Converting AF oscillator to use discrete components

[neazoi]

Hi I would like to convert the last circuit on this page https://www.learnabout-electronics.org/Oscillators/osc34.php to use only discrete components and remove the opamp somehow.
Is that possible and how?
I have made an ALC amplifier for RF **broken link removed** which I think it can be used for audio as well, but this will be a total of 4 transistors including the oscillator one. I need to do it as simple as possible.
Any ideas?

 

[FvM]

Wien bridge need a non-inverting amplifier with gain around +3, variation range can be relative small. For clear sine wave output, a linear control element like JFET is preferable, I would even refer to an additional linearization circuit that superimposes half of the drain-source AC voltage to gate-source voltage, can be found in many Wien bridge designs.

A two or three transistor complementary amplifier should serve the purpose.

 

[BradtheRad]

Wien-bridge oscillator made from 2 transistors. A schematic is at website:

https://www.circuitstoday.com/wien-bridge-oscillator



Component values aren't labelled so I tried playing with values until it started oscillating by itself. The potentiometers need to be adjusted within a narrow range, so that transistors receive just the right amount of bias.

 

[Easy peasy]

you could replace the op-amp with the bjt equiv - long tailed pair etc ... about 7 xtors

 

[betwixt]

I'm not sure Brad's schematic is a true Wien-Bridge design as it uses different RC values in the series and shunt arms of the feedback loop and it also has additional phase shifting between the amplifier stages. This probably accounts for the critical bias problem as the bias settings also change the RC time constants. It would be very difficult to tune when the R and C feedback components do not track in value.

It would be better to concentrate on a wideband amplifier with minimal phase shift then apply the frequency selection only in the feedback loop.

Brian.

 

[FvM]

The original circuitstoday design is a Wien bridge, the modified one strictly isn't, just some kind of RC oscillator, apparently working.

To achieve good sine quality, you need ALC. Or an oscillator topology with effective harmonic suppression like Bubba.

If you want variable frequency with tandem potentiometer as in the post #1 design, true Wien bridge is preferable.

 

[FvM]

To show the feasibility of discrete implementation, see the below circuit and simulation files.



It can be probably optimized in different ways, I believe that the original OP implementation is more straightforward and essentially simpler i any case.