What is the frequency of this phase shift oscillator?

[Jony130]

At first, the circuit under discussion is NOT intended to use as an oscillator - because sometimes it oscillates and sometimes not.
Both of your loop gain simulations are a good example - and you see that the condition of oscillation depends on the used RC time constant, in conjunction with the open-loop characteristic of the opamp.
Thus, no surprise that it "fails" sometime. I suppose for lowering the RC time constant the circuit will (perhaps) oscillate.

I simply want to check if this circuit is so prone to oscillation as it looks in simulation program. But in real live I wasn't able to to force this circuit to oscillation. All I get was damp oscillations.
See o-scope screen shot

 

[Winterstone]

I simply want to check if this circuit is so prone to oscillation as it looks in simulation program. But in real live I wasn't able to to force this circuit to oscillation. All I get was damp oscillations.
See o-scope screen shot

Yes, I got this information from your previous post already.
What about the RC time constant? Did you lower it?

Besides this, you have to be careful to transfer simulation results - in particular loop gain simulation using the L-C method - to closed-loop conditions.
In many cases, the L-C method works sufficiently - however, this simple method has one drawback: In closed-loop operation the opamp input resistance is in parallel to the capacitance and, thus, changes the phase function. However, in your loop gain simulation scheme hte capacitor is NOT loaded anymore.
This "loading error" (deviation from real conditions) may be responsible for the effect you have observed.
Note, in particular because of this error Middlebrook has developed another (more complicated) method for CORRECT loop gain simulation.

 

[Roff]

Yes, I got this information from your previous post already.
What about the RC time constant? Did you lower it?

Besides this, you have to be careful to transfer simulation results - in particular loop gain simulation using the L-C method - to closed-loop conditions.
In many cases, the L-C method works sufficiently - however, this simple method has one drawback: In closed-loop operation the opamp input resistance is in parallel to the capacitance and, thus, changes the phase function. However, in your loop gain simulation scheme hte capacitor is NOT loaded anymore.
This "loading error" (deviation from real conditions) may be responsible for the effect you have observed.
Note, in particular because of this error Middlebrook has developed another (more complicated) method for CORRECT loop gain simulation.

I think his latest post was referring to the circuit with the cap on the op amp output. His discrepancy between hardware and simulation point up the limitations of models.

 

[Winterstone]

I think his latest post was referring to the circuit with the cap on the op amp output. His discrepancy between hardware and simulation point up the limitations of models.

Hi Ron,

may be you are right, however, if I read post#39 my feeling is he speaks about the opamp with the RC-lowpass in the negative feedback path.
Anyway, perhaps Joni gives us some more information about his problem.
W.

 

[Roff]

Hi Ron,

may be you are right, however, if I read post#39 my feeling is he speaks about the opamp with the RC-lowpass in the negative feedback path.
Anyway, perhaps Joni gives us some more information about his problem.
W.

I think the first sentence of post #39 was referring to the original circuit, with the RC-lowpass in the negative feedback path.

 

[Jony130]

Initially I build this circuit to see if this circuit start osculation so easily as in the simulation.

**broken link removed**
I try with R1 = 220K; 100K; 33K ; 10K and C1 = 1nF ; 10nF; 100nF ; 1μF But with no luck. O-scope show no oscillation, sometime O-scope show damp-osculation at start-up.


Next I reconnect the capacitor directly to the op amp output.
**broken link removed**
And this circuit oscillate without any problem.
As we can see on a scope
[/QUOTE]

 

[Winterstone]

Hi Joni,
what shall/can I say?
As I told you already - the circuit with RC feedback is NOT intended to be used as an oscillator. Everybody who wants to use this circuit as a differentiating unit (input via C), which is the normal application, would be happy if it wouldn`t oscillate. However, as one cannot be sure if it is stable or not a suitable resistor is placed in series with C (as mentioned already earlier in post#35). So it is no surprise that it does not oscillate - it really depends on (a) the type of opamp and (b) the particular part (because of tolerances) and (c) on possible parasitic influences within the hardware.

Regarding the 2nd circuit I gave you an explanation. The observed waveform indicates that the slew rate (only 0.6V/µsec for the LM358) has major influence on frequency and amplitude.
Further questions?

 

[MrAl]

Hi,

I agree that these circuits dont look very good at all. It doesnt matter if they oscillate or not, what matters for an oscillator is that yes they must oscillate, BUT they must ALSO be designed properly which allows them to achieve other goals that make them a reliable circuit.

For example, the second circuit lacks some sort of forced hysteresis. With the non inverting terminal tied to ground the hysteresis depends on some hard to analyze and too variable of a parameter to make the design a viable alternative to what is already out there that is known to work well.

So my advice would be to study some of the other oscillators that are already out there to find out what their major design goals where in creating the circuit. That will tell you a mountain of information that will help you either design your own or modify another type or just use another type the way it is.

 

[Winterstone]

Hi MrAl,

I agree to your statements/recommendations.
On the other hand, I see that there is a good reason to study the behaviour of such circuits under real conditiones. Thus, somebody who is not so familiar with real-life opamp-based circuits can learn that the frequency-dependent gain (very often NOT contained in formulas, frequently even not mentioned!) and the resulting phase shift can have a severe influence on the desired function of the circuit.
* Example with RC negative feedback: This circuit resembles the classical differentiating circuit (input at the capacitor) in case of IDEAL opamps, but cannot be used without additional stabilization means in case of real amplifier units.
* Example with unity gain amplifier (capacitively loaded): Contained in each textbook and works as desired for ideal opamps, but may tend to oscillations for real opamps.

 

[MrAl]

Hi Winterstone,

Oh yes some good points. Also the slew rate which i hardly ever see mentioned which as a big effect on these circuits. It always surprised me that slew rate is almost never mentioned. I think what most authors assume is that the person designing with an op amp should have studied op amps before starting a design, and that course should have included the various subtleties of the device.
So i guess we can conclude that these circuits here in this thread are for study purposes only.

 

[Winterstone]

Also the slew rate which i hardly ever see mentioned which as a big effect on these circuits. It always surprised me that slew rate is almost never mentioned.
.

Oh yes, I totally agree. For example, the last picture shown in post#46 can be explained only as an unwanted result of the finite slew rate (around 0.6V/µs for LM358).
Otherwise (for not slew limited opamps) the frequency should be around the transit frequency (about 1 MHz). But it is only 25kHz (with a rather small amplitude).