Stabilising wienbridge sine wave using FET

[GeoffC]

Please can someone give me some advice using a ‘FET’

I am using a ‘FET-2N3819’ to stabilise the output of a ‘Wien-Bridge Oscillator’ which is in the resistor feedback from the output. (Sine-Wave), therefore there will be both positive and negative volts.

The FET’s Gate is connected to receive negative volts from the output via a diode, while the FET is between a feedback resistor and GND.

SO, which of the other two FET connections – Drain or Source - goes to GND?

GND will always be GND, while the resistor will be oscillating. (+/-) – Please can somebody help

Regards Geoff – GeoffC

 

[Alec_t]

The source should go to ground.
Can you post a schematic of your circuit?

 

[LvW]

Yes - the source must be grounded because the voltage VGS is the controlling quantity.
However, do you know that a diode alone is not sufficient for controlling the gate voltage?
You need a loading capacitor and a resistor which allows deloading.

 

[duke37]

Wikipedia give a circuit using a bulb for amplitude control, this is simpler than the fet route.

Horowitz and Hill give a circuit for a bulb and also for fet amplitude control.

 

[AnalogKid]

The sinewave oscillator that started the Hewlett-Packard company used a light bulb for stabilization. Still, there is nothing wrong with using a FET if done well.

ak

 

[GeoffC]

Hi to All'

Many thanks for the replies. What was confusing me, was that the output feedback, which is connected to the Drain is both ‘positive or negative’

I need to oscillate a LED near to 50KHz in order <> to a photo diode that is connected to a turned circuit of 50KHz (48.2KHz)

I have a schematic circuit that I can send to you, IF I can see how to send it. (I’m not very good on finding my way around the forum) The software that I use is ‘R2XL’ (This circuit I will convert to a PCB)
Regards, Geoff. - GeoffC

 

[duke37]

A 2N3819 is symmetrical so source and drain can be interchanged. In this application, it is run as a resistor and will pass current either way.

 

[GeoffC]

The source should go to ground.
Can you post a schematic of your circuit?

Hi Alec
Many thanks for the replies. What was confusing me, was that the output feedback, which is connected to the Drain is both ‘positive or negative’

I need to oscillate a LED near to 50KHz in order <> to a photo diode that is connected to a turned circuit of 50KHz (48.2KHz)

I have a schematic circuit that I can send to you, IF I can see how to send it. (I’m not very good on finding my way around the forum) The software that I use is ‘R2XL’ (This circuit I will convert to a PCB)
Regards, Geoff - (GeoffC)

 

[Alec_t]

If you want to upload a file, use the 'Upload a File' button next to the 'Post Reply' button.

 

[LvW]

What was confusing me, was that the output feedback, which is connected to the Drain is both ‘positive or negative’

For low voltages betqween D and S (VDS lower than app. 0.5V) the FET works as a voltage-controlled resistor (in both directions!).
In any case, you should combine the FET with another series resistor so that the voltage VDS remains within the mentioned limits.
I would not use a light bulb for stabilizing purposes - unless the time constant of the bulb goes well together with the desired oscillation frequency.

 

[AnalogKid]

Please upload a PDF or JPG of the schematic.

 

[GeoffC]

For low voltages betqween D and S (VDS lower than app. 0.5V) the FET works as a voltage-controlled resistor (in both directions!).
In any case, you should combine the FET with another series resistor so that the voltage VDS remains within the mentioned limits.
I would not use a light bulb for stabilizing purposes - unless the time constant of the bulb goes well together with the desired oscillation frequency.

please see my circuit below

The source should go to ground.
Can you post a schematic of your circuit?

I have got the source to GND, and I will try to 'Upload the circuit'




- GeoffC

 

[BobK]

A Wien Bridge oscillator is not the optimal circuit to be creating a frequency to drive an LED. It is great for putting out a pure sine wave, but you do not want a pure sine wave, you want a square wave.

A crystal oscalliator + divider makes much more sense.

Bob

 

[AnalogKid]

What is the application? What is the circuit that the photodiode drives?

ak

 

[LvW]

The circuit - as shown - will not work. The feedback network is not correct - the FET acts as a load to the opamp and not as part of the feedback path.

 

[duke37]

The circuit - as shown - will not work. The feedback network is not correct - the FET acts as a load to the opamp and not as part of the feedback path.

Well spotted.
Horowitz and Hill show a 1k feedback and a 6.8k shunted by the fet to ground.
There is a 1μF capacitor in series to stop any DC passing through the fet.

The time constant of a bulb should not be a problem at 50kHz.

 

[GeoffC]

To: BobK
A Wien Bridge oscillator is not the optimal circuit to be creating a frequency to drive an LED.

I’m inclined to think square waves are rather thumpy in analogue circuits. I’m not saying that you are wrong. I just think sine-wave is more of a gentle approach – GeoffC

To: AnalogKid
What is the application? What is the circuit that the photodiode drives?

All being well, it will transmit to a tuned circuit, using 2off op-amps (TL072).
After that to a precision rectifier again another two Op-Amps - Not yet constructed. -GeoffC


To: Lvw & duke37
The circuit - as shown - will not work. The feedback network is not correct - the FET acts as a load to the op amp and not as part of the feedback path.

Firstly:- The Circuit is basically a ‘Text-Circuit’

OK:
R1 & C1 with R2 & C2 are the engine of the circuit.
R3 4.7KΩ is as near I can get to the (Z) impedance of R1 & C1.
R3 and the FET is the feedback net work. Their resistance combined will be (4.7K+(4.7K/2)= 7KΩ
If the output max. is 10 volts / 7KΩ is a max. of: ………………………….. 1.4 m/Amps.
Regarding the resistance to control the FET 220KΩ + 100KΩ= ……….. 30μAmp (assuming a max 10volts)
Op-Amp - IC 1b is a voltage follower, with a FET - I/P
R5 47K I just stuck in (thinking of a bleed resistor)
So, I’m sure it will work. - GeoffC

 

[LvW]

Firstly:- The Circuit is basically a ‘Text-Circuit’

R3 and the FET is the feedback net work. Their resistance combined will be (4.7K+(4.7K/2)= 7KΩ
So, I’m sure it will work. - GeoffC

Are you really? Did you try?
My recommendation: Do not blindly trust every circuit you find in a "text".
Furthermore - do you know that the opamp input resistance may be regarded as (nearly) infinite?
In case you know that - what do you think is the current through R3?
Did you ever try to UNDERSTAND the circuit?

 

[(*steve*)]

HP published a really interesting paper on Wien bridge oscillators (well, they probably published a few, but I will refer to one) which showed that for maximum performance the opamp must be imperfect.

For interest I built a Wien bridge oscillator using some cheap opamp and got great results. I then built a version with a really nice modern high spec device and it was a pig.

I thought I had made an awful mistake somewhere until I found the article which showed the exact behavior I was seeing with an explanation.

I can't recall off hand now if the bulb was part of the problem and consequently if using another control element ( e.g. A fet) will fix it.

The "problem" is that during the last 20 years or so the quality of opamps has improved so much that it is a lot easier to get into the problem of the circuit being "too good".

 

[(*steve*)]

Here are the relevant posts

https://www.electronicspoint.com/th...-bridge-oscillator.260338/page-2#post-1573956

(and thanks to @The Electrician for pointing me to it)