others. I wouldn't say that that the oscillator is made into a bandpass filter. Technically it already is a bandpass filter. Anything added to an already functioning oscillator in order to clean up the signal is most likely considered a separate stage in the chain.

 

[antoker]

Averagesupernova: I think you misunderstood me, when I was talking about oscillator as a BP-filter, what intended to say was: Since, oscillators tend to produce more harmonics, tank circuits are introduced on collector side of transistor in order to clean up the signal.

Oscillators most certainly can be started by noise in the circuit.

How much noise are we talking about here? I would really like to see an oscillator with 0V Vcc being started & sustained by some random noise in the system. Again, my bad english tends to create misunderstanding.

Yep, shows 180 degree phase shift like you said. Says nothing about the amplifier being an inverting amp.

Yes, I was referring to 180 shift between the input/output network, I wasn't talking about the whole transistor stage

 

[Averagesupernova]

Averagesupernova: I think you misunderstood me, when I was talking about oscillator as a BP-filter, what intended to say was: Since, oscillators tend to produce more harmonics, tank circuits are introduced on collector side of transistor in order to clean up the signal.

I guess I'd have to see the circuit. I'd say that the tank circuit you refer to is simply the resonant element in the oscillator.

Oscillators most certainly can be started by noise in the circuit.

How much noise are we talking about here? I would really like to see an oscillator with 0V Vcc being started & sustained by some random noise in the system. Again, my bad english tends to create misunderstanding.

Take a step back. No one has said anything about an oscillator taking of and running with no power supply. If you power up an oscillator and then stop it by loading the inductor or some other subtle means, some oscillators can start on their own simply due to thermal noise.

Yes, I was referring to 180 shift between the input/output network, I wasn't talking about the whole transistor stage

A mistake you should avoid. Some oscillators have feedback from the emitter to the base. This arrangment has a zero degree phase shift through the transistor.

 

[antoker]

Well, here we have a oscillator, where collector resistance was replaced by L2&C3 to improve the oscillation by removing some harmonics, by adjusting tank circuit-values, better filtering can be achieved (Note: just made it in two minutes or so, didn't really calculate any values).

If you power up an oscillator and then stop it by loading the inductor or some other subtle means, some oscillators can start on their own simply due to thermal noise.

Now I can agree with you, having inductor loaded, kills the oscillation and then some noise will be sufficient to start the oscillation, while having constant current flowing. But that is not the point. Initially, when power is not on, transition between 0-vcc level starts the oscillation, i wasn't thinking about stopping the oscillator and then starting it again. We were not discussing the means by which the oscillation can be started and stopped, we're discussing the initial oscillation build-up.

 

[Averagesupernova]

I really don't care to argue over what we call what. It would generally be accepted though that the circuit you posted (or at least part of it) is a class C amplifier. By definition, this is a bandpass filter. Aso, in your example we could remove C1, C2, L1 and R7 and replace them all with a small toroidal transformer that shifts the phase 180 degrees and have virtually the same result. Or for that matter just replace those 4 components with another inverting amplifier. Then the capacitor and inductor in the collector circuit of Q1 are considered the resonant circuit. So we can twist things around as long as we want and argue about it. What's the point? I have never said that the startup transient (technically it IS noise) didn't contribute to an oscillator starting. My main points were that thermal noise can cause circuits to go into oscillation and that a circuit requires an approximate 360 degree phase shift TOTAL to oscillate.
-
P.S. Which tank circuit do you consider to be the resonant element in your posted circuit?

 

[waht]

You can do an experiment, build an oscillator on a breadboard, and make you sure it oscillates. Now disconnect the positive feedback path and power up the oscillator again. What happens now? It doesn't oscillate. (It shouldn't unless your fingers and the surroundings can couple back some positive feedback) Now, while the oscillator is still live, reconnect the positive feedback path and it magically starts oscillating again. There is no transients now.