Sine Oscillator

Discussion in 'General Electronics Chat' started by akis02, Sep 25, 2013.

  1. akis02

    Thread Starter Member

    Apr 30, 2011
    I would like to experiment with a simple, and adjustable frequency sinewave generator.

    I have built successfully a Wien bridge using parallel FETs in the feedback loop for good temperature stabilisation. However its frequency is determined through 2 resistors and 2 capacitors.

    I have searched and have not found a circuit for sinewave generation at a variable frequency, not analog anyway.

    There is the XR2209 but it does not have a sine output.

    I read a post hereabouts that a passive filter will convert a triangle or square wave into a sine wave - is that true and will the sinewave be of low distortion?

    Otherwise can you please point me to any circuit I could try?

  2. t06afre

    AAC Fanatic!

    May 11, 2009
  3. studiot

    AAC Fanatic!

    Nov 9, 2007
    It is difficult to design a good quality variable analog sine oscillator controlled by only one adjustable component.

    Such a circuit was published in Wireless World, Aug1974 p272

    "Single pot controlled Wein Bridge"

    Elektor July 1987, no 70

    However the range is restricted.

    A good place to find alternatives to traditional Wein is the Texas Instruments app note

    Another approach might be to look into voltage controlled oscillators.

    Last edited: Sep 25, 2013
  4. akis02

    Thread Starter Member

    Apr 30, 2011
    The oscillators I know of ; Bubba, Colpits, Wien, RC 3/4 stages - all are not stable enough, frequency wise and amplitude wise, and the circuit I have attached here is the best practical I have ended up with after long experimentation.

    This is a fixed frequency Wien bridge oscillator which I have built and been using for some time now. It has good amplitude and temperature stability, I have blasted it with a 3KW heater fan, until the components were at 70C, and there was very little change in the output. Final amplitude adjustment can be performed by the second op-amp.

    The frequency is determined by the RC (R1, C1, C2, R4, C3, C4) - I am not sure how to make this variable, even for a short range, short of using a dual-pot in place of R1/R4 ? But that would not be accurate and there's noise in the pot maybe ?

    Then I was thinking to use two LDRs in place of R1 and R4 and an LED to control them! How stable and how repeatable would that be? I have a bunch of LDRs here and they have quite a wide range (from a few KOhm to hundreds of KOhms), not sure how matched and stable that arrangement would be - but it seems a fantastic way to control the frequency.

    For better results I was thinking maybe some sort of PLL and/or a crystal to provide more stability, accuracy and repeatability? But I have not come across a practical circuit for a PLL oscillator, hence my questions here :)

    Edit: I have not found an analog VCO to produce sine waves yet.
    Last edited: Sep 25, 2013
  5. Shagas

    Active Member

    May 13, 2013
    I've made a roughly 20-6500 hz wien bridge sin wave oscillator using two fet optocouplers connected to one single potentiometer (using 10 nf caps and 500k pot if i remember correctly). You could ommit the optocouplers and use a dual pot to vary the resistors.
    If you want to extend the range then you will have to either use an analog multiplexer which will switch channels(once you reach the limit) on which diffrent value caps will sit or you can try making it with capacitor modelling circuits made out of an op amp , a fixed cap and a variable resistor which will give you a total requirement of 4 resistors to controll simultaniously . Since we don't have quad potentiometers you can use 4 fet optocouplers all connected to one single potentiometer
  6. studiot

    AAC Fanatic!

    Nov 9, 2007
    There are several separate issues when designing a high quality oscillator.

    1) Frequency accuracy is obviously controlled by the frequency determining components. For almost all types of design this requires the tracking variation of two or more components. For a Wein bridge this can be a pair of resistors or a pair of capacitors. Manufacturers buy or make special purpose tracking pots that are extremely difficult to obtain outside for this purpose. And as you note even this are less than ideal.
    Alternatively you can go for high impedance and use high quality tracking variable capacitors. This used to be a common method in the days of valve (hi Z) circuitry and has been resurected since with FETs or FET input op amps. Variable capacitor variation are much nearer perfect.
    A further alternative is to use switches and spot frequencies. J Linsley Hood published some simple very high spec oscillators on these lines for measurement of ultra low distortion. He also used a Twin T feedback network, rather than a Wein for this as it leads to lower distortion and better oscillation stability.

    Now it is up to you to tell us more detail of your requirements for sensible help at this level of design.
  7. daviddeakin

    Active Member

    Aug 6, 2009
    Here is an example of a simple quad-opamp sine generator that uses a single pot for tuning. Range is about 20Hz to 100kHz and distortion less than 1% over most of the range. I don't know about temperature stability though.
    Shagas likes this.
  8. akis02

    Thread Starter Member

    Apr 30, 2011
    Based on all the above suggestions, I breadboarded the Wien Bridge which I showed above but replaced the R1/R4 with two LDRs. I filed a green LED on both sides and super glued the two LDRs onto it, then black taped it as best I can. I used a normal pot across the rails and an op-amp to drive the LED. The LDRs range is from many MOhm to 500Ohm. With 1nF caps I got from a few KHz to about 250KHz, with 100pF I got all the way up to 1.5MHz - however the breadboard is dreadful and the curves, even without the LDRs, do not look great. Next step to do it again on a better quality LDR. However I might also try a dual pot, because the LDRs are not matched and do they receive same amounts of light resulting in differences.
  9. BobTPH

    Senior Member

    Jun 5, 2013
    Last edited: Sep 25, 2013
  10. Brownout

    Well-Known Member

    Jan 10, 2012
    Hi T,

    A VCO typically uses the control voltage to set the oscillator frequency. In the circuit you linked, the FET is used to automatically set the amplifier gain so that the amp isn't overdriven. That prevents distortion.