In short
A rectifier converts alternating current (AC) — which swings positive and negative 50 times a second on an Indian power socket — into direct current (DC) that a battery or a chip can actually use. The rectifier is built out of p-n junction diodes, which conduct current only in one direction.
Half-wave rectifier. One diode. It passes the positive half of each AC cycle and blocks the negative half. Output is a train of bumps at 50 Hz, with a DC value V_{DC} = V_m / \pi \approx 0.318 V_m and an RMS value V_{RMS} = V_m / 2. The ripple factor \gamma = \sqrt{(V_{RMS}/V_{DC})^2 - 1} = 1.21 — the ripple amplitude is larger than the useful DC.
Full-wave rectifier. Two diodes with a centre-tapped transformer, or four diodes in a bridge configuration. It flips the negative half of the AC wave upward, producing bumps at 100 Hz. V_{DC} = 2V_m/\pi \approx 0.637 V_m, V_{RMS} = V_m/\sqrt 2, and \gamma = 0.482 — much better.
Capacitor filter. A capacitor in parallel with the load charges up on each bump and slowly discharges through the load during the valley between bumps. This smooths the pulsating DC into nearly-flat DC with a small residual ripple. The ripple voltage
drops inversely with capacitance C and ripple frequency f_r — which is why full-wave rectifiers filter better (f_r = 100 Hz) than half-wave (f_r = 50 Hz) for the same capacitor. The ripple factor after filtering becomes \gamma = 1 / (2\sqrt 3\, f_r R_L C) — the working formula for designing power supplies.
Every charger plugged into an Indian 230-V, 50-Hz outlet — your Mi or Samsung phone brick, the Diwali LED string lights, the 12-V power pack for a Wi-Fi router, the inverter in your home solar rooftop — has a rectifier as its first stage. The later stages (switching regulators, voltage regulators) refine the DC further, but they all start with four diodes in a bridge.
Pull a 230-volt, 50-Hz wall outlet to pieces in your head. The voltage at the live terminal relative to the neutral terminal rises sinusoidally from zero to +325 volts in about 5 milliseconds, swings back through zero to -325 volts over the next 10 milliseconds, and returns to zero — completing one full cycle in 20 ms. Then it does it again. Fifty times every second. This is what "230 V AC" means: 230 is the RMS value of the sine wave, and the peak value is 230 \times \sqrt 2 = 325 V.
Now look at what your phone charger actually produces. Plug a multimeter into the brick's USB output while it's plugged into the wall. It reads a flat, steady 5.0 volts DC. No swinging. No 50-Hz hum. A current flows out of the "+" terminal of the USB connector, through your phone, back in through the "−" terminal, always in the same direction, whenever the phone needs charge.
Somewhere inside the brick — a few cubic centimetres of plastic and copper and silicon — the sinusoidal 325-volt-peak AC has been transformed into a steady 5-volt DC. The transformer inside steps down the voltage. A rectifier flips the negative half of the sine wave into positive, turning AC into a sequence of bumps. A capacitor smooths the bumps into almost-flat DC. A switching regulator — more modern chargers skip the linear regulator for efficiency — chops the DC into high-frequency pulses and filters them back to a precise 5 V.
The first two moves — rectification and capacitor filtering — are the subject of this article. They are where the fundamental physics happens. The subsequent stages are refinements. Every power supply that has ever sat on an Indian worktable, from the old "eliminator" that the rickshaw-wala's radio ran off in the 1980s to the ₹20 USB charger sold at every kirana shop in 2026, has the same first stage: a diode (or four), and a capacitor.
The move — use a diode as a one-way valve
A p-n junction diode conducts essentially freely when the anode (p-side) is at a higher potential than the cathode (n-side) — this is forward bias. It conducts essentially zero current when the anode is at a lower potential than the cathode — reverse bias. Current flows through a diode in one direction only.
Feed a sinusoidal voltage V_{in}(t) = V_m \sin(\omega t) across a diode in series with a load resistor R_L. During the positive half-cycle (when V_{in} > V_D \approx 0.7 V), the diode is forward-biased; current flows, and the load sees a voltage nearly equal to V_{in}. During the negative half-cycle (V_{in} < 0), the diode is reverse-biased; negligible current flows, and the load sees zero.
The output is one-directional — always either zero or positive — but far from smooth. It is a sequence of half-sine bumps separated by gaps. This is what the phrase pulsating DC describes, and it is the first step in converting AC to the flat DC you eventually want.
In the remainder of this article: derive the DC and RMS voltages for both half-wave and full-wave cases, show how the ripple factor quantifies "unsteadiness", and then compute what a capacitor does to reduce it.
Half-wave rectifier — the numbers
Input voltage V_{in}(t) = V_m \sin(\omega t) with \omega = 2\pi \cdot 50 rad/s = 314 rad/s. Assume an ideal diode (zero drop, zero reverse leakage — we'll come back to the 0.7-V silicon drop). The output voltage is
and repeats with period T = 2\pi/\omega = 20 ms.
DC value — the average
The DC value (also called the average value) is the time-average of V_{out} over one full period:
The output is zero for the second half, so only the first half contributes:
Step 1. Evaluate the integral.
Why: \omega(T/2) = \pi, so \cos(\omega T/2) = \cos \pi = -1, and \cos 0 = 1. The integral of sine over half a cycle is 2/\omega.
Step 2. Combine.
Why: \omega = 2\pi/T, so 1/\omega = T/(2\pi). The Ts cancel, and you are left with V_m/\pi.
For a 230-V-AC supply (V_m = 325 V), half-wave DC value is 325/\pi \approx 103 V. A third of the peak.
RMS value
The RMS (root-mean-square) is the square root of the mean of the squared voltage. Over one period:
Step 1. Use \sin^2(\omega t) = [1 - \cos(2\omega t)]/2.
Step 2. Integrate.
Why: the cosine integrates to sine over one full wavelength of itself, which is zero. Over half a period of the original signal, 2\omega t ranges over a full 2\pi, so the cosine averages to zero.
Step 3. Combine.
For a 230-V-AC supply, half-wave V_{RMS} = 325/2 = 162.5 V.
Ripple factor
The ripple factor \gamma quantifies how much of the output is AC (useless, for DC applications) versus DC (useful). It is the ratio of the RMS value of the AC component to the DC value:
The RMS of the AC component is \sqrt{V_{RMS}^2 - V_{DC}^2}, so
Why: any periodic waveform decomposes into a DC average plus a fluctuating AC component. Their mean-squares add: \langle V^2\rangle = V_{DC}^2 + \langle V_{AC}^2\rangle. Solving for the AC RMS gives the expression above.
For the half-wave rectifier:
Why: the ripple RMS is 21% larger than the DC value — the output is mostly ripple. Half-wave rectification on its own is nearly useless for powering a circuit; you need filtering.
Full-wave rectifier — flip the negative half
The waste in half-wave rectification is obvious: half the cycle does nothing. A full-wave rectifier uses the negative half too by flipping it upward. Two circuits achieve this: the centre-tapped two-diode circuit and the four-diode bridge rectifier. Both produce the same output waveform — the absolute value of the sine — but the bridge rectifier is by far the more common, because it does not need a centre-tapped transformer.
During the positive half-cycle, D_1 and D_4 conduct while D_2 and D_3 are reverse-biased. Current flows: AC(top) → D_1 → load → D_4 → AC(bottom). Load voltage is +V_m \sin \omega t.
During the negative half-cycle, D_3 and D_2 conduct while D_1 and D_4 are reverse-biased. Current flows: AC(bottom) → D_3 → load → D_2 → AC(top). The load current is again in the same direction even though the AC polarity has reversed. Load voltage is +V_m |\sin \omega t|.
Result: the output is |V_m \sin \omega t| — the absolute value of the input. A sequence of positive bumps at 100 Hz, twice as fast as the input.
DC and RMS values for full-wave
Both halves contribute, so the DC value is:
Why: exactly twice the half-wave value, because both halves of the cycle now contribute.
The RMS value uses the full period:
Why: |\sin|^2 = \sin^2, so squaring the absolute value gives the same integrand as the original sine squared. The RMS is exactly the RMS of the original AC — which makes sense, because the power delivered to a purely resistive load is the same as if the AC had flowed straight through (no diodes lost any).
Ripple factor for full-wave
Why: 0.482 is roughly a quarter of 1.21, the half-wave value. The full-wave rectifier produces far less ripple. But 48% ripple is still too much for almost any real load — a capacitor filter is what gets you the rest of the way.
Capacitor filter — the smoothing
Place a capacitor C in parallel with the load R_L at the rectifier output. The capacitor acts as a reservoir: it charges up to the peak voltage V_m during each rising portion of the rectifier output, then discharges slowly through the load during the valleys between bumps.
During the short interval when the rectifier output exceeds the capacitor voltage, the diode conducts and recharges the capacitor back to the peak. During the much longer interval when the rectifier output is below the capacitor voltage, the diode is off and the capacitor drives the load alone.
Ripple amplitude. During the discharge, the capacitor loses charge at rate I_{load} = V_{DC}/R_L. In time t, the charge lost is Q = I_{load} \cdot t, so the voltage drops by \Delta V = Q/C = I_{load} \cdot t / C. The discharge lasts for most of the period between bumps — roughly the ripple period T_r = 1/f_r, where f_r = 50 Hz for half-wave and f_r = 100 Hz for full-wave. Taking t \approx T_r:
Why: this is the peak-to-peak ripple amplitude. The RMS ripple is V_r/(2\sqrt 3) because the sawtooth ripple has an RMS equal to peak-to-peak divided by 2\sqrt 3. Dividing by V_{DC} gives the ripple factor after filtering:
— a workable design formula.
Notice that full-wave rectification gives f_r = 100 Hz, half-wave gives f_r = 50 Hz: full-wave halves the ripple for the same capacitor. This is the second reason (after the factor-of-2 in DC output) to use full-wave rectification. Combined, full-wave gets you roughly a factor of 4 better DC for the same component cost.
Worked examples
Example 1: DC and ripple of a 230 V bridge rectifier with filter
An Indian 230-V-RMS, 50-Hz supply drives a step-down transformer with turns ratio 10:1, producing 23-V-RMS AC. This feeds a bridge rectifier followed by a 1000-μF filter capacitor, loaded by R_L = 100 \text{ Ω}. Assume silicon diodes with 0.7 V drop each. Find: (a) peak transformer secondary voltage, (b) peak rectifier output (after diode drops), (c) load current, (d) ripple voltage, (e) DC output voltage including ripple.
Step 1. Peak transformer secondary voltage.
Why: V_{RMS} for a sine is the peak divided by \sqrt 2. Running it backward, peak is RMS times \sqrt 2. The transformer is assumed lossless with a 10:1 turns ratio.
Step 2. Peak rectifier output (after diode drops).
In a bridge rectifier, two diodes are in series at any time (one on each leg of the current path). Each drops 0.7 V, for a total forward drop of 1.4 V.
Why: real silicon diodes have a near-constant 0.7 V forward drop at room temperature. Two conduct in series for each polarity of the bridge, so 1.4 V comes off the peak.
Step 3. Approximate load current (using peak voltage as the DC estimate).
Why: with a properly-sized filter capacitor, the output stays close to the peak. Strictly, I_{load} uses the actual DC voltage, which will be slightly less than the peak by half the ripple — a small correction we'll check in Step 5.
Step 4. Ripple voltage (peak-to-peak).
Why: f_r = 100 Hz for full-wave rectification, C = 1000\,\mu\text{F} = 10^{-3} F. The denominator f_r C = 0.1 F/s. The ripple is the current times the discharge time (1/100 Hz = 10 ms) divided by the capacitance.
Step 5. DC output voltage (averaged over the ripple).
Why: the filtered output swings between V_{peak} at the top of each recharge and V_{peak} - V_r at the bottom. The average (DC value) sits halfway.
Step 6. Ripple factor.
Why: a sawtooth ripple has RMS equal to peak-to-peak / (2√3). The ripple factor of 3% is acceptable for powering a Wi-Fi router or a string of LED lights, but too large for an analog sensor circuit that needs a quiet rail.
Result: The supply delivers about 29.5 V DC at 0.31 A into a 100 Ω load, with 3% ripple. A standard ₹200 Diwali-lantern adapter rolled off a Tamil Nadu assembly line does exactly this, within rounding.
What this shows: Every number in the power-supply design drops out of three formulas — V_{peak} = V_{RMS}\sqrt 2, V_r = I_{load}/(f_r C), \gamma = V_r/(2\sqrt 3 V_{DC}) — plus bookkeeping on diode drops. No device-physics details are needed beyond knowing the diode forward drop, because the circuit topology does the rest.
Example 2: Half-wave vs full-wave ripple comparison
The same 23-V-RMS secondary and 1000-μF filter capacitor drive a 100-Ω load, but this time through a single-diode half-wave rectifier. Find the ripple voltage and compare to the bridge rectifier of Example 1.
Step 1. Peak output voltage (only one diode drop now).
Why: a half-wave rectifier passes current through one diode at a time, not two in series.
Step 2. Load current (approximate).
Step 3. Ripple frequency is 50 Hz (once per AC cycle, not twice).
Why: f_r = 50 Hz for half-wave rectification. The capacitor has to hold the load for twice as long between recharges, so the discharge voltage drop is twice as large.
Step 4. Compare to the bridge.
Why: the factor of 2 comes directly from the factor of 2 in ripple frequency. Ripple scales as 1/f_r, so doubling f_r halves the ripple.
Step 5. DC output.
Step 6. Ripple factor.
Why: twice the ripple factor of the bridge rectifier, as expected. Plus a slightly lower DC voltage because of the larger ripple valley.
Result: The half-wave version gives 28.6 V at 0.32 A with 6.4% ripple — more than double the ripple of the bridge rectifier for the same component cost (minus three diodes). This is the textbook reason why bridge rectifiers dominate all modern power supplies: four diodes is 3 more than a half-wave, and they cost pennies, and they halve the ripple.
What this shows: The ripple formula V_r = I/(f_r C) decides everything. The only way to reduce ripple is to increase the ripple frequency (use full-wave rectification) or increase the capacitance (bigger, more expensive, slower-response capacitor). Modern switching supplies take a third route — they generate a ripple frequency of 100 kHz or more, which lets them use tiny capacitors and shrink into the ₹20 USB brick you buy at a kirana store.
Common confusions
-
"A diode is a one-way wire." Not quite — a diode has a forward drop of about 0.7 V for silicon (0.3 V for germanium, 2-3 V for LEDs). This means the output peak is less than the input peak by the drop. In low-voltage supplies (5 V, 3.3 V), that 0.7 V per diode is a serious efficiency penalty, which is why Schottky diodes (0.3 V drop, very fast switching) are used in modern high-efficiency power supplies and synchronous rectification (using MOSFETs controlled to mimic diodes, with near-zero drop) is used in premium phone chargers.
-
"The capacitor stores DC voltage between bumps." More precisely, the capacitor holds the charge Q = CV, and the load pulls charge off at rate I_{load}. Voltage falls at rate dV/dt = I_{load}/C. Design for 1% ripple means the voltage falls by 1% between recharges. If R_L is small (heavy load), the capacitor drains faster and ripple is larger. If R_L is infinite (no load), the capacitor stays at peak — which is why you can measure an unloaded power supply and see the full peak voltage, but under load it drops.
-
"A rectifier produces DC." Strictly, rectification produces unipolar pulsating voltage — a voltage that does not change sign. "Pure DC" (flat, constant voltage) requires further filtering. The capacitor gets you close; for laboratory-grade DC, a linear regulator (LM7805 and cousins) or a switching regulator follows, trimming the residual ripple to microvolts. Your phone battery needs 4.2 V precisely, not "approximately 5 V with 3% ripple" — the precise regulation happens after rectification and filtering.
-
"Bridge rectifier is better than half-wave because it uses both halves of the cycle." It is true that it uses both halves, but the deeper reason for its better performance is the doubled ripple frequency. Even a centre-tapped full-wave circuit (which also uses both halves) has the same ripple-frequency advantage. The bridge is preferred over the centre-tapped design because it does not require a centre-tapped transformer — but its rectification performance is identical. The "both halves" intuition is right; the "doubled ripple frequency" insight is more useful when you design the filter.
-
"A bigger capacitor always gives a cleaner output." True up to a point, but a very large capacitor pulls huge inrush currents when you first plug in the supply — all the charge must flow in to bring V_C up to the peak, and the diodes and transformer winding see a brief but massive current spike. Real power supplies include a small series resistor or a thermistor (NTC inrush limiter) precisely to bleed that inrush safely. The trade-off between ripple and inrush is one of the fundamental design dials.
-
"The output voltage is the RMS value of the input." No — the unfiltered output is pulsating with DC value V_m/\pi (half-wave) or 2V_m/\pi (full-wave). The filtered output approaches V_m (the peak) when the load current is small. The RMS of the input is V_m/\sqrt 2 — neither of those. Which voltage you see depends on where in the supply chain (rectifier, filter, regulator) you put the voltmeter.
If you came here to understand how a diode converts AC to DC, how the ripple factor quantifies smoothness, and how a capacitor filter works, the above is enough. What follows is for readers who want the exact solution of the filtered output, the physics of the inrush current, and the design of modern switching power supplies.
The exact filtered output — transcendental equation
The approximation V_r = I_{load}/(f_r C) assumed the capacitor discharges for the full ripple period. In reality, the capacitor discharges for a time T_r - t_{charge}, where t_{charge} is the brief interval near each peak when the diode conducts and recharges the capacitor.
During the conduction interval, the capacitor voltage equals the rectified input: V_C(t) = V_m |\sin(\omega t)|. During the discharge interval, V_C(t) = V_1 \exp(-(t - t_1)/\tau) where \tau = R_L C and V_1 is the voltage at the start of discharge.
The conduction interval starts when the rectified input catches up to the decaying capacitor voltage. This is a transcendental equation:
For large \tau f_r (the well-filtered case), t_1 is near the peak (where \sin = 1) and t_2 is just after the next peak. Linearising gives
— matching the naive formula we used. For \tau f_r < 10 (lightly filtered), the linearisation breaks and you need to solve the transcendental equation numerically. Most real power supplies are in the well-filtered regime, so the simple formula is good to a few percent.
Peak inverse voltage and diode selection
When a diode is reverse-biased in a rectifier, the voltage across it — the peak inverse voltage (PIV) — is set by the circuit topology.
- Half-wave rectifier with capacitor filter: during the negative half-cycle, the capacitor holds +V_m at the cathode while the source is at -V_m at the anode. PIV = 2 V_m.
- Full-wave centre-tapped: the off-duty diode sees the full secondary voltage plus the filtered output — PIV = 2 V_m.
- Bridge rectifier: each off-duty diode sees one secondary voltage drop — PIV = V_m.
For a 230-V RMS supply with filter, the bridge rectifier needs diodes rated for 325 V PIV. The 1N4007 (1000 V PIV, 1 A) is the universal workhorse. For the half-wave or centre-tapped circuits, you need diodes rated for twice that. This is another reason bridges dominate — they put cheaper diodes in a workable configuration.
Inrush current and its limiting
When the power supply is first plugged in, the capacitor voltage is zero. On the first peak, current flows at I_{in} = (V_m - 0)/R_{ESR}, where R_{ESR} is the total series resistance of the transformer winding, the diodes (about 0.1 Ω each during heavy conduction), and the capacitor's equivalent series resistance. For a small transformer, R_{ESR} \sim 1 \text{ Ω}, so I_{in} \sim 325/1 = 325 A for an unprotected bridge into a 1000-μF capacitor.
That spike lasts only a few milliseconds (the time for the capacitor to charge through the resistance), but it is enough to blow diodes, weld transformer windings, or trip circuit breakers. Three remedies:
- Inrush thermistor (NTC): a resistor with negative temperature coefficient. Cold, it is ~5 Ω; as current flows, it heats up and drops to ~0.1 Ω in seconds. You see these in every desktop computer power supply.
- Active inrush limiter: a MOSFET that starts open and closes gradually under microcontroller control.
- Soft-start filter capacitor: a smaller capacitor in series with a relay that shorts it out after a few cycles. Common in industrial drives.
Switching-mode power supplies — where modern chargers actually live
The rectifier + capacitor + linear regulator topology described above produces steady DC, but wastes most of the power as heat in the regulator. A 5-V phone charger fed from 325-V-peak mains would waste 320/325 = 98\% of the input power as heat across the regulator. In the 1970s this was acceptable. Today, with a ₹20 charger powering a 5-W phone at 85% efficiency, it is not.
Modern chargers use switching-mode power supplies (SMPS). The topology is:
- Rectify the mains directly with a high-voltage bridge (325 V DC across a 470 μF capacitor).
- Chop the DC with a MOSFET switch at 100 kHz.
- Pass the chopped waveform through a tiny ferrite-cored transformer (the high frequency makes small cores viable).
- Rectify the transformer secondary (low-voltage, high-current — use Schottky diodes or synchronous MOSFETs).
- Filter with a small capacitor (100 μF, because 100 kHz ripple filters with tiny caps).
- Feedback-regulate the output voltage by adjusting the duty cycle of the MOSFET switch.
The SMPS inherits the rectifier-capacitor physics described in this article — it is used twice, once at the mains input and once at the transformer secondary — but the key insight is that 100 kHz ripple filters with a 100-μF capacitor the same way 100 Hz ripple filters with a 100-mF capacitor. Miniaturisation follows.
ISRO's launch-vehicle power systems use similar SMPS architectures built from radiation-hardened components — silicon carbide MOSFETs, wide-bandgap rectifiers — to pack 10-kW power conversion into the confined avionics bay of a PSLV. Samsung's mobile chargers, manufactured at Sriperumbudur, use gallium nitride switches operating at 1 MHz to achieve 100-W-in-a-thumb-sized-brick. The silicon physics is the same; the topology is the same; only the switching frequency moves.
Ripple factor from Fourier series
For a rigorous derivation, Fourier-decompose the full-wave rectifier output |\sin \omega t|:
The DC component is 2/\pi (matching our V_{DC} = 2V_m/\pi). The fundamental ripple is at 2\omega (i.e., 100 Hz) with amplitude 4/(3\pi). Higher harmonics decrease as 1/n^2.
The AC component's total RMS:
This converges slowly but to a known value, and you can verify V_{AC,RMS}^2 + V_{DC}^2 = V_{RMS}^2 = V_m^2/2, recovering the 0.482 ripple factor numerically.
Filtering is then trivial: each Fourier component passes through the filter 1/(j\omega RC)-style, so the 100-Hz component is reduced by 1/(2\pi \cdot 100 \cdot R_L C), and higher harmonics are reduced even further. The total residual ripple is dominated by the fundamental 100-Hz component, which is what our simple V_r = I/(f_r C) captures.
This is the standard Fourier argument in signal processing: a rectifier is a nonlinearity that generates harmonics; the filter is a low-pass that blocks them. Design the filter's cutoff frequency well below f_r, and the DC survives while the harmonics die.
Where this leads next
- The p-n Junction Diode — the device physics of the diodes that make rectification possible. The built-in potential, the depletion region, the I-V characteristic, and why the forward drop is approximately 0.7 V for silicon.
- Semiconductors — Intrinsic and Extrinsic — the crystal-level physics of n-type and p-type silicon, without which no diode, no transistor, no chip exists.
- AC Generation and RMS Values — the sinusoidal source that a rectifier converts. The RMS convention and the peak-to-RMS conversion used throughout this article.
- Capacitance and the Parallel-Plate Capacitor — the physics of the filter capacitor. Charge stored, time constant RC, and why a bigger capacitor stores more charge per volt.
- AC Through R, L and C — impedance in reactive components, and the Fourier picture of how a filter blocks AC while passing DC.