# Wien Bridge Sine Wave Generator Problem

Discussion in 'The Projects Forum' started by witssq, Aug 2, 2012.

1. ### witssq Thread Starter Active Member

Mar 29, 2009
48
0
The output(pk2pk) of Wien Bridge sine wave generator attached, drops gradually down to some level(1.84->1.80V) when power on.

http://www.piclist.com/images/www/hobby_elec/e_ckt18.htm

Kind response would be appreciated.
Best regards,
SunSung Hwang

File size:
1.2 MB
Views:
12
File size:
209 KB
Views:
9
File size:
27.1 KB
Views:
26
File size:
64.4 KB
Views:
9
2. ### Ron H AAC Fanatic!

Apr 14, 2005
7,050
657
How much time is "gradually", i.e., how long does it take for the output level to stabilize?

3. ### t06afre AAC Fanatic!

May 11, 2009
5,939
1,222
The drift 40 mv seems to be somewhat on the high side. It may be several factors that contribute to this. Like is your design made on a prototype board or a PCB. Or are you just simulating it. And like Ron asked. How long time do it need to reach stable conditions. If this only take a few seconds I would not take much notice of it.

4. ### witssq Thread Starter Active Member

Mar 29, 2009
48
0
i am using pcb. And not simulating. it take time about 1-3.
minutes. When isolate(Open output) from other circuit, it is same.

Best regards,

5. ### DickCappels Moderator

Aug 21, 2008
2,768
669
That is not a very stable circuit. The forward drops of all three diodes add together to make approximately 6 mv p-p drift in amplitude per degree C. That's can't be helping. The operating point of the FET's gate will vary by device and will change with temperature.

You might be able to compensate for all three of the diodes by putting three diodes in series with the source of the FET (cathodes toward ground) putting a high value resistor from the source of the FET to the positive power supply (try 47 k for starters; the value isn't really very important).

I don't know what you can do about the FET unless you have several, in which case, you can just sort through them and find one that is more stable than the others.

It would help to use a bridge with a zener inside it to set the amplitudes instead of just D2 and D3. The Zener can compensate for the drift in the diodes and having the bridge circuit there would increase the overall amplitude, which could then be divided back down to a volt or two, with less drift. The bridge idea is shown in this old circuit for the ancient and venerable CA3140.

6. ### witssq Thread Starter Active Member

Mar 29, 2009
48
0

Parts (D12, D11, D10, and RF31) are added circuit as to the above.

It is hard to find JPETs like BFT that I can buy JPET from MOUSER,
, Digikey, and other distributors. It seems that JPET is changed to MOSFET currently.
It seems that the circuit is to be changed to use MOSFET.
Please recommend suitable MOSFET for that circuit and how the circuit is revised.

CA3019 bridge seems to be disconnected. And I could not find bridge with
zener. So, I revised the circuit with D22(Bridge), D24(Diode), D23(Zener Diode), RF44, RF45, RV10(for voltage divider).
Please review circuit revised on thenbottom area of the attached pdf file,
and recommend part D22, D24 and D23, and the appropriate values for
FR44, RE45 and RV10.

File size:
92.9 KB
Views:
7
File size:
1.5 MB
Views:
8
File size:
40.1 KB
Views:
19
7. ### DickCappels Moderator

Aug 21, 2008
2,768
669

You added them correctly, but after some thought, there is a better way to compensate the forward drops of the diodes and that is to use a Zener with large positive temperature coefficient. See below.

I don't think a MOSFET can replace an JFET in such an application (somebody explain if I have this wrong). I think by raising the amplitude of the sine wave going into the FET circuit you can reduce the drift. You might want to try and add further compensation to the circuit after trying the Zener idea, if you still think it necessary.

In your revised circuit you would not need D16 or D24 because the bridge does the rectifying. You can make a bridge with four small signals diodes, precision matching and thermal tracking that you would get with ten CA3019 is not really needed in this case.

On looking at your revised circuit, I realized that there may be a simpler way to get the same result. Let me first say that I do not see any reason for using a full wave rectifier as the original designer showed (D8, D9, and opamp pins 1,2, and 3) since only one peak of the sine wave should be enough. But let's leave that alone because the more changes are made from a working circuit, the greater chance or breaking it or making it worse.

The Zener can be put in series with the output of the full wave rectifier circuit and fortunately, a Zener in the voltage range in which we are interested has a large positive temperature coefficient that will compensate for most of the change in diode drops of D7, D8, and D9. Referring to the attached "lower_TC_Sine.pdf": the added 4.7 k resistor is to assure enough current for the Zener to achieve a sufficient voltage drop.

Would you try this and let us know if any improvements to the percent change in amplitude?

File size:
36 KB
Views:
21
8. ### t06afre AAC Fanatic!

May 11, 2009
5,939
1,222
I think you may get better results by using a less crude precision rectifier. This may also give you some ideas. http://sound.westhost.com/appnotes/an001.htm
Also if the 2SK30ATM is hard to get. Try the NTE Equivalent NTE459. Other
small signal NChannel Silicon JFET Transistors may also be used. With some tuning of the values. This book also have a good section on oscillators http://www.ti.com/lit/an/slod006b/slod006b.pdf You could try the clasic method using a small light bulb to stabilize the amplitude. This document on page 3 may share some light on how to select the bulb http://www.ti.com/sc/docs/apps/msp/journal/aug2000/aug_07.pdf
Remember that all circuits will have some warmup time. Much like a car engine. And you must always do the tuning then stable conditions are reached
Another thing. If for some reason a opamp is hard to get. It may very well be that you can get something that is compatible. Most often it is enough to duplicate the principals given in datasheets or application notes. At least if you pick opamps that match the used ones fairly. You can ask the forum for help if you are in doubt

Last edited: Aug 4, 2012
9. ### witssq Thread Starter Active Member

Mar 29, 2009
48
0

I will try out according to your suggestion.
The result will be posted, too.

To : DickCappels

It seems to be considered that if 2SK30A is used, output voltage is 3.3V(p2p) and if BFT46 is used output voltage is 1.8V(p2p)
The drift response of 2SK30A is better than BFT46(SOT-23). But 2SK30A(DIP) is not easily purchased. and my PCB is SOT-23.
Please let me know about that This status may be negligible for deciding Zener value(1N5235).

I found some more JFETs at MOUSER. All things would be tested.

To : t06afre

I will read documentation suggested. After reviewed I would respond.

Best regards,

10. ### DickCappels Moderator

Aug 21, 2008
2,768
669
I would like to let you know that the basic principle of using an un-compensated JFET transfer function as a reference to stabilize anything is not the best approach. Even the light bulb feedback method would be better (note: the light bulb approach's main failing in terms of accuracy and stability is a frequencies close to and below the time constant of the filament).

There are some better approaches in Mancini's Op Amps For Everyone
www.ti.com/lit/an/slod006b/slod006b.pdf

Chapter 15 is devoted entirely to Sine Wave Oscillators.

And if you are still in a reading mood, try the National Semiconductor (Now TI) Application Note 263, Sine Wave Generation Techniques
http://www.ti.com/general/docs/lit/getliterature.tsp?literatureNumber=snoa665b&fileType=pdf

11. ### witssq Thread Starter Active Member

Mar 29, 2009
48
0
1)
Your Zener(1N5235) and 4.7K suggestion is tested. The result output scope graph is attached. The major characteristics is that amplitude is raised, but distortion and shift(amplitude center) is found. The used zener is
MMBZ5233B-6V.

2)
The 3 diodes to ground option is not oscillated.

Would you please diagonose this situation?

• ###### ZENER6V.jpg
File size:
126.4 KB
Views:
16
Last edited: Aug 7, 2012
12. ### DickCappels Moderator

Aug 21, 2008
2,768
669
I suspect the amplitude is getting a little high for the FET. As I mentioned, the circuit you are using has some basic problems and there are better solutions out there. Are you interested in stretching your horizon and trying one of the other options?

It seems that we have run into "conservation of problems"; "Problems can neither be created nor destroyed, only changed in form" -David Wilton.

13. ### witssq Thread Starter Active Member

Mar 29, 2009
48
0
Of course, I'm interested in better option. I will try if you suggested another option.

Regards,

14. ### DickCappels Moderator

Aug 21, 2008
2,768
669
Have a look through the pdf files that t06afre and I recommended and see if any of the circuit in there appeal to you. If you don't see what you are looking for, tell us what you are looking for and we'll hlep.