Create your own
Lesson illustration

Half-Power Cutoff: Understanding the 1/√2 Magnitude Ratio and −3 dB Point

Hello. In the previous lesson, you calculated the phase of a frequency response and saw that a first-order RC low-pass filter has a phase of at its cutoff frequency. We now connect the other familiar cutoff marker, , to a physical statement about power.

The key idea is simple but easy to misstate: half power does not mean half voltage. At a half-power cutoff, the voltage magnitude has fallen only to , about , of its passband value. This lesson derives that relationship, expresses it in decibels, and connects it to the cutoff point of an RC filter.


Voltage magnitude and power are related by a square

Your transfer function magnitude,

is a voltage-amplitude ratio. It tells you what fraction of the input voltage amplitude appears at the output at a particular frequency.

For a sinusoidal voltage across a resistor , the average power is

Thus, provided we compare voltages across the same resistance (or the same reference impedance), power is proportional to voltage squared:

Suppose that the passband voltage is , and that at some frequency the output voltage magnitude is a fraction of the passband value:

The corresponding power ratio is therefore

At the half-power point,

so the voltage magnitude ratio must satisfy

Taking the positive square root, since magnitude cannot be negative,

Therefore, the half-power cutoff condition is

For a unity-gain passive filter whose passband magnitude is approximately one, this is often written more simply as

A numerical picture

Imagine that a passband output is across a resistor. Its power is

At the half-power frequency, the voltage is not . It is

The resulting power is

ConditionVoltage ratioPower ratio
Passband reference
Half-power cutoff
Half voltage

So a half-voltage point would be a quarter-power point, not a half-power point.

In introductory filter analysis, “power” at cutoff is often shorthand for this squared-magnitude comparison. For a voltage measured across a reactive component, such as the capacitor in an unloaded RC low-pass filter, the capacitor itself does not dissipate average real power. The half-power terminology remains useful because is the normalized power-transfer quantity when the voltage is referenced to an equal resistive impedance or a specified load.


Why half power is called

There are two equivalent decibel formulas:

and, when the compared voltages have equal reference resistance,

The factor of , rather than , in the voltage form comes directly from the square relation between power and voltage.

At half power,

Therefore,

Using the associated voltage ratio gives exactly the same result:

In normal filter work, engineers round this value:

Be precise about the wording:

  • A drop of means half the power.
  • A drop of means the voltage magnitude is of its reference value.
  • A voltage magnitude of corresponds to

and hence one-quarter power.

Half-Power Point and the -3dB Point. Are they the same?

Watch “Half-Power Point and the -3dB Point. Are they the same?” by twdr for a visual derivation that begins with a halved power value and then obtains the associated voltage ratio.

Watch half power derivation. Follow the two representations of the same condition: P_o/P_i=1/2, then V_o/V_i=1/\sqrt{2}. Pause when the numerical value 0.707 appears and relate it to the square-law relationship P\propto V^2.


The cutoff point on a real filter response

An ideal low-pass filter would change instantly from full transmission to zero transmission. Physical passive filters do not: their magnitude response changes gradually. The cutoff frequency gives a standardized, reproducible location within that transition.

What Is a Low Pass Filter? A Tutorial on the Basics of Passive RC Filters

Read the “The Cutoff Frequency” section from All About Circuits to connect the mathematical -3\ \text{dB} definition with the gradual shape of a practical RC low-pass response.

In the section titled “The Cutoff Frequency,” begin with the explanation of gradual transition. Then continue through the remainder of that section. Focus on why a real RC filter has no perfectly sharp boundary between passband and stopband, and why the -3\ \text{dB} level supplies a conventional cutoff marker.

For a normalized low-pass response, the passband level is

which is

The cutoff level is below that passband level:

More generally, if a filter has passband gain in decibels rather than , the half-power cutoff level is

This “relative to the passband” wording matters. A passive filter often has a unity passband gain in simplified analysis, but component loading can make the actual passband gain lower than one. The cutoff is still below the relevant maximum or nominal passband level, not necessarily at an absolute level of .


Connection to the first-order RC low-pass filter

From the previous lesson, the standard RC low-pass transfer function is

Its magnitude is

The cutoff angular frequency is

Substitute :

Because

we obtain

That directly proves that the RC cutoff frequency is the half-power frequency. Expressed in decibels,

For this particular first-order RC low-pass circuit, cutoff has two useful simultaneous signatures:

The first two define the half-power magnitude condition; the phase value is a characteristic of this specific first-order low-pass topology.


Key takeaways

For equal reference resistances,

Therefore, at half power,

the corresponding magnitude ratio is

In decibels, the same condition is

Thus, “cutoff,” “half-power point,” and “ point” refer to the same standard magnitude condition. In the next lesson, you will use magnitude-response plots to identify passbands, stopbands, and cutoff frequencies directly from the graph.

Can't find a good explanation? Sign up and we'll make it for you

Sign up