Wein Bridge Oscillator

[bananasiong]

Hi,
I'm currently learning Wein Bridge oscillator, start from the one without amplitude control. From simulation, I can get the expected result. But when I did experiment, it didn't work. The CRO shows nothing, just a flat line. Even I've increased the feedback resistor to be more than twice of R1.
I think I know what's the problem, it is GROUND. From the attachment (red circled), if the GND is not connected, the output is flat line. But how should I connect a GND in practical? There is no GND with the power supply.

Thanks

 

[Nigel Goodwin]

You're confusing yourself with 'ground', 'chassis' would be a better word, as it doesn't need to be connected to 'ground'. The chassis connection is just a reference point that connects to the chassis (if there is one), and is grounded (if applicable) - for a real scope (just as in the simulation) the chassis of the scope should be connected to the chassis of the oscillator.

One obvious omission in your circuit is decoupling capacitors, place them from +ve to chassis, and -ve to chassis, as close to the chip as possible.

 

[audioguru]

Look in Google to learn that the gain of a Wien Bridge oscillator must be more than 3. So in your circuit the feedback resistor must be more than 20k.

Two power supplies are shown in your circuit with the circuit's ground connected beween them where they join. There is a positive 15V power supply and a negative 15V power supply. The circuit should work if two 9V batteries are used.
The circuit can be redesigned to use a single power supply.

When you tried the circuit, what was its load? An opamp can't drive a load that is less than about 2k ohms.

 

[bananasiong]

You're confusing yourself with 'ground', 'chassis' would be a better word, as it doesn't need to be connected to 'ground'. The chassis connection is just a reference point that connects to the chassis (if there is one), and is grounded (if applicable) - for a real scope (just as in the simulation) the chassis of the scope should be connected to the chassis of the oscillator.

Chassis? You mean the chassis is between the + and - of both the supplies? But the CRO and the oscillator are common grounded (not shown in attachment, but yes they are).
The end of the frequency selective network and the end of R1 is also grounded, should they be connected between the - of the top and + of the bottom supply?

One obvious omission in your circuit is decoupling capacitors, place them from +ve to chassis, and -ve to chassis, as close to the chip as possible.

What is decoupling capacitor?

The circuit can be redesigned to use a single power supply.

I thought only norton op-amp is using single supply?

When you tried the circuit, what was its load? An opamp can't drive a load that is less than about 2k ohms.

I don't have any load, I just connect the output to the CRO.

 

[audioguru]

The end of the frequency selective network and the end of R1 is also grounded, should they be connected between the - of the top and + of the bottom supply?

Of course the two power supplies should be connected together and to ground exactly as shown on the schematic.

What is decoupling capacitor?

A big filter capacitor across each power supply. They are very important if batteries are used because batteries have an internal resistance.

I thought only norton op-amp is using single supply?

Any opamp circuit can have a single supply. Simply disconnect the non-inverting input resistor from ground and connect it to half the supply voltage, then add a capacitor in series with the feedback resistor that went to ground.

 

[bananasiong]

Of course the two power supplies should be connected together and to ground exactly as shown on the schematic.

That's my problem, usually for only single supply, those need to be grounded are connected to -ve terminal. But for bipolar power supply, those need to be grounded are connected between 2 power supplies right? That's what I've done but still can't get the result.

A big filter capacitor across each power supply. They are very important if batteries are used because batteries have an internal resistance.

How if I'm not using battery? Are they still needed?

Thanks.

 

[on1aag]

Wien bridge oscillator.

Hi Bananasiong,

The conditions for oscillation, that the gain should be equal to three and
that the phase shift is zero, are only true if both 10k resistors and both 10nF capacitors are exactly equal ! Try increasing the loop gain, increase the 21k resistor to 22k or 24k. If you would have used smaller capacitors, the stray
capacitance of the circuit should also have been compensated as well.

on1aag.

 

[Hero999]

That scope isn't grounded, connect both of the dots on the right hand side to ground.

 

[Nigel Goodwin]

That's my problem, usually for only single supply, those need to be grounded are connected to -ve terminal. But for bipolar power supply, those need to be grounded are connected between 2 power supplies right? That's what I've done but still can't get the result.


How if I'm not using battery? Are they still needed?

Yes, ALWAYS NEEDED!.

 

[Hero999]

The power supply will have some internal impedance, whether it be the lead reisitance and inductance or the transient response of a linear regulator like the LM317. To compensate for this a small filter capacitor is always used, even if it's just a 100nf ceramic disc.

 

[bananasiong]

That scope isn't grounded, connect both of the dots on the right hand side to ground.

Where is the right hand side to ground?

In simulation, if the GND is removed, it doesn't work. In practical, I don't know which GND should I connect to. Power supply has only +ve and -ve terminal but has no GND.

Thanks.

 

[audioguru]

Don't you have two power supplies like on the schematic, a positive power supply and a negative power supply?
Don't you have the two power supplies connected in series and where they join together you connected to the circuit's ground, like on the schematic?

 

[bananasiong]

Yes, I do have 2 power supplies as in the attachment. But I don't know where should I connect their joint to. Or their joint is known as GND?
When doing other experiment with only single supply, the -ve terminal is always known as GND, but for bipolar, I don't know where should those should be connected to GND connected to.

Thanks

 

[audioguru]

Yes, I do have 2 power supplies as in the attachment. But I don't know where should I connect their joint to. Or their joint is known as GND?

Of course where the two batteries are joined is connected to the circuit's ground. Look at the schematic. The 'scope is also connected to the circuit's ground.

When doing other experiment with only single supply, the -ve terminal is always known as GND, but for bipolar, I don't know where should those should be connected to GND connected to.

The schematic shows what is connected to the circuit's ground as explained above. If you connect the -ve of the negative supply also to ground then the negative supply would be shorted.

 

[bananasiong]

Hi,
I've done Wein Bridge Oscillator, it works!!! I think the analog lab unit I used previously is not working well, that's why I kept finding where the hack is the GND.

From the theory, positive feedback makes the op-amp unstable and hence vibrate and start oscillating. But I can see that there are both -ve and +ve feedback in the oscillator circuit. I've tried swaping the inverting and non-inverting input with the same circuit using simulation, it oscillates for a very short period only, it goes to -ve before the amplitude is constant.

What do the decoupling capacitors do? What is compensate??

Thanks.

 

[audioguru]

Hi,
I've done Wein Bridge Oscillator, it works!!!

Good.

I can see that there are both -ve and +ve feedback in the oscillator circuit. I've tried swaping the inverting and non-inverting input with the same circuit using simulation, it oscillates for a very short period only, it goes to -ve before the amplitude is constant.

The negative feedback is AC and DC. The DC negative feedback causes the output DC voltage to be near 0V. The positive feedback through the Wien Bridge parts is AC only and at only one frequency.
If you reverse the inputs then the opamp has DC positive feedback then its output will saturate to near the positive supply or the negative supply and stay there.

What do the decoupling capacitors do?

They keep the supply voltage from swinging with the signal.

What is compensate??

Opamp frequency compensation? Wien Bridge Oscillator output level compensation? What do you mean?

Thanks.[/QUOTE]

 

[bananasiong]

Opamp frequency compensation? Wien Bridge Oscillator output level compensation? What do you mean?

I read the replies and saw that decoupling capacitors are used to compensate the circuit capacitance, I don't get it actually.

 

[audioguru]

Many ICs oscillate at a high frequency if the supply doesn't have a decoupling capacitor. Maybe you could call the capacitor a compensation capacitor. I call it a supply bypass capacitor.

 

[Hero999]

A compensation capacitor causes an artificial roll-off which causes less than unity gain at 180 degrees, preventing an amplifier from going into oscillation. This is totally different to a decoupling capacitor.

 

[audioguru]

Remember the old 709 opamp that needed an external compensation capacitor?