3
\$\begingroup\$

I am doing a project, in which I need a sine wave. I decided to use an RC oscillator because it's a simple circuit to make (I can switch to a different oscillator if needed).

I want to use only one power supply, so I need to make some modifications to the circuit to compensate for the absence of the negative power supply of the amplifier. I tried to add a bias to the positive input of the amplifier, but the simulation doesn't work for me.

The RC oscillator (with two power sources) I built is:

RC oscillator

The circuit worked in the simulation (oscillation of around 33 kHz).

A circuit to make an inverting amplifier with one power supply and a DC bias to the positive input is:

Amplifier with one power supply

This circuit also worked in the simulation.

I simulated several configurations to mix both circuits. One of them is:

Simulation that didn't work

The simulation didn't work here; I get a noisy output which is centered around 0 V. I added the noisy power source (B1) to the positive input to start driving the oscillation. It was needed when I simulated the RC oscillator alone.

If you guys could help me here that would be awesome. Is it even possible to make an RC oscillator with only one power supply? After playing with it, I would assume now that it's a bit different than making an inverting amplifier. For example, I can't really add a capacitor with the same functionality as C1.

If it's possible to make a single supplied RC oscillator, would you advice me as to how I should adjust my circuit to make it? A reference would also be good, if that would be simpler for you.

Edit:

With your help I figured out I didn't add the capacitor with a functionality similar to C1 from the single-supply non-inverting circuit that I drew. For some reason I assumed that adding it would ruin the oscillation. Of course it didn't.

In my project I actually need an oscillation of 1 kHz. I simulated it with 33 kHz because that was the circuit we solved in class and I already had all the needed values. In hindsight it wasn't a smart thing to do.

The circuit I have now is:

New Circuit The difference is the added capacitor C6 which is added before the resistor R5

The plot I get from the simulation is: enter image description here

Thank you all.

\$\endgroup\$
7
  • \$\begingroup\$ Try to simulate the passive network only, without the amp, to see the phase shift. Can you simulate the amp alone, i.e. is the amp model loaded and working? Btw, this oscillator structure seems rather unusual. Too many Rs and Cs. It is good only for learning by experimenting, but there again too many parameter to be constantly aware of. Check the Wien-bridge oscillator if you want sinusoidal output, or some relaxational ones with transistors/comparators. \$\endgroup\$ Aug 30, 2022 at 18:39
  • \$\begingroup\$ RC oscillators are not good for a sine wave generator because the natural response is not a sine wave. \$\endgroup\$
    – Miss Mulan
    Aug 30, 2022 at 20:46
  • 1
    \$\begingroup\$ @MissMulan that is false in phase shift RC oscillators as there is no pulse to respond. \$\endgroup\$ Aug 31, 2022 at 0:21
  • \$\begingroup\$ @AradHilel If you are intending actually building one of these then, when you do, you will discover the practical problems which must be overcome as pointed out in the answers here. Some of these practical issues do not arise in your simulation. \$\endgroup\$
    – user173271
    Aug 31, 2022 at 4:32
  • \$\begingroup\$ @AradHilel if we are done here, you should take the 2 minute tour to understand how to select an answer as your most preferred and close down the session. If you still have queries related to this question then raise a comment under the most appropriate answer. \$\endgroup\$
    – Andy aka
    Aug 31, 2022 at 10:51

2 Answers 2

2
\$\begingroup\$

If you guys could help me here that would be awesome.

R4 is the problem <-- it needs to go to mid-rail else it "off-biases" the op-amp. Or, maybe you forgot to add C1 from your 2nd circuit <-- that will stop the "off-bias" situation.

Is it even possible to make an RC oscillator with only one power supply?

Definitely.

However, when you do get it going you will need an additional circuit to stabilize the amplitude (else the output looks like a clipped sine wave). This can be done with a controlled amplitude clipper using diodes and resistors or with a JFET circuit. Both are shown on the internet but, they are normally associated with Wien-bridge oscillators. However, the same principle can be equally used for the phase-shift oscillator: -

enter image description here

Image from ANALYSIS OF A DIGITALLY CONTROLLED WIEN-BRIDGE OSCILLATOR. Here's a typical version with diode amplitude control: -

enter image description here

Image from here.

\$\endgroup\$
2
\$\begingroup\$

You also need to bias the bottom of the three 200 R resistors to mid-supply but do it using two resistors and a buffer amp (unity gain amp). A buffer is required because those three resistors will sink quite a bit of current which would disturb the reference voltage if just two resistors were used.

In fact those 200R resistors are significantly too low in value. The op amps might not be able to source the required current so I would recommend increasing them well above 200R.

The high pass version of the phase shift oscillator has some flaws. It can create spikes in the sine wave (if the output of the inverting amplifier is allowed to saturate). So, I would recommend swapping the positions of the three resistors and three capacitors to use the low pass version instead. This then has the advantage that you connect the bottom of the three capacitors straight to 0 V and no mid-rail bias is needed for them, however you would then need a buffer amplifier between the phase shift network and R5. Then tap your undistorted output from the output of this buffer and amplify it as necessary.

Changing the high pass version over to the low pass version will alter the oscillation frequency and you will need to recalculate the R & C values to get the required frequency.

For the same phase shift network R & C values, the low pass version will oscillate at about 6X the frequency of the high pass version.

\$\endgroup\$
2
  • \$\begingroup\$ C6 in the second schematic allows the opamp to operate as a DC follower with a DC gain of 1 so it is biased at half the supply voltage. \$\endgroup\$
    – Audioguru
    Aug 30, 2022 at 20:48
  • \$\begingroup\$ @Audioguru Yes, C6 acts to block dc and couple ac and, importantly in this application, the high pass filter formed by the combination of R5 & C6 adds less than 1 degree of phase lead at 1 kHz. \$\endgroup\$
    – user173271
    Aug 31, 2022 at 0:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.