# Wien Bridge Oscillator

#### sparky 1

Joined Nov 3, 2018
648
RE Audio Guru
Using simulation distorsion analyzer of a first stage Op Amp 60.0 Hz a few adjustments and 100k resistor.
Noise source it should be possible to get 1 or 2% Thd. Additionally a jfet or thermistor or LDR could be used for AGC on a second stage
to minimize the tendency to build up to increase stability. I think with simulation it is better to build stage at a time.
On the breadboard building and testing both the high pass and the low pass. R3 and R4 can be trimmed with a multiturn 0-10 Ohm
It was popular to use the high impedance type liked the TL072 others liked the NE5532 low noise they typically got 2% Thd.

Last edited:

#### Ian0

Joined Aug 7, 2020
4,082
A Jfet used for AGC reduces the distortion much better.
True, but:
i) the distortion increases at low frequencies as the time Period approaches the time constant of the amplitude feedback loop
ii) the amplitude feedback loop is not always the easiest to stabilise
and
iii) the output amplitude is dependent on the characteristics of the JFET used - and isn’t any circuit that is dependent on particular transistor characteristics considered a ”bad circuit”?

The usual one-op-amp circuit needs a low impedance virtual earth to work on a single supply. The dual op-amp circuit doesn‘t, so it doesn’t actually need any more components.

Choice of circuit depends on the application. It is actually feasible to make a dual-op-amp Wien bridge using a pair of 74HCU04 inverters instead of the op-amps. It’s not great from the distortion point of view, but it might be as good as the XR2206 which is what usually gets recommended on this forum.

Last edited:

#### Wendy

Joined Mar 24, 2008
22,700
Finally! I got this sucker to work, after I figured out my wiring errors. I used a 10KΩ variable for R7 final resistance was 7.5KΩ I measured 7.5KΩ. As expected it was very finicky about the tweaking on the resistor, I will try to figure out an AGC circuit later. But a thought occurred to me set it forward just barely clipping and then take the waveform from pin 5.

Last edited:

#### Wendy

Joined Mar 24, 2008
22,700
No sooner did I think it then I did it:
...........................................................................................................................................................................pin 7

..................................................................................................................................................................................pin 5

Now to figure out an AGC circuit that will work.

#### Audioguru again

Joined Oct 21, 2019
4,103
Texas Instruments and Elliot Sound Products show a Wien Bridge oscillator with a Jfet doing its AGC. It has fairly low distortion of from 0.001% to 0.3% and one circuit avoids amplitude bounce when the frequency is changed.
Many years ago I used a Wien bridge oscillator with a lightbulb for AGC and its amplitude bounced a lot when the frequency was changed.

#### KeithWalker

Joined Jul 10, 2017
2,093
Texas Instruments and Elliot Sound Products show a Wien Bridge oscillator with a Jfet doing its AGC. It has fairly low distortion of from 0.001% to 0.3% and one circuit avoids amplitude bounce when the frequency is changed.
Many years ago I used a Wien bridge oscillator with a lightbulb for AGC and its amplitude bounced a lot when the frequency was changed.
I used a low voltage incandescent bulb from an old string of Christmas tree lights. It worked very well but it did bounce around on frequency changes. I still prefer using back to back diodes.

#### AnalogKid

Joined Aug 1, 2013
9,621
iii) the output amplitude is dependent on the characteristics of the JFET used - and isn’t any circuit that is dependent on particular transistor characteristics considered a ”bad circuit”?
It's been a while since I did a tolerance audit, but I think a transconductance amp has better parametric repeatability than a discrete JFET. Both the LM13700 and the CA3080 have all of the features needed for an AGC stage.

ak

#### Wendy

Joined Mar 24, 2008
22,700
I'm thinking more along the lines of a MOSFET, with that mind I came up with this:

when I measured R5 it was 4.663KΩ. It is after 11:00PM time for me to go to bed.

#### Bordodynov

Joined May 20, 2015
2,948

#### Attachments

• 2.6 KB Views: 1

#### Bordodynov

Joined May 20, 2015
2,948
60Hz

#### Attachments

• 1.7 KB Views: 1

#### Wendy

Joined Mar 24, 2008
22,700
finding E34 lamp is almost impossible without spending $20. low voltage incandescent lamps are basically being replaced by LEDs. #### Bordodynov Joined May 20, 2015 2,948 There are many lamps made in Russia. I'm from the Republic of Belarus and I'm linking to a branch of a Russian firm. Keep in mind that$1=2.5BYN. I am not suggesting to buy from this store. I'm sure you will find your suppliers. I only show the names of the lamps and their approximate cost.
https://www.chipdip.by/catalog-show/incandescent-lamps?from=suggest_group
I have parametric lamp models in my collection. You can set the operating voltage, power and cold filament resistance, and thermal time constants.

#### Alec_t

Joined Sep 17, 2013
12,456
In place of an incandescent bulb, has anyone here had success with a tiny PTC bead thermistor for the AGC?

Edit:
finding E34 lamp is almost impossible
'Grain of wheat' bulbs seem readily obtainable.

#### Bordodynov

Joined May 20, 2015
2,948

#### Ian0

Joined Aug 7, 2020
4,082
There always used to be glass-cased NTC thermistors called RA53 that seemed to be made specially for the job, designed to maximise self-heating and minimise the effects of external temperature. I bet no-one has seen one for sale recently!

#### AnalogKid

Joined Aug 1, 2013
9,621
I'm thinking more along the lines of a MOSFET, with that mind I came up with this:
The decoupling capacitors, especially C5, should be as close as possible to U2, not as far away as possible.

ak

#### Alec_t

Joined Sep 17, 2013
12,456
..
There always used to be glass-cased NTC thermistors called RA53 that seemed to be made specially for the job
I have one, in the WB oscillator I built donkey's years ago.

#### Audioguru again

Joined Oct 21, 2019
4,103
I think a 240V incandescent bulb's resistance with only a few volts of signal in it will have its resistance change more by ambient temperature than by the signal level.

On ebay they sell 100 6V "grain of wheat" light bulbs for only \$8.99US. But they don't know who makes them and don't know if they are incandescent or LED. They also do not know their current.
They appear to be incandescent but say they have an anode and a cathode.
Amazon also sells some "grain of wheat bulbs" but the only specs are the size and weight.

Wendy, I was wondering why your sinewave has no crossover distortion on its downward swing. Maybe it is your low frequency because it is shown in the LM358 datasheet at 5kHz:

#### Attachments

• 133.2 KB Views: 7