Why Does My Op-Amp Oscillate?
The classic DIP-8 op-amp components you’ll find with old lab equipment are incredibly useful and versatile components. More powerful and specialized op-amps are not normally supplied in the older DIP package, and instead they are most often used as surface-mount components in small packages. Although the package size for op-amps has become more modern, there are some things about op-amps that will never change.
Some op-amp circuits will have a problem relating to their stability, where they will output a signal that was not intended, normally a specific sinusoidal signal. When these op-amp oscillations occur, the designer should take steps to eliminate these by inspecting the feedback loop. When trying to drive more advanced amplifiers into higher frequency ranges, there can be problems in the circuit or on the PCB that cause these oscillations. We’ll look at these challenges in this article.
The Causes of Op-Amp Oscillations
When an op-amp circuit is unstable, it will exhibit a continuous sinusoidal oscillation on the output from the amplifier circuit. Oscillations in an op-amp circuit are easy to identify in measurement and simulation. For example, most commercial op-amps will oscillate at a frequency that can often be measured by low-cost oscilloscopes. An example of an unintended oscillation being excited as a transient response to an input AC wave is shown below.
In some cases, the op-amp output will have this kind of oscillation when driven with a sinusoidal input. The undesired sinusoidal oscillation frequency might be very close to the desired sinusoidal output, so the output should be checked to see if it is the desired signal. There are a few ways to identify whether a sinusoidal output is an unstable oscillation.
- Measure the frequency and compare this to the expected output frequency
- Vary the input voltage level/amplitude; unstable oscillations will be very insensitive to changes in the input voltage level
- For an AC input, vary the input frequency; if the output frequency does not change then the output may be oscillating
The easiest oscillation to identify is in an op-amp that is amplifying a DC voltage. A DC input into an op-amp would normally give an expected DC output, but this will typically be applied with a step function input. If you apply the step input to the required DC value, and you see a sinusoidal output on top of the DC output, then the circuit is oscillating.
From the above examples, we can see that the output produces a sustained oscillation when the circuit is unstable. The output does not necessarily need to be a sine wave; it could be a general periodic waveform. What causes these oscillations and why do they persist? This has to do with the phase margin along the feedback loop.
Feedback Loop Delay and Phase Margin
Instability is produced through an unintentionally large phase delay in the negative feedback loop of an operational amplifier. The phase delay between the amplifier’s output and the inverting feedback node determines the stability of an op-amp circuit. This will happen at certain frequencies (poles) in the loop gain spectrum.
In general, we can define the following relationship for the output voltage and the loop gain of an op-amp. The loop gain can be a complex number, so we can define three possible cases for the loop gain and resulting oscillations.
Op-amp oscillation conditions.
When the output from the op-amp propagates back to the input, it can experience a phase shift, both due to the action of the open-loop gain and any additional reactance in the feedback loop. For example, if there is excess capacitance in the feedback loop, this can result in a phase shift as the output propagates back to the input along the negative feedback loop.
When the phase delay from the output to the inverting input reaches 180°, the input will be out-of-phase with the non-inverting input signal. In this case, when the op-amp shifts its output to equalize the two inputs, the 180° causes the negative feedback to transform into positive feedback. This newly formed positive feedback causes the closed-loop gain will diverge to infinity, and we get a sustained oscillation.
We can now define a certain quantity known as phase margin to define how close the feedback loop’s phase shift has approached 180°. The phase margin can be defined as follows:
Phase margin definition.
In the definition of phase margin, the value of 𝜑 is the phase of the loop gain spectrum, and -180° < φ < 0°. Typically a value for 𝜑 can be found as a function of frequency by looking in a component datasheet. This may be available in a phase shift plot, which ideally should show 𝜑 as a function of frequency.
In some datasheets, such as for a transimpedance amplifier used with a photodiode, the input capacitance on the photodiode will set the cutoff and phase margin as instability is approached. The example below shows a transimpedance gain spectrum with typical values for a photodiode input capacitance, feedback resistor, and photocurrent. The frequency response shows a large peak where we would expect a strong oscillation right at ~450 MHz. Adjusting the input capacitance and feedback resistor would produce a different curve.
The peak in this transimpedance gain plot corresponds to the frequency where we would expect instability.
In general, if there is some reactance somewhere in the feedback loop, this can add to the phase shift (𝜑) or PM value we would expect from the open-loop gain spectrum, and the phase margin can start to approach 0°.
Why would this inversion occur? There are typically three reasons for these instabilities to arise, as summarized in the table below.
Addition of reactance (such as a filter) to a circuit
If a filter is added to a stable op-amp, the filter could make the circuit unstable
Reactance (parasitics) in the PCB layout
This could produce the same effects as adding reactive components
Driving capacitive loads is well-known to cause instabilities
At very high frequencies, it is possible for power amplifiers to experience instability due to output-to-input coupling.
The classic form of amplifier instability is a sustained oscillation. For a designer, the goal is to be able to predict under which conditions the instability will occur. This can typically be done by overlaying the feedback factor (ꞵ) and the open-loop gain on the same graph.
From here we can see where we have loop gain = 0 dB and the resulting phase margin; this occurs where the 1/ꞵ curve and the open-loop gain curve intersect. The resulting phase margin will tell you how close the op-amp is to running at full positive feedback. A standard rule used by designers is to run the op-amp circuit such that the phase margin is kept greater than 45°.
If this produces a growing oscillation, this oscillation could rail out the amplifier, and the device would saturate with the pole in the gain spectrum producing a square wave output. This is basically the same result as if intentional positive feedback was applied through a purely resistive feedback loop (square wave generator).
Underdamped Oscillation (Ringing)
In the above discussion, we need to distinguish between “stability” in terms of sustained oscillations and ringing. The idea behind ringing is a momentary transient response, where an initial oscillation is excited and then dies off relatively quickly in time (underdamped oscillation). Some designers and engineers will refer to the ringing observed below as “instability” or an “unstable” op-amp.
The condition for ringing (see the green curve above) is:
Op-amp underdamped oscillation condition for the loop gain.
Here we have a phase response in both quantities that make up the loop gain, and together they can produce a negative magnitude for the loop gain. In general, this will happen when there is some phase shift along the feedback loop, but not enough phase margin to create a sustained or growing oscillation.
The reality is that these ringing waveforms can arise even without additional reactance and phase margin directly placed in the feedback loop. For example, filters on op-amps can produce these oscillations when the inputs are fed with a step pulse. The same can occur in square wave generation or sudden changes on the input/output in general.
Op-amps experience instability and oscillation due to undesirable delay and phase margin in the feedback loop. The additional phase margin can come from the design of the op-amp or from the PCB layout where the op-amp is placed. Some of the common op-amp circuits where these oscillations can be observed include the following list:
- Buffers for stable reference voltages
- Transimpedance amplifiers
- Transient suppressors on op-amps
- MOSFET gate drivers
- Line driver amplifiers
The other point to note is about terminology. Although an engineer would typically describe an oscillating op-amp as “unstable”, it does meet the mathematical definition of “stable” as defined in stability analysis for dynamical systems. Although the circuit is considered stable mathematically because the oscillation always remains at fixed amplitude, that does not mean the oscillation is desirable. The oscillation should still be eliminated by means of removing excessive phase margin along the feedback loop.
When you’re ready to create and simulate your op-amp circuits and solve problems with oscillations, you can design and simulate your circuits with the simulation tools in PSpice from Cadence. PSpice users can access a powerful SPICE simulator as well as specialty design capabilities like model creation, graphing and analysis tools, and much more.
Subscribe to our newsletter for the latest updates. If you’re looking to learn more about how Cadence has the solution for you, talk to our team of experts.