Hi everyone,

I'm trying to implement a Wien Bridge oscillator running at 500kHz sine wave for my project. I'm using a MCP629x operational amplifier.

http://ww1.microchip.com/downloads/en/DeviceDoc/21812e.pdf

Calculating f = 1/2πRC, I've chosen a C = 10nF and R = 33Ω for my circuit.

But after I have implemented the circuit on a breadboard, i'm only able to get 335kHz as shown above. And the sine wave doesn't seems to be a nice sine wave with some distortion. I tried to use lower values of C and higher values of R but the problem arises. Does that mean is there a problem in the operational amplifier or the problems lies with my drawn schematic which is shown below. Thank you very much.

 

"Gain Bandwidth Product: 300 kHz (typical)"
Maybe the op-amp only has a gain of 1 at 335KHz

 

I didn't realize that I posted the wrong IC that I've used. It is MCP629x instead of MCP623x. So sorry. I've edited the main post. Thanks.

 

Maybe the amp has difficulty driving 33Ω. Try raising the resistance and lowering the capacitance - 1k & 330pF

 

Here's the datasheet:
Aack! You already did that.

 

The specified maximum output current from that '629 opamp is 25mA. At 335kHz the load (2 x 33R + 10nF) is trying to draw peaks of ~36mA. No wonder the waveform is distorted . You need higher resistor values.

 

Maybe the amp has difficulty driving 33Ω. Try raising the resistance and lowering the capacitance - 1k & 330pF

Thanks I'll keep you updated if it is working. As I only tried using 1.0nF to 10nF of capacitors on my circuit.

The specified maximum output current from that '629 opamp is 25mA. At 335kHz the load (2 x 33R + 10nF) is trying to draw peaks of ~36mA. No wonder the waveform is distorted . You need higher resistor values.

May I know how did you calculate the peaks of ~36mA?

 

May I know how did you calculate the peaks of ~36mA?

You may. I ran an LTspice simulation and that's what it showed .

 

You may. I ran an LTspice simulation and that's what it showed .

Is it possible to send me the LTspice simulation file that you have designed? Thanks because I don't have knowledge on LTspice software.

 

Not all oscillator topologies are appropriate for all frequencies. Besides the problem of driving the network impedances, an opamp introduces a phase shift (time delay) into the signal. As the frequency increases, this phase shift becomes a larger and larger percentage of one cycle of the signal of interest. As you have seen, this can pull the oscillator frequency away from the calculated value. Bottom line, a Wein circuit is difficult to achieve at 50 kHz, let alone 500 kHz.

ak

 

Is it possible to send me the LTspice simulation file that you have designed?

Here's the screen shot. The .asc file is attached.

 

Maybe the amp has difficulty driving 33Ω. Try raising the resistance and lowering the capacitance - 1k & 330pF

The specified maximum output current from that '629 opamp is 25mA. At 335kHz the load (2 x 33R + 10nF) is trying to draw peaks of ~36mA. No wonder the waveform is distorted . You need higher resistor values.

Thanks for the help. The signal is not so distorted like previously as shown in the picture below. But I'm only able to get 365kHz than rather 500kHZ that I wanted. Is there any ways I can improve my frequency level?

 

See post #10

 

See post #10

Thanks Albert for highlighting the post to me!

Not all oscillator topologies are appropriate for all frequencies. Besides the problem of driving the network impedances, an opamp introduces a phase shift (time delay) into the signal. As the frequency increases, this phase shift becomes a larger and larger percentage of one cycle of the signal of interest. As you have seen, this can pull the oscillator frequency away from the calculated value. Bottom line, a Wein circuit is difficult to achieve at 50 kHz, let alone 500 kHz.

ak

Thanks ak! So my Wien-bridge oscillator circuit is ok just that because of the phase shift, it affects my theory calculated value. Am I right to say that?

 

So my Wien-bridge oscillator circuit is ok just that because of the phase shift, it affects my theory calculated value. Am I right to say that?

The opamp's phase shift is a big part of the reason, yes. Also, the opamp has a GBW of only 10 MHz, so at 500 kHz its open-loop gain is only 20-- barely adequate for this circuit. Both problems can be alleviated by choosing an opamp with a much higher GBW, such as an LM6171 (100 MHz GBW). Note that the LM6171 needs at least +/- 5V to operate, though.

 

The opamp's phase shift is a big part of the reason, yes. Also, the opamp has a GBW of only 10 MHz, so at 500 kHz its open-loop gain is only 20-- barely adequate for this circuit. Both problems can be alleviated by choosing an opamp with a much higher GBW, such as an LM6171 (100 MHz GBW). Note that the LM6171 needs at least +/- 5V to operate, though.

Thanks for the information OBW0549! I have to change my op-amp in order to satisfy my 500kHz sine wave signal. That means I have to look for an op-amp that has more than 100MHz GBW? Am I right to say that? Thanks for recommending LM6171 but my circuit needs to run on battery so I need to find a smaller voltage to operate the op-amp.

 

Thanks for the information OBW0549! I have to change my op-amp in order to satisfy my 500kHz sine wave signal. That means I have to look for an op-amp that has more than 100MHz GBW? Am I right to say that?

I don't know that you need more than 100 MHz GBW; but IMO 10 MHz is inadequate and a much faster opamp would certainly help. I only cited the LM6171 because I'm familiar with it and it's my "go-to" part when I need something fast.

Thanks for recommending LM6171 but my circuit needs to run on battery so I need to find a smaller voltage to operate the op-amp.

You should be able to find something suitable from Texas Instruments, Analog Devices, Maxim or Linear Technology.

 

Completely agree with OBW. As a rule of thumb, to close a negative feedback loop you want a minimum of 20 dB of negative feedback at the highest frequency of interest. For distortion-sensitive applications such as audio, 40 dB minimum is better.

ak

 

I don't see any bypass capacitors in your photo. You _must_ have a ceramic cap from the power to ground pin of the op-amp.
This cap should have a value of about 0.1uF and the leads must be very short -- a total of less than 1/4" to the IC pins.
You should also put an electrolytic cap to bypass the power buses. The lead length is not so critical for this cap. Its value can be anywhere from 10uF to 100uF.

You have done pretty good with the lead lengths of the other parts. At these frequencies, especially on a solderless breadboard, shorter the better. Never use a wire when the part can connect directly.

 

I don't see any bypass capacitors in your photo. You _must_ have a ceramic cap from the power to ground pin of the op-amp.
This cap should have a value of about 0.1uF and the leads must be very short -- a total of less than 1/4" to the IC pins.
You should also put an electrolytic cap to bypass the power buses. The lead length is not so critical for this cap. Its value can be anywhere from 10uF to 100uF.

You have done pretty good with the lead lengths of the other parts. At these frequencies, especially on a solderless breadboard, shorter the better. Never use a wire when the part can connect directly.

Hi RichardO, thank you for the valuable information.

Sorry, what do you mean by putting a ceramic capacitor of 0.1uF from the power to ground pin of the op-amp? Is it something like this as shown in the picture?

"You should also put an electrolytic cap to bypass the power buses." Is this what you mean by putting an electrolytic cap at the power buses as shown in the picture.

Thanks in advance!