How to remove this problem (Bubba oscillator - Proteus error)

[Iiest2021]

Hi,

I want to generate a sine wave by using Bubba oscillator. But when I do the simulation then showing this error in Proteus.

Please help me !!

pls help me .

 

[FvM]

Looks like you are reproducing the problem you produced in PSpice before. Using a LM348 component without simulation model.

 

[KlausST]

Hi,

Additionally
* no feedback of U2
* positive feedback of U4, U5
* wrong power supply polarity of U3, U1A, B2
* U1A at all
* some part values
* no power supply decoupling capacitors
make no sense.

Klaus

 

[danadakk]

Single supply Wein Bridge oscillator -

Online Circuit Simulator with SPICE

PartSim is a free and easy to use circuit simulator that includes a full SPICE simulation engine, web-based schematic capture tool, a graphical waveform viewer that runs in your web browser.

www.partsim.com

Regards, Dana.

 

[Iiest2021]

Looks like you are reproducing the problem you produced in PSpice before. Using a LM348 component without simulation model.

please, can you say what do you mean" using LM348 component without simulation model."

 

[Iiest2021]

Hi,

Additionally
* no feedback of U2
* positive feedback of U4, U5
* wrong power supply polarity of U3, U1A, B2
* U1A at all
* some part values
* no power supply decoupling capacitors
make no sense.

Klaus

beasically I am design of bubba oscillator for sine wave generation .
i am design same circuit in protius .
I am not unterstanding what is the problem ?
and file is attached

 

[FvM]

please, can you say what do you mean" using LM348 component without simulation model."

Just read the warning thoroughly and follow the suggestion: Use a symbol from the PSpice simulation library. Orcad schematic entry is used for simulation and PCB design. Most libraries are for PCB design only, they have associated footprints but no SPICE model. The other option is to download the missing model from the internet and attach it to the symbol.

If there's no LM348 simulation component shipped with PSpice, you might use uA741 instead, it's essentially the same OP as single type.

 

[danadakk]

Pspice model here for LM348 :

https://www.ti.com/lit/zip/sloj044

Wein bridge solutions with AGC. https://en.wikipedia.org/wiki/Wien_bridge_oscillator


Regards, Dana.

 

[Audioguru]

I think the circuit has a problem with its bias voltage:

 

[FvM]

The bubba oscillator circuit as shown in post #6 has overall negative DC feedback, stable bias and can work well. It doesn't need improvement like the suggested AC coupling.

 

[KlausST]

Hi,

where is the feedback at the right bottom OPAMP?

Klaus

 

[FvM]

Right, one feedback connection is missing. Apparently I was supplementing it in my mind.

The topology is correct nevertheless.

--- Updated Oct 31, 2022 ---

In contrast to the various errors in post #1 that you listed in post #3. It turns out that the same errors are also in the PSpice version of the same circuit, caused by a rather sloppy way of drawing the simulation schematic. https://www.edaboard.com/threads/help-with-pspice-error.404596/
Correcting these errors is necessary to make the circuit work but doesn't remove the problem reported in post #1 which is solely caused by using OP symbols without attached simulation model.

 

[Audioguru]

Here is another wrong circuit from Texas that I got my wrong idea from:

 

[Easy peasy]

What about the lack of connection to the input at lower right ... ?

surely this is an error - should be a unity gain buffer ?

 

[FvM]

surely this is an error - should be a unity gain buffer ?

Yes, that's what KlausST adressed in post #11.

Here is another wrong circuit from Texas that I got my wrong idea from:

The wrong bias voltage can be still found at TI website. https://www.ti.com/lit/pdf/slyt164 See Fig.11.

--- Updated Nov 1, 2022 ---

A similar phase shift oscillator Fig.7 has a correct bias of Vcc/2 = 2.5 V.

 

[KlausST]

Ot:
Analyzing where the error in schematic of post#13 comes from.

The according OPAMP has a DC gain of 1+Rf/Rg = 1 + 1500k/360k = 5.17.
So maybe the first idea was that it amplifies the 0.5V x 5.17 = 2.58V
But this idea was wrong, because it ignores that the loop across the other OPAMPs causes a DC feedback of 1.
Indeed the overall DC gain becomes close to 1, thus making the bias voltage close to 0.5V (instead of 2.5V)

Klaus

 

[Easy peasy]

This does appear to be correct ....

 

[Audioguru]

Easy peasy, the 0.5V is wrong and should be 2.5V (half the supply voltage) for the output to have filtered but symmetrical clipping.
The output of the 4th RC filter would have the lowest distortion but they selected the output from the 2nd filter probably for its higher output level.
I wonder why they selected an output from the fairly high impedance RC filter instead the very low impedance from the output of the opamp filter buffer?

Notice that the filter resistors haver higher resistance than before (then the filter capacitors have equally less capacitance than before) so that the 360k gain-setting resistor does not load-down the 4th filter as much.

Does anybody have this circuit simulated so that we can see how much clipping distortion exists at each filter stage?

 

[FvM]

I wonder why they selected an output from the fairly high impedance RC filter instead the very low impedance from the output of the opamp filter buffer?

They didn't think.

Distortion of Bubba oscillator has been previously discussed:

https://www.edaboard.com/threads/simulation-of-bubba-oscillator.311292

 

[Audioguru]

I failed to turn the page of the application note where distortion was shown. Cosine was 0.1% when the Cmos quad opamp was used. The antique LM348 quad opamps have an input resistance too low for the 1.5M or 2.2M feedback resistor.