Maker Pro
Maker Pro

How can I set a frequency in a Wien bridge oscillator

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
So why don't you let a sweet young thing like Jeri tell you how to do it without using any lame parts you might have.

When researching lamp stabilised Wien bridge oscillators some time ago I came across a great article from HP describing issues they had in designing the original oscillators they made.

They found that some oscillators were less stable than they should have been. Perversely, these were the ones with the best distortion figures. What they discovered was that it was the small non-linearities in the amplifying element that contribute to stability. I found this after making a rock-steady lamp stabilised sine wave oscillator using a cheap rubbishy op-amp, only to have the circuit become unstable and largely unusable when an ultra linear, low distortion, gold plated, hand assembled by nubile virgins, purchased at the cost of your first born child style of op-amp was transplanted into the circuit.

I have a hard copy of that article at home, I can pull it out as an interesting reference as to why you might actually prefer to use "lame" parts in your Wien bridge oscillator.

And curiously, after watching that video (I'm pretty sure I've watched it previously) we're actually talking about similar things.
 

Ratch

Mar 10, 2013
1,099
Joined
Mar 10, 2013
Messages
1,099
When researching lamp stabilised Wien bridge oscillators some time ago I came across a great article from HP describing issues they had in designing the original oscillators they made.

They found that some oscillators were less stable than they should have been. Perversely, these were the ones with the best distortion figures. What they discovered was that it was the small non-linearities in the amplifying element that contribute to stability. I found this after making a rock-steady lamp stabilised sine wave oscillator using a cheap rubbishy op-amp, only to have the circuit become unstable and largely unusable when an ultra linear, low distortion, gold plated, hand assembled by nubile virgins, purchased at the cost of your first born child style of op-amp was transplanted into the circuit.

I have a hard copy of that article at home, I can pull it out as an interesting reference as to why you might actually prefer to use "lame" parts in your Wien bridge oscillator.

And curiously, after watching that video (I'm pretty sure I've watched it previously) we're actually talking about similar things.

So the inferior op amp had the right kind of nonlinearities that contributed to better stability in the circuit, and the circuit needed more or different nonlinearity than the lamp was able to provide when the superior op amp was used. Just speculating. Anyway, the article sounds interesting.

Ratch
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Whilst I'm pretty sure this 1960 article from HP is going to go way beyond what tburn wants or needs, some of us might find it of interest.

The most important conclusion is in the second paragraph:

"It is shown that amplifier non-linearity so slight as to cause insignificant waveform distortion has a profound effect on envelope stability. In fact, if it were not for this slight non-linearity, it would be virtually impossible to build a simple lamp-stabilized RC oscillator with good stability over a wide frequency range."

And it goes on to relate the correlation with low distortion and envelope instability seen in practice.
 

Ratch

Mar 10, 2013
1,099
Joined
Mar 10, 2013
Messages
1,099
Whilst I'm pretty sure this 1960 article from HP is going to go way beyond what tburn wants or needs, some of us might find it of interest.

The most important conclusion is in the second paragraph:

"It is shown that amplifier non-linearity so slight as to cause insignificant waveform distortion has a profound effect on envelope stability. In fact, if it were not for this slight non-linearity, it would be virtually impossible to build a simple lamp-stabilized RC oscillator with good stability over a wide frequency range."

And it goes on to relate the correlation with low distortion and envelope instability seen in practice.

That's a lot of heavy math and feedback control. Only a lamp was analyzed. What about other forms of AGC, such as back-to-back diodes, FET's, bridges, etc? Looks like making a Wien with very low distortion is not easy.

Ratch
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
It is very interesting. My own tests indicate that the "modulating" frequency (the envelope instability) is related to the time constant of the lamp.

If you use any other control device, there may also a time constant (such as a capacitor maintaining a voltage on gate of a jfet). I suspect that similar behaviour might be exhibited using such control elements, but I have not investigated it.

As for other control elements (like back-to-back diodes) I can't comment other than to suggest that these don't produce a high purity sine wave.

Incidentally the first op-amp I used was a reasonably well performing one and I did note that the Wien bridge took maybe a few seconds to stabilise after changing frequency, powering up, etc. I assumed that a better op-amp would improve this (and that lead me down a very interesting rabbit hole. (And I didn't expect to find WIlliam Hewlett and Charles De Gaulle along the way)
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
ahhh yes
I learnt a long time ago that a thermistor or light bulb was needed in the R2 position
Dave

It depends on the sign of the temperature coefficients.
Your goal must be to LOWER the gain (which at t=0 must be larger than "3") for rising amplitudes leading to rising temperatures..
That means either R2 must be INCREASED (light bulb) or R1 must be DECREASED (NTC thermistor).
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Getting back to the original question of setting the frequency, I found this page interesting when I was building up my oscillators.

This page covers a heap of practical issues and explains how the oscillator works. It also suggests a method of frequency control (over a limited range -- a decade) by varying a single resistor.

Of note is that the op-amp is given a beefier output stage to allow it to drive the low resistance feedback path (the part where the lightbulb is).
 
  • Like
Reactions: LvW

Ratch

Mar 10, 2013
1,099
Joined
Mar 10, 2013
Messages
1,099
Getting back to the original question of setting the frequency, I found this page interesting when I was building up my oscillators.

This page covers a heap of practical issues and explains how the oscillator works. It also suggests a method of frequency control (over a limited range -- a decade) by varying a single resistor.

Of note is that the op-amp is given a beefier output stage to allow it to drive the low resistance feedback path (the part where the lightbulb is).

And here is another circuit that uses two op amps and one tuning resistor. A better AGC could be substituted for the two back-to-back diodes.
Wien.jpg

Ratch
 

LvW

Apr 12, 2014
604
Joined
Apr 12, 2014
Messages
604
Getting back to the original question of setting the frequency, I found this page interesting when I was building up my oscillators.
This page covers a heap of practical issues and explains how the oscillator works. It also suggests a method of frequency control (over a limited range -- a decade) by varying a single resistor.
Of note is that the op-amp is given a beefier output stage to allow it to drive the low resistance feedback path (the part where the lightbulb is).

I think, the method as proposed in the referenced paper is very useful - in particular for fine tuning of the oscillator. That means, when the required tuning range is not too large (e.g. less than 10...20%).
 
Top