Don't get an inverting sinus.

[unitt]

Hi all,
I have running a bubba sinus oscillator on 20Khz. The picture show the input and output singals. The quad opamp is a LME49740. I buffer the sinus wiht 2 voltage followers. The reason for this is that I will get a non-inverting and an inverting sinus.
And here goes it wrong and no idea how to get it.
I have much time spend wiht try-outs. but
every time no result.
Is here anyone who can explain what to do.
thank you.

 

[crutschow]

Follower means the output follows (is the same) as the input is so obviously you can't get an inverted signal with that. Use an inverting op amp configuration.

 

[Audioguru]

Your 'scope shows the output signal is inverted from the input signal.
The first opamp in a Bubba oscillator is an inverting opamp. There are four RC networks that each provide 45 degrees phase shift so their total is 180 degrees.

 

[kam1787]

BTW I think you mean SINE, and not sinus. Sinus is a cavity within a bone or other tissue. Most are commonly found in the bones of the face and connecting with the nasal cavities

 

[Analog Ground]

Sinus can also mean a curve or a bend (from Latin). So, maybe the waveform is a "sinus"? Unusual use of the word but interesting.

 

[LvW]

Hi all,
I have running a bubba sinus oscillator on 20Khz. The picture show the input and output singals. The quad opamp is a LME49740. I buffer the sinus wiht 2 voltage followers. .

1.) Input and output? An oscillator needs no input. It creates a signal by self-excitement

2.) You don`t need two buffers. The opamp´s output signals can be used without buffering.

 

[unitt]

Hi
Thank you for your reactions.
A few corrections,
It is indeed sine, sinewave.
An oscillator have no In-Out. I agree. wiht the picture
I will show that the phase shift is 180 degrees between the input and the output,of the opamp with amplification, the 4 times 45 degrees shift is correct
Vcc is 12Vdc, and the wave alter around Vcc/2 =
6Vdc.
I still have one confusion, The high value of the resistors for amplification.
Can someone explain?
thanks

 

[LvW]

Unitt, where is your circuit?
The high resistor values are neceesary to prevent heavy loading of the last RC section (which would result in an unacceptable frequency shift).

 

[unitt]

Hi LvW
Thank you for your reaction.
The rc is 1500R and C is 4.7nF. calculate 22Khz
Rf is 2.2M and R1 is 470 K
This bubba oscillator running wihtout a problem.
I have read a few sites wiht info about opamp's.
I guess the reason why it fails are the resistor value's.
I used resistors of 10 K for Rf.
I will a non-inverted and an inverted sine of the small sinewave
of the picture in post # 1
An amplification of 3 or 4 times is ok.
What must I do?

 

[Audioguru]

What must I do?

Use 2.2M and 470k for the feedback, then the input resistance is high which does not load down the RC and the gain is 2.2M/470k= 4.68 times.
The gain of the first opamp must be more than 4 times for the circuit to oscillate because each RC reduces the level 0.707 times.

 

[Audioguru]

I have used your part values and added a non-inverting output amplifier with a gain of 3 times.

 

[tonny_t]

make the loop gain >= 1 and phase condition that oscillation happened.

 

[LvW]

I have used your part values and added a non-inverting output amplifier with a gain of 3 times.

Why not using the additional opamp (as another unity gain amplifier) between the last RC section and the inverter?
This has the additional advantage of buffering the last section - and we can use lower R values for the inverter.

 

[unitt]

Hi LvW,
Your advise for add 2 opamps will give result. An inverting opamp is running.
the non-inverting come soon. and I show a few pictures.
Indeed high value's for Rf and R1

 

[LvW]

Hi LvW,
Your advise for add 2 opamps will give result. An inverting opamp is running.
the non-inverting come soon. and I show a few pictures.
Indeed high value's for Rf and R1

Did you understand what I mean?
The circuit in post#11 uses another opamp.
OK - but, for my opinion, it is better to use this opamp as a noninverting buffer (gain=+1) in front of the inverting opamp.
This avoids loading of the last RC section and allows lower resistor values for the inverter.

EDIT: Ohh - I forgot the most important thing: This opamp gives you the wanted out-of-phase output.

 

[unitt]

Hi LvW
The point is I can't image about the reason for the values of Rf and R1.
I add two amp''s, one for an inverter and one for a non-inverter.
The input wave is the sine after the fourth amp of the oscillator.
The inverter amp running wiht an amplification of 3 x.
Rf= 2.2M and R1=750K ( 2 1.5M parallel) .
See the picture. ( It's almost the same as the inverter in the oscillator.)
So I have one wave .
The non-inverter don't work. Rf= 2.2M and R1= 1.1M ( 2 2.2M parallal).
There is an instability, but I have no idea to solve it.
Will you explain what to do.
Thanks

 

[Audioguru]

The text on the screen of your 'scope vibrates the same as the waveforms. Then is the 'scope unstable or is picking up interference?

If you made the circuit on a solderless breadboard then it oscillates due to the capacitance between rows of contacts and the messy jumper wires acting like antennas.

Please post your schematic.

 

[LvW]

Hi LvW
The point is I can't image about the reason for the values of Rf and R1.

These parts load the last RC sections - and change the wanted oscillation frequency (which is 1/RC).
However, if the values are very large, this change (error) is small.
But you can reduce this error to zero using an additinal buffer. This effect is the only reason for all the other unity gain amplifiers inthe circuit.

I add two amp''s, one for an inverter and one for a non-inverter.
The input wave is the sine after the fourth amp of the oscillator.
The inverter amp running wiht an amplification of 3 x.
Rf= 2.2M and R1=750K ( 2 1.5M parallel) .
See the picture. ( It's almost the same as the inverter in the oscillator.)
So I have one wave .
The non-inverter don't work.

Why not? What does your screenshot show?
What means "not work"? Such a description does not say anaything. Perhaps you forgot to power the unit (joke)?

Rf= 2.2M and R1= 1.1M ( 2 2.2M parallal).
There is an instability, but I have no idea to solve it.
Will you explain what to do.

See audiogurus´s answer

 

[Audioguru]

I added a non-inverting opamp that has a high input impedance to the 4th RC filter. The high input impedance does not load down the RC and this new opamp has a voltage gain of 3 to increase the output level but not enough to cause clipping. I reduced the gain of the inverting opamp to 1.5 so that the total gain is 3 x 1.5= 4.5 which is a little higher than 4 so the oscillator starts reliably.

 

[LvW]

I added a non-inverting opamp that has a high input impedance to the 4th RC filter. The high input impedance does not load down the RC and this new opamp has a voltage gain of 3 to increase the output level but not enough to cause clipping. I reduced the gain of the inverting opamp to 1.5 so that the total gain is 3 x 1.5= 4.5 which is a little higher than 4 so the oscillator starts reliably.

Because each section at f=fo has a damping factor of 1/SQRT(2) a total gain slightly larger than G=4*/SQRT(2)=2.83 should start the oscillation. Didn´t it start at G=3?

EDIT: Sorry for this silly error. Of course, we have [1/SQRT(2)]^4=1/4 - hence G=4 as a theoretical gain requirement.