Twin-T Notch Oscillator

After I worked out how the circuit functions, I liked the circuit that much I've updated it. http://everycircuit.com/circuit/4883274114793472

Excellent Jason, i continue here on your circuit, lets make a little step back and think a little why its now working and why on this frequency +-163 and not 159 like the simple math said on a perfect notch, which anyway can not work. Check this http://everycircuit.com/circuit/4977419797921792 i hope you can see the link 4977419797921792

Sorry, it might look pedantic, but its importante to understand and to get the best THD from such an oscillator, i hope you continue to follow me on this, we are not finished, but almost 😁

So i need a little signal to continue with my last circuit. See you

@hurz thanks get yiur important point. it is interesting,, that the notch oscillator oscillates at a frequency that is suppose to attenuate when is filtering. ( lowest gain!) one!?

Think @jason has gone to melt some nails for his unfinished inductor. Any way thx. @ jadon for the post, with @ hurz input, made it worth while for "i" at least.

Wait kiani, we are not finished and do not underestimate Jason! He will follow quickly.

So as Kiani cant wait, here is the final part lets see if Jason likes it. Jasons feedback is importante for me to see how close/good my explanations are. Anyway, any technical feedback is welcome for all readers i hope. Here the final circuit http://everycircuit.com/circuit/4792484747280384

Such an oscillator works at the most attenuated frequency, because everything in it is inversed. When working with the positive input terminal you make your circuit with a high Q bandpass filter (positive notch). When working with the negative terminal the you use a notch filter. Also, the notch filter introduces a 180dgs phase shift at the center frequency, and the op amp another 180dgs totalling in 360dgs positive feedback, which is one of the requirements for oscillation. Anyway, don't keep the op amp with open loop gain, introduce some negative feedback. Also, lowest THD+N can be attained with a lower Tau AGC than the oscillation frequency. With diode AGC, you work with each cycle to control the gain, which introduces noise and distortion into each cycle. To get better THD rating, you need to work with a few cycles at a time. For instance, you've heard of a simple bulb AGC. When the output voltage increases, so does its resistance, and when places on the proper spot it works as an AGC. The secret as to how it gives way better distortion figures, than almost any other methods is because the filament inside the bulb heats up and down quite slowly compared to the frequency of oscillation. You may have 1kHz oscillations, but the bulb time constant would only be a few Hz.

Now buggzy turns the world upside down. Lets see what Jason thinks about this statements buggzy made. I dont want to start all over as we ( i and buggzy) had so many times in the past years.

Phase shifts both varies a amplitude varies a little. EC does not register frequency change if it is happening.suppose the change in phase is to correct change in amplitude.

EC does register the exact frequency and phase, but it does not show us the result in numbers. 3 digits only for any numerical value, which i think is in 99.9% of circuits sexy. Sometimes i wish to overrule this temporary e.g. 6 or 9 digits. Anyway, its not Xmas.

It oscillates at the most attenuated spot because that is the only place where the 180 degree phase shift required for oscillation is present. Too high frequency or too low frequency and it’s phase shifted too much or not enough and won’t oscillate at those frequencies. It also oscillates at 163Hz instead of where there is 180 degree phase shift because of the AGC messing with it. When I got rid of the AGC, set the op-amp to have no voltage limits, and then set the gain of the op-amp to a value that made the oscillations neither increase nor decrease at any significant rate (55V/V) it then oscillated at 157Hz (180 degree phase shift is between 157 and 158Hz).

As as far as I can tell, everything @thebugger said is correct, except for the part where he says that it oscillates at the most attenuated frequency because the negative terminal is being used and that inverts everything. It really oscillates at the most attenuated frequency because that is where there is the 180 degree phase shift. The negative terminal is then used for the second 180 degree phase shift for a complete 360 degree phase shift, but only at that one specific frequency. Same thing for the band pass filter, but where it let’s the frequency through the most is where it phase shifts at 0 degrees and then the positive terminal is used because it also gives 0 degrees phase shift which is equivalent to 360 degree phase shift. The oscillations just happen to be at the most let through/attenuated frequency because that is where the 0/180 degree phase shift is in the band pass/notch filter.

I think there is a lot chaos in all the used words like feedback positive negative open loop closed loop negative gain or positive exact phase shift infinite attenuation.... etc. My main statement why i spend so much time is, feedback to the positive input terminal, which @thebugger i guess calls positive feedback, while two sentences later he calls also 360° at total loop phase shift “positive feedback“, OpAmp resistors from output to an + input terminal are not needed for such kind of oscillators, they are even dangerous! Everything might get out of control. @thebugger came with his circuit and a perfect 159.1549.... Hz Twin Notch oscillator which can not work with this extra “positive“ feedback network which is what i think, is just wrong. Even you can find it on many webpages. Just wrong. Thats the main thing why i made so many explanations for such oscillators. All other statements... maybe another day.

"i" think @ thebugger is talking about rule of thumb,.

Unbookmarking

I think I likely put it between the two ends of the twin-T because that seemed like the most sensible solution. After all, it oscillates at that frequency because that’s where the twin-T cancels the waveform completely and because of the negative feedback the OP-amp tries to attain a state where the waveform is canceled completely. When adding a resistor there (one that’s only “triggered” at higher voltages thanks to the diodes) it diminishes the ability of the twin-T to cancel out the waveform and the negative feedback causes the OP-amp to eliminate that waveform.

Oh; that's very cool. I didn't think much about that. Thanks again @jason9 (I wonder how many thanks I've put. Anyway.)