When you need to drop a signal's amplitude without distorting its waveform or messing up your impedance match, you build a resistive pad. While there are several topologies (Pi, Bridged-T, L-pad), the unbalanced T-pad is the workhorse of both 50 Ω RF bench setups and 600 Ω audio lines. To design one, you need the attenuator formula, which translates a target decibel (dB) loss into exact physical resistor values while maintaining the characteristic impedance ($Z_0$) of your system.

The Core Attenuator Formula and Symbol Definitions

The design of a T-pad attenuator relies on a two-step mathematical process. First, you convert the desired decibel attenuation into a linear voltage ratio ($K$). Second, you apply that ratio alongside your system impedance ($Z_0$) to calculate the series and shunt resistor values. Below is the primary attenuator formula set for an unbalanced T-network.

Step 1: Linear Voltage Ratio
K = 10^(dB / 20)

Step 2: T-Pad Resistor Values
Series Arms: R_1 = R_3 = Z_0 * ((K - 1) / (K + 1))
Shunt Arm: R_2 = Z_0 * (2K / (K^2 - 1))

Symbol Definitions for the Attenuator Formula
Symbol Definition Standard Unit
dB Target signal attenuation (expressed as a positive number for loss) Decibels (dB)
K Linear voltage attenuation ratio (V_in / V_out) Unitless (V/V)
Z_0 Characteristic impedance of the source and load Ohms (Ω)
R_1, R_3 Series resistors on the input and output legs of the T-pad Ohms (Ω)
R_2 Shunt resistor connected to ground between the series legs Ohms (Ω)

Rearranged Forms for Reverse Engineering

On the bench, you frequently encounter an unmarked attenuator pad on a salvaged PCB, or you need to verify a commercial pad's specs. By algebraically rearranging the attenuator formula, you can extract the design parameters from physical measurements. According to standard RF design references like RF Cafe, these are the most useful rearranged forms:

  • Solving for K (given R_1 and Z_0):
    K = (Z_0 + R_1) / (Z_0 - R_1)
  • Solving for Z_0 (given R_1 and R_2):
    Z_0 = sqrt(R_1^2 + (2 * R_1 * R_2))
    Note: This is invaluable for identifying the intended system impedance of an unknown pad just by measuring the resistors with a multimeter.
  • Solving for dB (given K):
    dB = 20 * log10(K)
  • Solving for R_2 (given Z_0 and R_1):
    R_2 = (Z_0^2 - R_1^2) / (2 * R_1)

Worked Examples with Unit Tracking

Let's run through two practical design scenarios. Tracking units through the attenuator formula ensures you haven't accidentally dropped a decimal or mixed up power and voltage ratios.

Problem 1: 50 Ω RF System, 10 dB Attenuation

Given: Z_0 = 50 Ω, dB = 10
Find: R_1, R_2, R_3

  1. Calculate K:
    K = 10^(10 / 20) = 10^0.5 ≈ 3.162 (Unitless)
  2. Calculate Series Arms (R_1, R_3):
    R_1 = 50 Ω * ((3.162 - 1) / (3.162 + 1))
    R_1 = 50 Ω * (2.162 / 4.162)
    R_1 = 50 Ω * 0.5194 = 25.97 Ω
  3. Calculate Shunt Arm (R_2):
    R_2 = 50 Ω * ((2 * 3.162) / (3.162^2 - 1))
    R_2 = 50 Ω * (6.324 / (10 - 1))
    R_2 = 50 Ω * (6.324 / 9) = 50 Ω * 0.7027 = 35.14 Ω

Bench Reality Check: You won't find 25.97 Ω in your parts bin. You will use the closest E24 standard value (26 Ω) or E96 value (26.1 Ω). This slight deviation introduces a minor impedance mismatch, degrading your return loss, but it remains perfectly acceptable for most sub-GHz RF work.

Problem 2: 600 Ω Audio Line, 6 dB Attenuation

Given: Z_0 = 600 Ω, dB = 6
Find: R_1, R_2, R_3

  1. Calculate K:
    K = 10^(6 / 20) = 10^0.3 ≈ 1.995 (Unitless)
  2. Calculate Series Arms (R_1, R_3):
    R_1 = 600 Ω * ((1.995 - 1) / (1.995 + 1))
    R_1 = 600 Ω * (0.995 / 2.995)
    R_1 = 600 Ω * 0.3322 = 199.3 Ω
  3. Calculate Shunt Arm (R_2):
    R_2 = 600 Ω * ((2 * 1.995) / (1.995^2 - 1))
    R_2 = 600 Ω * (3.99 / (3.98 - 1))
    R_2 = 600 Ω * (3.99 / 2.98) = 600 Ω * 1.3389 = 803.3 Ω

Bench Reality Check: Use a 200 Ω and an 806 Ω (E24 series) 1% tolerance metal film resistor. For audio, parasitic capacitance is negligible, so standard through-hole 1/4W resistors work perfectly.

When the Formula Applies (and When It Breaks)

Warning: The attenuator formula assumes a purely resistive, matched environment. Applying it blindly to reactive loads or high-frequency microwave circuits without compensation will result in severe signal reflections and a skewed frequency response.

Core Assumptions

The mathematics above assume that both the source and the load are purely resistive and exactly equal to Z_0. If your load is a complex impedance (e.g., an antenna with a reactive component at your operating frequency), the pad will not provide the calculated attenuation, and your VSWR will spike. Furthermore, the formula assumes DC or low-frequency AC where the physical components behave as ideal resistors.

Unit Mistakes That Break the Math

The most common error when using the attenuator formula is confusing power decibels with voltage decibels.

  • The 10 vs 20 Trap: Attenuation in terms of power uses 10 * log10(P_out / P_in). Attenuation in terms of voltage uses 20 * log10(V_out / V_in). Because K in our resistor formula is a voltage ratio, you must use the 20 multiplier. If you use 10, your calculated K will be the square root of what it should be, resulting in wildly incorrect resistor values.
  • Feeding dB directly into the resistor equation: K is a linear ratio (e.g., 3.162), not the dB value itself (10). Plugging '10' directly into the R_1 equation instead of '3.162' will yield a completely wrong pad.
  • Peak-to-Peak vs RMS: As long as you are consistent (both V_in and V_out are RMS, or both are peak-to-peak), the ratio K remains identical. Mixing them breaks the ratio.

Realistic Answer Magnitudes

How do you know if your calculator output makes sense? For a standard 50 Ω system, R_1 and R_3 will always be less than 50 Ω (typically between 5 Ω and 45 Ω). R_2 will typically range from 10 Ω to 300 Ω. If you calculate a series resistor of 450 Ω for a 50 Ω pad, you have made a math error. Conversely, for a 600 Ω audio system, expect series resistors in the 100 Ω to 500 Ω range, and shunt resistors in the 300 Ω to 2 kΩ range. As All About Circuits notes, if the calculated resistance exceeds your system's Z_0 for the series arms, your K calculation is inverted.

Frequently Asked Questions

How does the attenuator formula change for a Pi-pad network?

In a Pi-pad topology, the positions of the series and shunt elements are inverted compared to the T-pad: you have one series resistor and two shunt resistors to ground. The linear ratio K is calculated identically (K = 10^(dB/20)), but the resistor equations change to:
Shunt Arms: R_1 = R_3 = Z_0 * ((K + 1) / (K - 1))
Series Arm: R_2 = Z_0 * ((K^2 - 1) / 2K)
Note that the Pi-pad shunt resistors are mathematically the inverse of the T-pad series resistors.

Why does my calculated attenuator formula result in negative resistor values?

If your math spits out a negative ohm value, you have likely inverted K. The variable K must represent the ratio of the larger voltage to the smaller voltage (V_in / V_out), meaning K must always be greater than 1 for a passive attenuator. If you accidentally calculated V_out / V_in, your K will be a decimal (e.g., 0.316 instead of 3.162). Plugging a K < 1 into the term (K - 1) yields a negative numerator, resulting in impossible negative resistance.

Can I use the attenuator formula for high-frequency RF microwave designs?

The DC/low-frequency formula gives you the baseline values, but above 100 MHz, parasitic effects dominate. A standard 0805 SMD resistor has roughly 0.1 pF to 0.3 pF of parasitic parallel capacitance. At 2.4 GHz, this capacitance creates a low-pass filter effect with your shunt arm, causing the actual attenuation to drop off at higher frequencies. For microwave designs, you must use specialized high-frequency thin-film resistors, minimize pad sizes to reduce stray capacitance, and simulate the layout in an EM solver to compensate for the PCB trace inductance.

What is the difference between the power and voltage attenuator formula?

The physical resistor network doesn't change, but the way you define the decibel target does. Power attenuation uses the 10 * log10 multiplier, while voltage attenuation uses 20 * log10. Because power is proportional to voltage squared (P = V^2 / R), the logarithm brings the exponent down as a 2, which cancels out the 10 to make it 20. In a matched system (where source and load impedances are equal), a 10 dB power attenuation is exactly the same physical pad as a 10 dB voltage attenuation. The confusion only arises if you are measuring voltage across mismatched impedances.