why doesn't this circuit oscillate

[iVenky]

I have an inverting amplifier and in the feedback path I have a series combination of inductor and capacitor and Resistor

Then the transfer function of the feedback part is sL + 1/sC +R.

So that the 180 degree phase shift frequency is given by

tan-1( img/real) = 180

=> img/real=0
=> img part to zero

so that wL - 1/wC=0
so that

frequency w = 1/sqrt(LC).

Also if the amplifier gain is quite large then A* beta seems to be greater than 1.
But the truth is that it doesn't oscillate. Why is that so?.I know that I am missing something basic.

Thanks in advance.

 

[Sajjadkhan]

I dont know what type of oscillator you are using, i once used buba oscillator and it didn't oscillate in the begining. first thing you should check the basics.
1. starting gain must be greater than 1 so that it amplifies the noise and start oscillating.
2. after that gain must be maintained to 1 so that it wont converge or diverge.

a little error in cap, res and inductor values makes a huge difference. try calculating the component values with multimeter and i bet cap value wont be exact and if you are using 5% tolerance resistors then what can i say. use good capacitors like tantalum and 1% or 0.1% resistors and use a pot to precisely calibrate you gain and 180 degree shift. ceramic capaciotrs wont be good in this case.
Thats all i can say.

 

[FvM]

I have an inverting amplifier and in the feedback path I have a series combination of inductor and capacitor and Resistor

Then the transfer function of the feedback part is sL + 1/sC +R.

Strictly spoken, both statements seem mutual exlusive. You would need an ideal current source to represent the said transfer characteristic with an RLC series circuit and the RLC circuit as shunt. Obviuosly you should show a circuit.

Assuming you actually insert a transfer function of the said type in an inverting feedback path. Then the 180° phase would be varied over 90 to 270° maximum, thus never touch the oscillation condition.

 

[guguwuwu]

You should confirm the following:
(1) (It is an amplifier or an OP amplifier).
(2) The amplifier gain in the oscillation point is large enough too.
(3) The phase shift of the amplifier is 180 degreee too in the oscillation point.
You can make the bandwith and gain of amplifier a vety large value.The best is to post the circuit.

 

[iVenky]

I have attached the circuit

 

[engrMunna]

I have attached the circuit

I am no sure what software you are using to simulate this circuit.....but if its possible can you give some initial values to your nodes? then give one rail value to your output and the complementary rail value to the base. also try simulating with some transient noise enabled.

 

[FvM]

The transistor in the shown circuit won't oscillate for the simple reason, that it misses a DC bias. In this case, you don't need to think further about fulfillling the oscillation condition.

The other point is that a transistor isn't an ideal inverting amplifier. It has additional parameters, e.g. input and output capacitances. Thus I won't exclude, that the circuit can possibly oscillate after adding a bias network. But if so, it's not due the RLC feedback allone.

 

[BradtheRad]

To sustain oscillations, the LC loop (or string) must do these things in a cycle:

* Alternate between delivering a little (or no) bias current to the transistor, then a lot of bias

* Receive just enough jolt from somewhere, to gain back whatever energy it lost

Considerations to help achieve the above:

* The control network needs a certain degree of isolation from the supply rails

* A large capacitor needs more current than a small capacitor, to create the same voltage swing.

* A small coil needs more current than a large coil, to create the same emf and the same flux intensity.

* The transistor usually needs a certain amount of bias current in order to get the action started. The bias level needs to be adjustable, hence a potentiometer is handy for this purpose.

-------------------------

Your layout has some resemblance to a Clapp oscillator. It is distinguished by having a coil and cap in series. Operational circuits have a few more components added.

You may be able to get your setup to work, if you can fulfil all the conditions. Don't be surprised if it requires some reconfiguring. See if you can jolt it into operation somehow. There may be a few oscillations afterward, even if they decay quickly.

I recognize you are using Falstad's simulator. It is easy to use. I like it and I use it as an alternate to my own homebrew animated simulator.

I have a video of Clapp oscillators in animated simulation on Youtube:

https://youtu.be/wKnarrvynIw

 

[FvM]

In a short: If you discuss the circuit as Clapp oscillator implementation, than you should add, that it exactly misses the additional capacitors that make a common emitter Clapp circuit oscillating.

 

[BradtheRad]

In a short: If you discuss the circuit as Clapp oscillator implementation, than you should add, that it exactly misses the additional capacitors that make a common emitter Clapp circuit oscillating.

Yes. I only made vague mention of this by saying 'operational circuits have a few more components added.'

If the OP continues to experiment, he may find a way to implement a series LC oscillating loop, in a different way than the Clapp oscillator. I keep a collection of oscillator schematics, and only the Clapp type has the series LC.

Nevertheless a different method may be discovered by someone who never heard it couldn't be done. So I don't want to say it has to be done a certain way.

Checking into its close cousin, there are simple oscillators built around an LC parallel loop (tank loop). I list a few below.
However two of them seem to require input of some kind in order to operate:

Page 77. Figure 2-10A. (Scroll about 1/3 of the way down.)
www.hnsa.org/doc/neets/mod09.pdf


and


...while the third seems to need something I have not yet figured out, since oscillations fade in my simulator.

www.falstad.com/circuit/e-eclosc.html

 

[FvM]

I agree, that the Pierce Oscillator can be seen similar to the original circuit, if you replace the crystal with a RLC series circuit. Also in this case, that additional capacitors are a necessary condition to make the circuit oscillate.

 

[godfreyl]

Hi iVenky

Your circuit doesn't oscillate because the voltage gain (and thus the feedback) is negative, not positive.

frequency w = 1/sqrt(LC).

At that frequency, the phase of the impedance of your network is zero degrees, not 180 degrees.

An easy way to make an oscillator with a feedback network like that is to use a long tailed pair for the amplifier, as in the pic below.

Cheers - Godfrey

 

[BradtheRad]

I can confirm the layout in post #12 works.

A simulation shows it provides alternating pulses at the outputs.

I substituted a value of 100nF for the capacitor.