A fullwave rectifier converts both the positive and negative halves of an AC sine wave into pulsating DC. Compared to a half-wave design, it yields double the ripple frequency (120Hz on a 60Hz grid) and roughly 40% higher average DC output voltage, drastically reducing the filter capacitance required for a clean rail. While offline switch-mode power supplies (SMPS) dominate high-wattage applications today, the fullwave rectifier remains the mandatory front-end for ultra-low-noise linear power supplies used in precision ADCs, audio DACs, and RF receiver chains.
Fullwave Rectifier Topologies: Bridge vs. Center-Tapped
There are two primary ways to achieve fullwave rectification: the four-diode bridge and the two-diode center-tapped configuration. The choice dictates your transformer cost, diode voltage stress, and conduction losses.
The bridge rectifier uses four diodes and a standard two-wire transformer secondary. Current flows through two diodes in series during every half-cycle, resulting in a higher forward voltage drop (typically 1.4V to 2.0V depending on current and silicon type). However, the Peak Inverse Voltage (PIV) across each non-conducting diode is only equal to the peak secondary voltage ($V_m$).
The center-tapped (CT) rectifier uses two diodes and a three-wire transformer. Current flows through only one diode at a time, cutting the conduction voltage drop in half (0.7V to 1.0V). The trade-off is that the non-conducting diode must withstand a PIV of $2V_m$, requiring higher-voltage-rated diodes. Furthermore, the transformer utilization factor (TUF) is poor because only half the secondary winding conducts at any given time, forcing you to buy a physically larger, more expensive transformer for the same DC output power.
| Parameter | Bridge Rectifier (4x 1N5408) | Center-Tapped (2x 1N5408 + CT Transformer) |
|---|---|---|
| Diode Count | 4 | 2 |
| Peak Inverse Voltage (PIV) | 21.2V ($V_m$) | 42.4V ($2V_m$) |
| Forward Voltage Drop ($V_f$) | ~2.0V (2 diodes in series) | ~1.0V (1 diode in series) |
| Conduction Loss (Heat) | 4.0W | 2.0W |
| Transformer Utilization Factor | 0.812 (High) | 0.693 (Low) |
| Relative Component Cost | $3.50 (Standard transformer + bridge) | $8.00+ (Custom CT transformer) |
Source: All About Circuits - Full-Wave Rectifier Circuit
For bench and DIY builds under 50W, the bridge rectifier wins on total BOM cost and physical footprint. The 2W extra heat dissipated in the bridge is easily managed by selecting an over-specced square bridge package (like the KBPC3510) that acts as its own heatsink.
Linear vs. Switching: When to Use a Fullwave Front-End
Why build a bulky linear supply with a fullwave rectifier in 2026 when a $2 GaN-based offline switching module exists? The answer is the noise floor.
A switch-mode power supply operates by chopping DC at frequencies between 50kHz and 2MHz. Even with extensive LC filtering and Faraday shields, high-frequency common-mode noise inevitably couples into sensitive analog stages. If your load is a 24-bit audio DAC, a precision strain-gauge amplifier, or an RF superheterodyne receiver, that switching noise will manifest as heterodyne whines, elevated noise floors, or spurious ADC counts.
A fullwave rectifier feeding a bulk capacitor and a linear regulator produces a DC rail with only 120Hz (or 100Hz) ripple. This low-frequency ripple is easily rejected by the Power Supply Rejection Ratio (PSRR) of a linear regulator and falls well outside the audible and RF bands of interest.
Choose Fullwave + Linear when: Load is <10W, absolute minimum EMI is required, and the application involves precision analog, audio, or RF.
Choose Fullwave + Switching Buck when: Load is >10W, efficiency and thermal management are primary concerns, and the load is digital (microcontrollers, motors, relays).
The expected ripple on the bulk capacitor of a fullwave linear supply is calculated using the standard approximation:
$V_{ripple(p-p)} = \frac{I_{load}}{f \times C}$
Where $f$ is the ripple frequency (120Hz for 60Hz mains). For a 2A load and a 4700µF capacitor, the peak-to-peak ripple is $2 / (120 \times 0.0047) = 3.54V$. This dictates the headroom your linear regulator must have to prevent dropout.
12V / 2A Design Example: Input Protection, Headroom Math, and Derating
Let’s design a 12V DC, 2A linear power supply using a fullwave bridge topology. We will calculate the exact component values, regulator headroom, and thermal requirements.
1. Transformer and Input Protection
We need 12V DC at 2A. To ensure the linear regulator doesn't drop out during the ripple valley, we select a 15VAC RMS, 3A transformer (e.g., Triad Magnetics F-456P). The extra 3VAC provides the necessary headroom.
Input Protection: Never connect a transformer directly to the mains without protection. Install a 3A 250V slow-blow fuse in series with the primary to handle the inrush current of the transformer magnetizing and the bulk capacitors charging. Place a 130VAC Metal Oxide Varistor (MOV) (e.g., Bourns MOV-14D201K) across the primary lines, downstream of the fuse, to clamp transient grid spikes.
2. Rectification and Filtering
Instead of four discrete diodes, use a KBPC3510 bridge rectifier module. It is rated for 35A and 1000V. While this is massive overkill for a 2A load, the KBPC3510 costs roughly $2.00, features an integrated metal case with a mounting hole, and will run barely warm at 2A without an external heatsink, eliminating a failure point.
Peak DC Voltage: $15V_{RMS} \times 1.414 = 21.21V$.
Bridge Drop: ~2.0V at 2A.
Peak Capacitor Voltage: $21.21V - 2.0V = 19.21V$.
To keep the ripple valley above the regulator's dropout threshold, we need a bulk capacitance that limits ripple to roughly 3V. Using $C = I / (f \times V_r)$, we get $C = 2 / (120 \times 3) = 5555\mu F$. We will use two 4700µF 35V electrolytic capacitors in parallel (total 9400µF). This yields a real-world ripple of about 1.77V, dropping the valley voltage to $19.21V - 1.77V = 17.44V$.
3. Regulator Selection and Headroom Math
A standard LM317 is limited to 1.5A and requires up to 3V of dropout. For a 2A load, we need a higher-current Low Dropout (LDO) regulator. The LM1084-12 (fixed 12V, 5A) is ideal. According to the LDO dropout specifications, the LM1084 has a maximum dropout voltage of 1.5V at 5A (typically 1.2V at 2A).
Headroom Check:
Valley Input Voltage = 17.44V.
Required Output = 12.0V.
Available Headroom = $17.44V - 12.0V = 5.44V$.
Since 5.44V is well above the 1.5V maximum dropout requirement, the regulator will maintain a clean 12V rail even at the bottom of the 120Hz ripple cycle.
4. Thermal Derating and Heatsink Sizing
Linear regulators dissipate excess voltage as heat. We must calculate the worst-case continuous power dissipation to size the heatsink.
Average Input Voltage: Peak (19.21V) - (Ripple / 2) = $19.21V - 0.88V = 18.33V$.
Power Dissipated ($P_d$): $(V_{in\_avg} - V_{out}) \times I_{load} = (18.33V - 12.0V) \times 2A = 12.66W$.
The LM1084 in a TO-220 package has a junction-to-ambient thermal resistance ($\theta_{JA}$) of roughly 50°C/W without a heatsink. Dissipating 12.66W bare would raise the junction temperature by 633°C, instantly triggering thermal shutdown and potentially destroying the silicon. You must use a heatsink.
To keep the silicon junction below 110°C (giving a 15°C safety margin below the 125°C absolute max) in a 25°C ambient room:
Max Allowable $\theta_{JA}$ = $(110°C - 25°C) / 12.66W = 6.71°C/W$.
The TO-220 package has a junction-to-case ($\theta_{JC}$) resistance of ~3°C/W. Accounting for a silicone thermal pad ($\theta_{CS}$ ~0.5°C/W), the heatsink itself ($\theta_{SA}$) must be rated for:
$6.71 - 3.0 - 0.5 = 3.21°C/W or lower$.
Select an extruded aluminum heatsink like the Aavid Thermalloy 530602B02500G (rated ~2.5°C/W with natural convection). Apply a thin layer of thermal compound, torque the LM1084 to the heatsink to 0.5 Nm, and verify the case temperature with an IR thermometer after 30 minutes of full-load burn-in. If the case exceeds 85°C, add forced air or step up to a larger finned profile.
