Audio: low distortion wien bridge

I do need a audio signal source. i had a beautiful BIG signal generator, but it's too large and had setup time for amplitude stabilize and  with a little bit AM modulator component.

My first poor man generator

*VCC=9V

*U$3,4 use for adjust the frequency

*U$2 use for adjust feedback

*U$1 with a buffer give a adjustable output level control

total THD about 1%. it can be better if a suitable lamp is used. but gain stable will take a noticeable time period, so it not so smooth to use as signal source. and occasionally it stop work,  because there is no voltage regulator, some time power by wrong supply voltage, this is my fault not the circuit, but obliviously signal generator should had a reliable power.  

@2012/9/10

this is a popular configuration several famous ammeter project based on this type Wein bridge oscillator.  I just re-combine them.

you might notice there are something strange things:  

R3:(R6+R7)  should be  2:1, but my circuits use 1:1, this value get from a working  circuits. 

I dig into  this problem for several days, i finally notice the C1, C4, they are Monolithic type capacitor, they are so much UN-stable use as audio  "tank", one of them only 7nF another one is 10nF, simulating confirmed this unsymmetrical  bridge gian loss is 2:1,  not 3:1 and frequency also go up to 3khz, not expected 700Hz.

Finally i use 2 6800pF Mica replaced them, and replace  R6  with a 4.7k resistor too.

drawback

the supply voltage is relatively low, my version use +- 5V supply voltage.  so, if  you want to get better THD like 0.1%  even 1%,  the output amplitude will stay low to keep low THD.  in such a situation JFET's gate almost reach 0V, the low THD mainly provide by resistor R6  [definitely need a POT here].   this make  circuit hard to  alignment and some kind of missing it's design purposes, JFET almost losing it's functionality. 

if try increase amplitude, THD get worse, and output can not generate enough negative bias  [to Vcut-off]  for JFET to reduce THD. the best THD in this situation get  around 3%.

so suggest deploy this circuit by improve the supply voltage.  both simulation and working circuit  confirm it.

use POT replace R,  MPF102 gate bias reach to 0.3V ,  and use a supply  >+8V 

then get THD < 0.1%,  this is a acceptable.  simulation also show just same result.  

Summary:

1. hard align for best performance when supply voltage is lower than +-8V

2. not suitable for mass built.

3. low output amplitude on low THD

Improved the JFET stabilized  low THD oscillator

use additional  gate to amplify the signal.  simulation show while output from U2A is 200mV,  achived  THD <0.05%.

Summary:

1. easy  align for best performance

2. low output amplitude on low THD (common for all JFET controlled wein bridge)

Updated 2012/11/12:

I'm very sorry to find this , the method i use to measure THD is wrong.  the harmonics  dispear in the noise floor,  because the output signal level is too low.   An low pass active filter might a good idea.  this result only indicate the harmonics is  -60dBc (  (-10dBc)  -  (-70dBc) ) , or  THD 0.1%. 

 and another problem is JFET introduce a little bit AM component.

An Even More better Wein Bridge