Audio: low distortion wein 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
wein bridge built @Feb/24 2011.
*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.
JFET stabilized low THD oscillator
@2012/9/10
this is a popular configuration several famous ammeter project based on this type Wein bridge oscillator. I just re-combine them.
Fig.1 BAT to +- supply converter
Fig.2 the wein bridge oscillator
mpf102 zero base bias typical resistance : 150R
optimized: Rf: 9.44k
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.
Get the best performance
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.
test result by my SA: PC sound card oscope.
even better THD 0.07%: no obvious harmonics, noise floor 80dB
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%.
a working hw circus confirm that while output reduced to 50 mV-100 mV the THD stay below 0.1%, i.e, 0.06% to 0.09%.
its easy tune R3 to a value achieved this performance. in my built, R3=120ohm, cause R4 is not perfectly 10k, it's 9.6K. practically, to keep THD on this level, JFET gate voltage will drop to around 200 mV, U2A output is around 50 mV. this is a reproducible circuit.
Fig. screen shot of 0.08 THD.
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.
Updated 2012/11/25:
good new this time. it turns out i used a bad audio buffer and bad cable for my sound card SA. after change to "dummy buffer", i re-check the hw again, this time THD is around 0.01% (-80dBc)!!!!
bad things mislead me so much, i even begin to wonder all result other guys posted on INTERNET!.
An Even More better Wein Bridge
to build this block, check this errata and notes:
1) 15pF should be 15uF, i use 10uF.
2) 100k resistor change to 1k
3) Q1 use MPF102
4) LT1115/LT1055 replace with dual op NE5532
5) insert a POT in the 2.4k and 560R resistor, this key for get lowest THD.
6) C1/C2 use 6800pF Mica. R1 R2 chose as your need.
7) LT1004-1.2 use 3.3V Zender and use resistor voltage divider feed to LT1055 + input.
8) 22k resistor could be replace with a 100k resistor for tune the feed back.
THD -88dBc, 0.003% THD.