Nichols chart

In the field of control systems engineering, achieving a delicate balance between stability and performance is a constant challenge. While classic tools like Bode and Nyquist plots offer powerful insights, they often present a fragmented view of a system's behavior. The need for a single, comprehensive dashboard that synthesizes this information has led to the development of one of the most practical tools in the engineer's toolkit: the Nichols chart. It offers a masterful graphical interface that merges frequency and phase information, transforming abstract equations into a tangible landscape for analysis and design.

This article serves as a guide to navigating and utilizing this powerful tool. We will explore the core concepts that underpin the Nichols chart, bridging the gap between open-loop response and closed-loop behavior. You will learn not only how to read the chart but also how to use it as a dynamic workbench for shaping a system's personality. The following chapters will first delve into the foundational ​​Principles and Mechanisms​​ of the chart, explaining its construction and its inherent ability to reveal system stability. Subsequently, we will explore its real-world value through ​​Applications and Interdisciplinary Connections​​, demonstrating how it is used to tune controllers, predict performance, and solve complex design challenges.

Imagine you are a pilot flying through a storm. Your instruments show you your altitude, your compass heading, your speed. Each instrument is useful, but what you’d really love is a single map that not only shows your current position but also instantly reveals how close you are to dangerous mountain peaks, and how a simple change in your throttle will affect your path relative to those dangers. In control systems, we have such a map: the ​​Nichols chart​​. It’s a masterful synthesis of information, a graphical tool that blends the insights of Bode and Nyquist plots into a unified, practical dashboard for system analysis and design.

So, how do we build this wondrous map? We start in the world of the Nyquist plot, where a system's frequency response, $L(j\omega)$, is a curve drawn on the complex plane. Each point on that curve is a complex number, which we can think of in polar coordinates as having a magnitude $r = |L(j\omega)|$ and a phase angle $\phi = \arg(L(j\omega))$.

The Nichols chart is simply a clever change of coordinates. Instead of a Cartesian grid of real and imaginary parts, we create a new grid. The horizontal axis is the phase angle $\phi$, typically measured in degrees. The vertical axis is the magnitude $r$, but plotted on a logarithmic scale—specifically, in ​​decibels (dB)​​, given by $20 \log_{10}(r)$.

Why this particular transformation? Because it turns multiplicative relationships into additive ones. As we shall see, this simple trick makes the effect of many design choices astonishingly clear.

Every point on the map has meaning. A point at $(0 \text{ deg}, 0 \text{ dB})$ represents a response of $1 e^{j0}$, or simply the number 1. A point at $(-90 \text{ deg}, 20 \text{ dB})$ corresponds to a response with a magnitude of $10$ (since $20 \log_{10}(10) = 20$) and a phase lag of $90$ degrees.

Of all the points in the Nyquist plane, the most critical one for stability is the point $-1$. Where does it land on our new map? The number $-1$ has a magnitude of $1$ and a phase of $-180^\circ$. Its magnitude in decibels is $20 \log_{10}(1) = 0$ dB. So, the Nyquist critical point $-1$ becomes the ​​Nichols critical point​​ at $(-180^\circ, 0 \text{ dB})$. This is the point we must keep a very close eye on.

Plotting the open-loop response $L(j\omega)$ is just the first step. The true power of the Nichols chart lies in a pre-drawn set of contours that are overlaid on it. These contours are like the elevation lines on a topographic map; they reveal the characteristics of the closed-loop system, $T(j\omega) = \frac{L(j\omega)}{1+L(j\omega)}$, without you having to calculate a single thing.

The most important of these are the contours of constant closed-loop magnitude, or ​​M-circles​​. Each M-circle is the locus of all open-loop points $(|L|,\angle L)$ that result in the exact same closed-loop magnitude $|T|$. For instance, if your system's plot at some frequency $\omega_0$ happens to fall on the M-circle labeled +2.0 dB, you know instantly that the magnitude of your closed-loop response at that frequency is $|T(j\omega_0)| = 10^{2.0/20} \approx 1.26$. The chart has done the complex calculation for you!.

This is a profound advantage over using separate Bode plots. To find the peak closed-loop resonance, $M_p$, you don't need to perform any calculations. You simply look at your plotted open-loop response and see what is the highest-value M-circle it becomes tangent to. The value of that contour is your peak resonance, $M_p$.

But where do these elegant, nested oval contours come from? They are the shadows of a beautiful geometric property in the Nyquist plane. The condition for a constant closed-loop magnitude, $|T|=M$, can be rewritten as:

This equation states that the ratio of the distance from the point $L(j\omega)$ to the origin (0) and its distance to the critical point ($-1$) is a constant, $M$. The ancient Greeks knew that the locus of points satisfying such a condition is a circle, known as a ​​Circle of Apollonius​​. For each value of $M$, we get a different circle in the Nyquist plane. When these circles are transformed into the Nichols chart's log-magnitude vs. phase coordinates, they form the distinctive M-contours.

A particularly important contour is the one for $M=1$, or $0$ dB. This is the boundary between closed-loop amplification ($|T|>1$) and attenuation ($|T|<1$). In the Nyquist plane, the condition $|L| = |1+L|$ defines the perpendicular bisector of the line segment from 0 to -1, which is the vertical line $\Re\{L\} = -1/2$. On the Nichols chart, this transforms into a large, tear-drop shaped contour that encloses the critical point $(-180^\circ, 0 \text{ dB})$. Any point of your open-loop plot that falls inside this 0 dB contour corresponds to a frequency where your closed-loop system is amplifying the input signal.

The true genius of the Nichols chart reveals itself when you start to tune your controller. Many common adjustments translate into beautifully simple geometric transformations of the entire plot.

• ​​Adjusting Proportional Gain ($K$)​​: Let's say your open-loop transfer function is $L(s) = K \cdot G(s)$. What happens if you double the gain $K$? Since $K$ is a positive real number, it has no effect on the phase. But its effect on the magnitude is additive in the decibel scale:

This means that changing the gain $K$ results in the entire Nichols plot shifting vertically up or down as a rigid shape, with no change in its horizontal position or form. This makes gain adjustment incredibly intuitive. Want a specific gain margin? The gain margin is simply the negative of the plot's magnitude (in dB) at the $-180^\circ$ phase crossing. To achieve a desired margin, you just need to shift the whole curve up or down until it crosses the $-180^\circ$ line at the right level. The amount of that shift in dB tells you exactly how to change your gain $K$. This is far more direct than on a Nyquist plot, where changing gain causes a radial stretching that deforms the plot's shape relative to the M-circles.

• ​​Introducing a Time Delay ($T$)​​: Many real-world systems, from chemical processes to networked robotics, involve time delays. A pure time delay is represented by the transfer function $\exp(-sT)$. What does this do to our plot? At a frequency $\omega$, its frequency response is $\exp(-j\omega T)$. This term has a magnitude of $|\exp(-j\omega T)| = 1$ (or 0 dB) for all frequencies. However, it introduces a phase lag of $-\omega T$ radians. The effect on the Nichols chart is thus a purely horizontal shift: each point on the original plot is shifted to the left by an amount $\omega T \cdot (180/\pi)$ degrees. Unlike the uniform shift from a gain change, this is a "warping" of the plot, as higher frequencies get pushed further to the left, often towards the unstable region around the critical point. This graphically demonstrates why time delays are so often a source of instability.

Finally, the Nichols chart is a complete tool for stability analysis, inheriting the full power of the Nyquist Stability Criterion. The criterion, in its essence, is a method of counting. It relates the number of unstable poles in the open-loop system ($P$) and the number of encirclements of the critical point by the frequency response plot ($N$) to the number of unstable poles in the closed-loop system ($Z$) via the famous equation $Z = P + N$. Stability requires $Z=0$.

On the Nichols chart, "encircling the critical point $(-180^\circ, 0 \text{ dB})$" serves the same role as "encircling -1" on the Nyquist plot.

• ​​The Simple Case​​: If your open-loop system is stable to begin with (so $P=0$), the stability condition simplifies to $Z = N$. To have a stable closed-loop system ($Z=0$), you just need zero net encirclements of the critical point ($N=0$). In most simple cases, this just means "don't let your plot loop around the $(-180^\circ, 0 \text{ dB})$ point."

• ​​The General Case​​: But what if your plant is inherently unstable ($P > 0$)? Then you need the plot to encircle the critical point in a specific way to achieve stability. We need $N = -P$ counter-clockwise encirclements. The Nichols chart gives us a clear way to count these. An encirclement is formed when the plot crosses the $-180^\circ$ phase line at a magnitude greater than 0 dB. The direction of the crossing tells you the sign of the encirclement:

  • Crossing the $-180^\circ$ line from right to left (phase decreasing) above 0 dB contributes ​​one counter-clockwise encirclement​​ ($N=-1$).
  • Crossing from left to right (phase increasing) above 0 dB contributes ​​one clockwise encirclement​​ ($N=+1$).

By adjusting the gain $K$, we shift the plot vertically. This allows us to control which of these phase crossings occur above the 0 dB line. We can therefore "turn on" or "turn off" encirclements to achieve the exact number needed for stability. This makes the Nichols chart an indispensable tool for designing controllers for complex, conditionally stable systems, transforming an abstract algebraic problem into a tangible, visual puzzle.

From a simple coordinate change, we have built a powerful analytical landscape. The Nichols chart is more than a graph; it's a unified framework where system performance and stability are not just numbers, but visible features of a terrain we can learn to navigate and reshape.

After our journey through the principles and mechanics of the Nichols chart, you might be thinking of it as an elegant but abstract mathematical construction. Nothing could be further from the truth. In the hands of an engineer or a scientist, the Nichols chart transforms into a powerful, dynamic tool—a sort of all-in-one dashboard for understanding, predicting, and shaping the behavior of complex systems. Imagine an autopilot system for a large ship; the Nichols chart is like the main instrument panel in the control room, telling the captain at a glance not just where the ship is heading, but how stable its course is, how quickly it responds to commands, and how close it might be to dangerously erratic behavior.

Let's explore this "instrument panel." From the intricate dance of the open-loop curve on the chart, we can read off several critical performance indicators directly. Three of the most important are the Gain Margin, the Phase Margin, and the Bandwidth. The Gain Margin tells you how much you could crank up the system's power before it goes haywire and becomes unstable. The Phase Margin tells you how much time delay or phase lag the system can tolerate before it starts oscillating uncontrollably. And the bandwidth tells you how "fast" the system is—the range of frequencies it can respond to faithfully. For instance, analyzing the Nichols plot for a ship's autopilot might reveal a gain margin of $11.8$ dB, a phase margin of $42.5$ degrees, and a bandwidth of $4.25$ rad/s, giving the engineers a complete, immediate summary of the system's robustness and responsiveness.

The most fundamental question we can ask about a feedback system is: Is it stable? The Nichols chart provides a beautifully intuitive answer. As we've learned, the entire drama of stability plays out in relation to a single, critical point on the chart: the point at $0$ dB gain and $-180^\circ$ phase. This is the "danger zone." The Nyquist stability criterion, in the language of the Nichols chart, essentially says that for most common systems, you are safe as long as your open-loop frequency response curve does not encircle this critical point. The Gain and Phase Margins are simply geometric measurements of how far away the curve is from this point. A Gain Margin of $11.8$ dB means that at the moment the system's phase lag hits the critical value of $-180^\circ$, its gain is still a comfortable $11.8$ dB below the critical $0$ dB line. It's a safety buffer.

But nature is often more subtle than we expect. Common sense might suggest that if a system is stable, adding more gain (amplifying the control action) will eventually push it toward instability. And if you reduce the gain, it should only become more stable. Usually, this is true. But not always! Some systems exhibit a strange and fascinating behavior known as conditional stability. They are stable at very low gains and at very high gains, but become unstable for a range of gains in between.

How can this be? The Nichols chart makes this puzzle visually clear. Imagine a system whose frequency response curve loops around and crosses the critical $-180^\circ$ phase line twice—once from the right at a low gain (say, $-12$ dB) and once again from the left at a higher gain (say, $+8$ dB). If we start with a very low overall gain, the entire curve is shifted far down, and the critical point is safely to the side. If we apply a very high gain, the entire curve is shifted far up, and again, the critical point is avoided. But for an intermediate range of gains, the critical $(0 \text{ dB}, -180^\circ)$ point gets trapped inside the loop of the plot. The curve now encircles the danger zone, and the system becomes violently unstable! The chart tells us precisely what this treacherous range of gain is: in this hypothetical case, any gain adjustment between $-8$ dB and $+12$ dB would spell disaster. This is a profound insight that is difficult to grasp from equations alone but is immediately obvious on the chart. The core task is always to ensure the plot, no matter its shape, keeps the critical point on the "safe" side.

The true power of the Nichols chart lies not just in analysis, but in design. It allows an engineer to become an artist, sculpting the response of a system to meet specific goals. The simplest tool in this toolkit is a proportional gain controller, which is like a simple volume knob.

Let's say a servomechanism is too "springy"—it overshoots its target and oscillates. On the Nichols chart, this behavior is visible as the plot getting too close to the critical point, producing a large peak in the closed-loop M-contours. This peak is called the resonant peak, $M_p$. If the plot for our servo is just tangent to the $M=5.80$ dB contour, we know it has a pronounced resonant peak. The design specification, however, demands this peak be no more than $2.00$ dB. What do we do? Here is the magic: changing the proportional gain $K$ simply shifts the entire Nichols plot vertically up or down by an amount $20\log_{10}(K)$. The shape doesn't change at all! So, to reduce the peak resonance from $5.80$ dB to $2.00$ dB, the engineer simply needs to slide the whole curve down by $5.80 - 2.00 = 3.80$ dB. It's an astonishingly simple and intuitive way to tune a system's behavior.

But a simple gain knob isn't always enough. To achieve more sophisticated goals, engineers use compensators, which are like special filters that reshape the Nichols plot in specific regions.

• To improve a system's long-term accuracy (for example, to eliminate the error when tracking a constant target), an engineer might add a Proportional-Integral (PI) controller. The "integral" part of this controller works its magic at very low frequencies. On the Nichols chart, it grabs the low-frequency end of the plot (which might have started at some finite gain and $0^\circ$ phase) and pulls it down toward $-90^\circ$ phase while stretching it up toward infinite gain. This infinitely high gain at zero frequency is what forces the steady-state error to zero.
• To improve a system's transient response—making it faster and less oscillatory—an engineer might use a Proportional-Derivative (PD) controller. The "derivative" part is sensitive to the rate of change and provides a predictive kick. On the Nichols chart, this appears as a phase lead at high frequencies. It grabs the high-frequency tail of the plot, which might have been heading toward a large phase lag of $-180^\circ$ or more, and bends it back toward $-90^\circ$. This added phase lead is precisely what increases the phase margin, making the system more stable and better damped.

These tools lead to profound questions of design philosophy. Suppose you need to improve both steady-state accuracy (requiring a low-frequency gain boost) and transient response (requiring a phase margin boost). You could use a lag-lead compensator, which first uses a lag element to boost the low-frequency gain and then adds a lead element to fix the phase margin at the new, lower crossover frequency. Or, you could use a lead-lag compensator, which first uses a lead element to set the phase margin and push the crossover frequency higher, and then adds a lag element at very low frequencies to get the required gain. On the Nichols chart, you can see the result of these two strategies. The lead-lag design results in a system with a higher gain crossover frequency. And because a higher crossover frequency generally means a larger closed-loop bandwidth, the lead-lag system will be faster and more responsive. The chart reveals the deep consequences of these engineering trade-offs.

Perhaps the most beautiful aspect of the Nichols chart is its ability to act as a Rosetta Stone, translating between the abstract world of frequency response and the tangible world of time-domain behavior that we experience directly.

When we look at the M-contours on the chart, we are looking at a prediction of the system's "peakiness." A Nichols plot that rises to a high M-contour value, say $M_r$, before falling again, corresponds to a closed-loop system that will have a sharp resonance at a particular frequency. In the time domain, this translates to a "bouncy" or oscillatory step response. There is a deep, mathematical connection between the height of this resonant peak, $M_r$, and the damping ratio, $\zeta$, of the system. The damping ratio is a measure of how quickly the oscillations die out. A low damping ratio means a very bouncy system, while a damping ratio of 1 means no oscillation at all. For a standard second-order system, a peak resonance of $M_r = 1.94$ dB on the Nichols chart directly implies a damping ratio of about $\zeta=0.447$. This powerful link allows an engineer to look at a plot in the frequency domain and say, "I see that peak... this system is going to overshoot by about 20% and oscillate a few times before it settles."

The connection to system speed is just as direct. We've mentioned bandwidth, which is a measure of the range of frequencies a system can follow. How do we find it on the chart? It's simply the frequency at which the open-loop curve crosses the M-contour corresponding to $-3$ dB of closed-loop gain. If a motor controller's Nichols plot crosses the $-3$ dB M-contour at a frequency of $12.5$ rad/s, then its bandwidth is $12.5$ rad/s, plain and simple.

This bridge between worlds shows the unifying power of the frequency-domain perspective. But the story doesn't end here. The fundamental idea of maintaining a "margin" of safety from the critical point has evolved. The classical Gain and Phase Margins are like checking for danger by walking along two straight lines—one of pure gain change, one of pure phase change. But what if gain and phase vary simultaneously, as they often do in the real world due to complex uncertainties? Modern robust control theory generalizes this idea into concepts like ​​disk margins​​. Instead of just a "distance," we define a whole "disk" of complex gain variations around the nominal point. Stability is guaranteed as long as the critical point $-1$ remains outside the locus of forbidden regions defined by the system's response. This leads to a single robustness number, $\rho^\star$, which guarantees stability for an entire family of simultaneous gain and phase changes. This modern perspective is a direct intellectual descendant of the insights first made visible by Nichols and his chart, demonstrating once again that a truly powerful idea never becomes obsolete; it simply becomes the foundation for the next leap in understanding.