Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Wien bridge oscillator feedback loop

Status
Not open for further replies.

Eight

Member level 2
Joined
Mar 8, 2013
Messages
53
Helped
1
Reputation
2
Reaction score
1
Trophy points
1,288
Activity points
2,084
Hi.

I'm trying to build a Wien bridge oscillator for producing an AC signal of about 4 kHz. I've done a lot of searching, and I found a number of circuit designs. Most of them are very similar, but there are a lot of differences when it comes to the feedback loop modification. Actually there are too many of them because I'm being confused and don't know which one to use. So I'm looking for some recommendations. The output frequency will be static of about 4 kHz, and the opamp (TL071) supply voltage will be a dual +/-9V. The output doesn't have to be *exactly* rail-to-rail, but I'd be happy, if the sine peaks reach at least +/- 7v. The output will go through a 100 kOhm potentiometer to GND. The adjustable part will be connected to another opamp, so we won't be drawing much current.

Can someone suggest me which design to use?

This one is the most basic design. I'd like to avoid using the lamp.
1-Wien-Bridge-basic-lamp.gif

The lamp can be replaced with resistor. This is supposedly unstable.
2-Wien-Bridge-resistors.gif

Some people suggested that two diodes be added for stabilization.
3-Wien-Bridge-diodes.gif

Another diode placement. This is where it gets confusing.
4-Wien-Bridge-diodes-2.gif

Yet another different design:
5-Wien-Bridge-diodes-3.png

I've also seen some designs with JFETs.
6-Wien-Bridge-JFET.png
 
Last edited:

The schematics are different regarding the amplitude stabilization mechanism.

Every oscillator needs a loop gain > 1 to start oscillations. If no amplitude control with variable gain (incandescent lamp PTC, FET as variable resistor) exists, clipping the oscillator output is the only available mechanism to stabilize the voltage, involving respective waveform distortions. You can adjust the gain carefully so that the initial loop gain is only slightly above unity, resulting in less distortions. Similarly, choosing the resistor values in the below schematic carefully makes the gain step switched by the diodes small and generates less distortion.

In a short, if you intend a pure sine waveform, go for the circuits with automatic gain control. Otherwise simple clipping might be sufficient.

The breaking point in the below schematic is incorrect. You have two parallel feedback paths which together make the total loop gain. Thus you need to break at the OP output.

image002.gif
 

It depends on how adverse you are to distortion.

If distortion is not a big deal, then back to back diodes (or inverse series zeners) are a simple effective way to stabilise the output amplitude.

Where lowest possible distortion is a priority, more sophistication is required in reducing the gain of the whole waveform uniformly, not just clipping off the peaks.
The first filament lamp circuit is very ancient, but it is also very good.
Finding a suitable lamp can be tricky, but not impossible.
The Jfet circuit is not too bad either for low distortion.

Not shown, but a light dependant resistor (LDR) illuminated by a small lamp also works very well as a very low distortion gain control.
 

Yes - I agree with FvM`s selection. For moderate distortion requirements amplitude stabilization with a pair of diodes is recommended. More than that - this method offers a pretty simple method for calculating resp. designing the desired output amplitude.
Remember: The resistors in the feedback path (voltage divider) must have a 2:1 ratio under operational conditions. For calculation of suitable resistor values satisfying your amplitude requirements you can assume that the peak voltage across the diodes is approximately 0.6 volts.
The beaking pont (mentioned by FvM) is important only in case you want to simulate the loop gain (check of the oscillation condition).
 

Thanks for the comments.

Perfect sine wave is not a requirement, so I'd prefer go with design simplicity. I was using the setup in picture 5 on my first post, but replaced the resistor R4 with a variable one, and the diodes were both 1N4001. The circuit seems to perform quite well on the oscilloscope and a 10kOhm load. By the way, what is the difference, if I use two diodes in a parallel configuration (same as on image 5) vs two serially connected diodes (back to back)?

Cheers!
 

Notice the difference, the serial connected diodes must be zener diodes, the parallel connected are regular rectifier diodes.

You can expect smoother voltage clipping and less distortion if the diodes bypass only a small part of the total feedback resistor, as in schematic 4 and 5.
 

The first filament lamp circuit is very ancient, but it is also very good. .

I have seen this, in one of the HP's (ancient) function generators (eons ago) and I got curious about the lamp. It was a tube circuit and the filaments did provide sufficient illumination. What a lamp is doing inside?
 

The filament lamp is used as PTC, self-heated by the signal. This usually a small indicator lamp, e.g. rated 12 or 24 V, 20 to 100 mA, glowing dim red at the steady state operation point. An incandescent lamp has a resistance variation of 1:10 between cold and hot state.
 

The lamp is a high impedance tungsten filament thing that runs maybe just red hot with a very few volts.

The idea is that as the oscillation amplitude builds up, the temperature of the filament increases. The resistance of the tungsten filament increases fairly dramatically with temperature.

The lamp is placed in series with a resistor which forms part of a voltage divider in the feedback path.
As oscillation amplitude increases, the lamp glows brighter and the change in resistance stabilises the amplitude.

The clever thing about it is the thermal inertia of the filament is long enough to not change during one cycle of oscillation, so it does not distort the waveform, but its fast enough to rapidly stabilise the amplitude.

It works down to a few Hz where the lamp may be seen to visibly flash, and distortion increases. But at higher frequencies it works pretty well. Back in the vacuum tube era, before silicon diodes and Jfets, it was pretty common.

*edit*

You beat me to it.
 

Both are correct.
We have two alternatives for opening the loop(s) because we have two stable loops under open conditions: (a) open both feedback loops or (b) open the positive loop only.
Both loop gain functions are different - however, both will pass the critical point at the same frequency in case the loop gain is unity at this frequency.
This is the only information that matters for designing an oscillator.
But it should be mentioned, that - for practical reasons - it is recommended to follow the approach (b).
Simulations will be very critical for case (a) because the loop contains a very high gain stage (1E4...1E5) multiplied by a very low feedback factor (1E-5....1E-4).
 
  • Like
Reactions: CataM

    CataM

    Points: 2
    Helpful Answer Positive Rating
I must confess that I didn't think thoroughly before replying.

My answer would be correct for determining the amplifier loop gain, as it's performed to check OP stability. But here you want to check the oscillation condition for the Wien bridge, presuming that the overall OP loop is stable. Then you'll break the positive loop.
 
  • Like
Reactions: CataM

    CataM

    Points: 2
    Helpful Answer Positive Rating
Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top