Bode Plot

A Complete Step-by-Step Guide with Theory

This document covers everything you need to construct a Bode plot from scratch — what a Bode plot
is, why each rule exists, the mathematics behind it, and a fully worked example using G(s) = 10/(s +
2).

1. What is a Bode plot and why do we draw it?

2. The big idea — break the transfer function into pieces
3. The 5 basic factors and their rules
4. Why a pole gives exactly −20 dB/decade (proof)
5. Step-by-step worked example: G(s) = 10/(s + 2)
6. Phase plot explained
7. The −3 dB correction
8. Second-order factors and resonance
9. Complete summary and memory tricks
1. What is a Bode plot and why do we draw it?
When you build a circuit, amplifier, or control system, a signal goes in and a (possibly changed) signal
comes out. The key question is: how does the system treat signals at different frequencies?

A Bode plot answers this with two graphs drawn on the same frequency axis:

• Magnitude plot — does the system make the signal louder (gain > 1) or quieter (gain < 1) at each
frequency? Plotted in dB on the y-axis.
• Phase plot — does the system delay or advance the signal at each frequency? Plotted in degrees
on the y-axis.

Both plots use a logarithmic frequency axis. Real systems can span from 0.001 Hz to 1,000,000
Hz. A log scale lets us see the entire range at once — compressing the wide span into an even,
readable spread.

Why? We use dB (decibels) for magnitude because multiplication of gains becomes addition
in dB. If stage 1 has gain K1 and stage 2 has gain K2, total gain = K1 × K2. In dB: total =
20log(K1) + 20log(K2). Addition is always easier to draw than multiplication.

K>1

K<1

→ ω (log scale)
Magnitude plot: a constant gain K shifts the plot up or down as a flat line.

2. The big idea — break the transfer function into pieces

Every transfer function G(s) is a product of simpler pieces. The key insight is:

log(A × B × C) = log(A) + log(B) + log(C) → Plot each piece separately, then add them up.

This works because magnitude in dB is just 20·log of the gain. Multiplication in the frequency domain
becomes addition on the Bode plot. So you never have to deal with the full transfer function at once
— just its pieces.

The pieces you will encounter in every transfer function:

• A constant K — shifts the plot up or down uniformly.

• Poles/zeros at the origin — terms like s or 1/s, giving constant slopes.
• Simple (first-order) poles or zeros — the most common type, creates a bend at one corner
frequency.
• Second-order (quadratic) factors — creates a steeper bend and possibly a resonance peak.
3. The 5 basic factors and their rules

3a. Constant gain K

A plain multiplier like K = 10. It contributes the same gain at every frequency.

Magnitude contribution: 20·log(K) dB — a flat horizontal line.

Phase contribution: 0° if K > 0, or −180° if K < 0 (phase flip).

Why? K doesn't depend on ω at all. So it shifts the entire plot up or down equally at every
frequency — hence a flat line, not a slope.

3b. Pole or zero at the origin — (jω)^N

These are pure s terms in numerator (zero) or denominator (pole). They affect all frequencies equally
with a constant slope.

Factor Magnitude slope Phase (all frequencies)

Zero: (jω) +20N dB/decade +90° × N (constant)

Pole: 1/(jω) −20N dB/decade −90° × N (constant)

Why? At the origin, there's no 'corner' — the term is just (jω)^N which is proportional to ω^N
everywhere. That's a straight line on a log-log plot, passing through 0 dB at ω = 1, with slope
±20N dB/decade.

3c. Simple (first-order) pole — 1/(1 + jω/ωc)

This is the most common building block. The number ωc is the corner frequency — the frequency
where the plot bends.

Region Magnitude Phase

Below ωc 0 dB (flat) ≈ 0°

At ωc (exact) −3 dB −45°

Above ωc Slope −20 dB/dec → −90°

Phase transitions over two decades: starts changing at 0.1·ωc, finishes at 10·ωc.
 0 dB
-3 dB at ωc

-20 dB/dec

ωc →ω
Magnitude of a simple pole: flat at 0 dB until ωc, then falls at −20 dB/decade. Real curve dips 3 dB at the corner.

0°

-45° at ωc
-45°

-90°

0.1ωc 10ωc
Phase of a simple pole: starts at 0°, transitions to −90° over two decades, passing through −45° exactly at ωc.

For a first-order zero, (1 + jω/ωc), simply flip all signs: +20 dB/decade above ωc, phase goes from 0°
to +90°.

3d. Pure time delay — e^(−jωT)

A system that simply delays the output by time T has no effect on magnitude (it's always 1 = 0 dB),
but phase decreases linearly: Phase = −ωT radians.

Why? A pure delay doesn't change how loud anything is — it just shifts it in time. But at
higher frequencies, the same time delay T represents a larger fraction of the period, so the
phase lag −ωT keeps growing without bound.

3e. Second-order (quadratic) pole

These arise from systems with two energy-storing elements (like an LC circuit or a
mass-spring-damper). The standard form is:

1 / (1 + 2ζ(jω/ωn) + (jω/ωn)²)
where ωn is the natural frequency (corner) and ζ (zeta) is the damping ratio.

Region Magnitude Phase

Below ωn 0 dB (flat) ≈ 0°

At ωn Peak if ζ < 0.707 −90° always

Above ωn −40 dB/decade → −180°

Resonance peak magnitude ≈ −20·log(2ζ) dB above the 0 dB line. The smaller ζ, the sharper and
taller the peak.

Why? A second-order system is like a spring-mass. Drive it at its natural frequency ωn and it
resonates. With low damping (small ζ), very little energy is absorbed, so the response grows
large — a tall resonance peak. High damping kills the peak smoothly. The slope doubles to
−40 dB/decade because there are two poles acting together.
4. Why a pole gives exactly −20 dB/decade — the proof
This is the most important 'why' in Bode plots. Here is the complete explanation, step by step.

Step A: Write out the magnitude of a simple pole

For the factor 1/(1 + jω/ωc), the magnitude is:

|G| = 1 / √(1 + (ω/ωc)²)

Step B: Look at what happens at high frequency (ω >> ωc)

When ω is much larger than ωc, the '1' inside the square root becomes tiny compared to (ω/ωc)²:

√(1 + (ω/ωc)²) ≈ √((ω/ωc)²) = ω/ωc

So the magnitude simplifies to:

|G| ≈ 1 / (ω/ωc) = ωc / ω

Step C: See what happens as ω increases by 10× (one decade)

At frequency ω: |G| = ωc / ω

At frequency 10ω: |G| = ωc / (10ω) = (1/10) × (ωc/ω)

The magnitude dropped by a factor of 10. In dB:

20 · log■■(1/10) = 20 × (−1) = −20 dB

Every time the frequency increases by 10× (one decade), the magnitude drops by exactly 10×
= exactly −20 dB. This is the origin of the −20 dB/decade rule.

-20dB

-20dB

-20dB

-20dB

10¹ 10¹ → ω10¹

(each step = ×10) 10¹
Each step right on the log axis is ×10 frequency. Each step down is exactly −20 dB. The rule is a direct consequence
of the 1/ω relationship.

Step D: Why is it FLAT before the corner?

At low frequency (ω << ωc):

1 + (ω/ωc)² ≈ 1 → |G| ≈ 1 → 0 dB
The (ω/ωc)² term is negligible, so the pole contributes nothing — it's 'sleeping'. The flat region exists
because the pole hasn't woken up yet. It only starts mattering once ω approaches ωc.
This is why we draw two separate lines meeting at ωc: a flat line before the corner (where the pole is
inactive) and a −20 dB/decade line after (where the pole dominates). The corner frequency ωc is the
exact boundary between these two regimes.

Step E: Why is the phase exactly −45° at ωc?

The exact phase of a simple pole is:

Phase = −arctan(ω/ωc)
At ω = ωc: Phase = −arctan(1) = −45°. This is exact, not an approximation.

Why? arctan(x) goes from 0° (at x=0) to 90° (at x→∞). The transition is smooth and
symmetric on a log scale — spread over two decades centered on ωc. At 0.1·ωc: arctan(0.1)
≈ −5.7°. At 10·ωc: arctan(10) ≈ −84.3°. The phase is essentially done transitioning outside
this range.
5. Fully worked example: G(s) = 10 / (s + 2)
We will now construct the complete Bode plot for this transfer function, explaining every single step.

Step 1 — Convert to standard form

Replace s with jω to get the frequency response:

G(jω) = 10 / (jω + 2)
Factor out the 2 from the denominator so it looks like (1 + jω/something):

G(jω) = 10 / [2 · (1 + jω/2)] = 5 · 1/(1 + jω/2)

Now we can clearly see two pieces:

• Constant K = 5
• Simple pole at ωc = 2 rad/s

Why? We factor it this way so the corner frequency appears directly as the number in the
denominator. If we left it as (jω + 2), we'd have to remember that the corner is at ω = 2. In
standard form, it's obvious: ωc = 2.

Step 2 — Compute magnitude of K = 5

20 · log■■(5) = 20 × 0.699 = 13.98 dB ≈ 14 dB
This constant shifts the entire magnitude plot up by 14 dB. Draw a flat horizontal line at +14 dB
across all frequencies.

Step 3 — Plot the pole contribution

The pole is 1/(1 + jω/2) with ωc = 2.

• For ω < 2: magnitude = 0 dB (flat line at 0 dB)

• For ω > 2: slope = −20 dB/decade (line falling to the right)
• Transition happens at ωc = 2

Step 4 — Add the two contributions to get the total

Total magnitude = K contribution + pole contribution (in dB, just add them):

• For ω < 2: total = 14 dB + 0 dB = +14 dB (flat)

• For ω > 2: total = 14 dB + (pole slope) = +14 dB line falling at −20 dB/decade
The total plot is a flat line at +14 dB that bends downward at ωc = 2 and falls at −20 dB/decade
thereafter.
 14 dB K=5

0 dB
total

-20
pole

ωc=2
→ ω (log)
Magnitude plot for G(s) = 10/(s+2). Blue dashed = K=5 (+14 dB flat). Red dashed = pole. Green solid = total
asymptote. Corner at ωc = 2 rad/s.

6. Phase plot for G(s) = 10/(s + 2)

Phase contributions from each piece:

• K = 5 (positive): contributes 0° — no phase shift.

• Simple pole at ωc = 2: phase goes from 0° → −90° over two decades.

The phase transition for the pole:

Frequency Phase Reason

ω < 0.2 (= 0.1 × ωc) ≈ 0° Pole not yet active

ω = 0.2 to 20 Transitions 0° → −90° Smooth arctan curve

ω = 2 (= ωc) Exactly −45° Always true for simple pole

ω > 20 (= 10 × ωc) ≈ −90° Pole fully active

0°

-45°
-45°

-90°

ωc=2
Phase plot for G(s) = 10/(s+2). Phase transitions from 0° to −90° over two decades (0.2 to 20 rad/s). Exactly −45° at
ωc = 2.
 Why? The phase transition is gradual (over 2 decades) because arctan(ω/ωc) is a smooth
mathematical function, not a step. It's nearly flat far from the corner, and steepest right at the
corner. The two-decade spread is a consequence of the shape of the arctan function on a
logarithmic frequency axis.

7. The −3 dB correction
The asymptote method is an approximation. The real curve differs slightly:

Location Asymptote Actual value

At ωc (corner) Asymptote says: 0 dB Reality: −3 dB

At 0.5·ωc or 2·ωc Asymptote says: 0 dB Reality: −1 dB

Far from ωc Asymptote accurate < 0.3 dB error

For our example G(s) = 10/(s+2), the total magnitude at ω = 2 (corner):

Asymptote prediction: +14 dB

Actual value: 20·log■■(|10 / (2j + 2)|) = 20·log■■(10/2√2) =
20·log■■(3.535) ≈ 10.97 dB
That's 14 − 10.97 ≈ 3 dB below the asymptote, exactly as predicted.

Why? The asymptote makes a sharp corner at ωc, but the real curve is smooth — it 'rounds
the corner'. The maximum error always occurs right at ωc and is always exactly 3 dB for a
simple pole or zero. For most engineering purposes, the asymptote is accurate enough.
Apply the correction only when you need precision.

8. Second-order factors and the role of damping ζ

When a transfer function has a quadratic denominator, the damping ratio ζ determines whether the
response is smooth or has a resonance peak.

Damping ζ System type Bode magnitude behavior

ζ > 0.707 Overdamped No peak, smooth −40 dB/dec bend

ζ = 0.707 Critically damped No peak, flattest possible response

0.5 < ζ < 0.707 Lightly damped Small peak ≈ few dB

ζ = 0.1 Very underdamped Large peak ≈ +14 dB spike!

The resonance peak height (above 0 dB line) is:

Peak ≈ −20 · log■■(2ζ) [dB above asymptote]

Examples: ζ=0.5 → peak ≈ 0 dB (no real peak). ζ=0.1 → peak ≈ +14 dB.
Why? A second-order system is mathematically identical to a mass-spring-damper. The
natural frequency ωn is the resonant frequency (where it 'wants' to oscillate). When driven at
ωn with low damping, energy builds up faster than it's lost — the amplitude grows large. This
is resonance. At exactly ωn, the phase is always −90°, regardless of ζ. Above ωn, the system
falls at −40 dB/decade because there are effectively two poles acting together.
9. Complete construction procedure summary
Follow these steps in order for any transfer function:

Step 1 Standard form

Replace s → jω. Factor denominator so every term looks like (1 + jω/ωc). This reveals
the corner frequency directly.

Step 2 List corner frequencies

Every first-order factor contributes one corner at ωc. Every second-order factor
contributes one corner at ωn. Mark them all on the frequency axis first.

Step 3 Starting line

Compute 20·log■■(K) and draw a flat horizontal line at that dB level. If there are
(jω)^N terms at the origin, tilt the starting line by ±20N dB/decade.

Step 4 Bend at each corner

Scan left to right. At each corner: pole → slope −20 dB/dec; zero → slope +20
dB/dec; second-order pole → slope −40 dB/dec. Keep a running total of slopes.

Step 5 Phase plot

Sum phase from each factor. Simple pole: 0° → −90° over 2 decades centered on ωc,
passing through −45° at ωc. Simple zero: 0° → +90°. Pole at origin: constant −90°.

Step 6 Apply corrections

If precision matters, subtract 3 dB from the magnitude asymptote at each corner
frequency (for poles; add 3 dB for zeros). The actual curve rounds the corners.

Step 7 Check for resonance

For second-order factors, check ζ. If ζ < 0.707, add a resonance peak of −20·log(2ζ)
dB at ωn. The phase still passes through −90° at ωn.

Memory tricks
Rule Meaning

Each corner, adjust the slope Scan left to right; every corner changes slope by ±20 dB/dec

−3 dB at the corner Real curve dips 3 dB below the asymptote exactly at ωc

Phase is lazy Phase starts changing one decade before the corner, finishes one dec

Second-order doubles everything −40 dB/dec slope, −180° total phase, and check ζ for resonance peak

Key formula: 20·log■■(K) gives magnitude in dB. Corner frequency: ωc = the number in the
denominator of (1 + jω/ωc). At ωc: always −3 dB and −45° for a simple pole.