# problem with oscillator

#### shreyas_bhat

Joined Jul 26, 2004
47
I am trying to design a Twin T oscillator. The Twin T network was connected between the inverting i/p and the output of an op-amp, and the non-inverting input was grounded. I looked up this design over the net, but this doesnt seem to oscillate. However, when a feedback resistor is connected between the output and non-inv input, it started to oscillate. Could somebody please explain the working ?

Also, I expected that the frequency of oscillation be a function of the resistor and capacitor in the Twin T circuit. However, even a variation in the positive feedback resistor seems to vary the frequency, which is not what I want. Can somebody suggest certain steps to overcome this problem.

Thanks

#### Brandon

Joined Dec 14, 2004
306
Originally posted by shreyas_bhat@Feb 2 2005, 04:58 PM
I am trying to design a Twin T oscillator. The Twin T network was connected between the inverting i/p and the output of an op-amp, and the non-inverting input was grounded. I looked up this design over the net, but this doesnt seem to oscillate. However, when a feedback resistor is connected between the output and non-inv input, it started to oscillate. Could somebody please explain the working ?

Also, I expected that the frequency of oscillation be a function of the resistor and capacitor in the Twin T circuit. However, even a variation in the positive feedback resistor seems to vary the frequency, which is not what I want. Can somebody suggest certain steps to overcome this problem.

Thanks
[post=4986]Quoted post[/post]​
Only way to get oscillations is through feedback normally, especially with an opamp.

Its like a swing. If you have negative feedback, it likes someone pushing the swing from one side while on the other side someone grabs the swing and slows it down. The whole negative thing is that the signal is 180 degrees out of phase compared to the orignal. Now, if you happen to have positive feedback is likes both people pushing the swing in the same direction so it keeps going higher and higher and your oscillator starts. For the swing, its frequency of oscillations has to do with its length of chain, for your circuit it has to do with component selection.

The feedback resistor shouldn't have much of any effect on oscillations, rather it should effect the rate at which your oscillations reach their peak value.

Could you possibly post a schematic? Wouldn't mind checking it out..

#### vineethbs

Joined Nov 14, 2004
56
one sec ,shreyas ! u hav a twin T n/w that is basically a notch filter , i can't really understand the ckt , so correct me if i am wrong .

the notch filter is your freq selection ckt , so actually to get an oscillation at a particular freq u can subtract the signal from the o/p of the notch filter , this wud give u a bandpass response

so the oscillation wud occur only if u feedback the o/p (no filtering at + terminal ) - the notch filter o/p , so the addition of a feedback resistor is important , but it shudn't really affect the freq , i think

thanx brandon for the exp of feedback!

#### shreyas_bhat

Joined Jul 26, 2004
47
Originally posted by Brandon@Feb 2 2005, 06:37 PM
Only way to get oscillations is through feedback normally, especially with an opamp.

Its like a swing. If you have negative feedback, it likes someone pushing the swing from one side while on the other side someone grabs the swing and slows it down. The whole negative thing is that the signal is 180 degrees out of phase compared to the orignal. Now, if you happen to have positive feedback is likes both people pushing the swing in the same direction so it keeps going higher and higher and your oscillator starts. For the swing, its frequency of oscillations has to do with its length of chain, for your circuit it has to do with component selection.

The feedback resistor shouldn't have much of any effect on oscillations, rather it should effect the rate at which your oscillations reach their peak value.

Could you possibly post a schematic? Wouldn't mind checking it out..
[post=4990]Quoted post[/post]​

#### shreyas_bhat

Joined Jul 26, 2004
47
I tried to attach the schematic but it doesnt go through. Anyways, lemme explain once again. Between the output of the op-amp and its inverting input, I have the Twin T filter circuit. And in between the o/p of the op-amp and the non-inverting i/p is a resistor (R4), and there is also a resistor between the non-inv pin and the ground. The only source of power is the DC to drive the op-amp.
Variation in R4 causes the oscillation frequency to vary, which should not ideally happen. If I had to retain the Twin T circuit and build an oscillator out of it, are there any suggestions ?

Thanks
Shreyas

Originally posted by shreyas_bhat@Feb 2 2005, 04:58 PM
I am trying to design a Twin T oscillator. The Twin T network was connected between the inverting i/p and the output of an op-amp, and the non-inverting input was grounded. I looked up this design over the net, but this doesnt seem to oscillate. However, when a feedback resistor is connected between the output and non-inv input, it started to oscillate. Could somebody please explain the working ?

Also, I expected that the frequency of oscillation be a function of the resistor and capacitor in the Twin T circuit. However, even a variation in the positive feedback resistor seems to vary the frequency, which is not what I want. Can somebody suggest certain steps to overcome this problem.

Thanks
[post=4986]Quoted post[/post]​

#### pebe

Joined Oct 11, 2004
626
If I read you correctly, you had a circuit which had the +ve input (non-inverting) grounded, and it didn't work.

Now you have connected a resistor from the o/p to the +ve input. That's a receipe for disaster because you are introducing uncontrolled positive feedback and the results will be unpredictable. I suggest you go back to your original circuit and find out why it is not oscillating.

It looks like it could have been wrong biasing on the inputs. Measure the DC voltage at the output. If it is running with correct bias it should be about half the supply voltage. If it is very much higher or lower it is likely that you have a bias problem, ie. the DC voltages difference between the two inputs is too big.

#### vineethbs

Joined Nov 14, 2004
56
but pebe the ckt actually takes the difference between the notch filter output and the positive feedback ,so is it uncontrolled ?

i think mebbe the gain of the twin T network mebbe a problem

#### pebe

Joined Oct 11, 2004
626
Hi vineethbs,
You are most probably right. I don't know the circuit but as originally designed it, presumably, worked.

I was thinking that as the original circuit had its non-inverting input earthed, it seemed rather drastic to introduce positive feedback which would probably not improve matters.

It's a pity we couldn't see the circuit. shreyas_bhat has tried to upload without success. I've tried to upload with two different file formats - without success. Can anyone say which file extensions are acceptable to this forum?

#### vineethbs

Joined Nov 14, 2004
56
hello all , i simulated a ckt with a twin T n/w in the feedback path , but what am getting is a square wave , too much gain i think , reducing the gain of the amp is not really solving the problem . what to do :blink: