Rectifiers: Half-Wave & Full-Wave

Rectifiers are the quiet workhorses of the electrical age: essentially every device that plugs into a wall and every one that charges from USB contains one. The four-diode bridge in a phone charger, laptop brick or LED bulb driver converts mains AC to DC before a switching regulator trims it to the exact voltage a battery or chip needs. On the grid's largest scale, banks of high-power rectifiers form the sending end of HVDC (high-voltage direct-current) transmission lines that move gigawatts between regions and across undersea cables more efficiently than AC could. Electroplating tanks, aluminium smelters and electric-railway substations run on industrial rectifiers delivering thousands of amps. And Schottky's own low-drop, fast diode is everywhere power efficiency matters — in the switching supplies inside computers, in solar-panel charge controllers, and in radio-frequency detectors that reach right back to Braun's galena crystal. The next article turns from diodes that merely conduct to diodes with a job — the LED that emits light, the photodiode and solar cell that drink it.

Start with the raw material: a sinusoidal voltage $v = V_m\sin\theta$, where $\theta = 2\pi f t$ marches steadily forward and the voltage swings from $+V_m$ down to $-V_m$ and back, $f$ times a second. In India and Europe $f = 50\ \text{Hz}$; in North America $60\ \text{Hz}$. A resistor fed this voltage carries a current that swings both ways equally, so its average is exactly zero — useless as a supply. The rectifier's job is to break that symmetry. Drop a single diode in series with the load. On every half-cycle where the source drives the diode's p-side positive, the diode is forward-biased: once the swing passes the knee voltage it conducts, and current flows through the load in one direction. On the opposite half-cycle the source reverses, the diode is reverse-biased, and it blocks — only a whisper of leakage gets through. The load therefore sees current during one half of every cycle and nothing during the other. The negative humps have simply been erased. What remains is a train of one-way bumps: pulsating DC — direct, because it never reverses, but far from steady.

The crude output above is genuine DC in the sense that matters — its current never reverses — but it is a long way from the flat line a circuit wants. Two problems are visible at once. First, for half of every cycle the load gets nothing at all, so the average current is low. Second, even during the humps the current climbs and falls all the way from zero to $I_m$ and back, so the output is drenched in ripple — an unwanted AC wobble riding on the DC. The rest of this article is a campaign against both problems: fold the wasted half back to raise the average (full-wave), then park a capacitor across the load to fill the valleys (filtering).

With a single diode the current through the load is a sine hump for the conducting half-cycle and flat zero for the other half. To describe such a lumpy waveform with one number, we ask for its average — the steady DC current that would deliver the same charge per second. The workhorse fact is that the average height of a single sine hump is $2/\pi$ of its peak (a standard result quoted below in the derivation). Because the half-wave output is a hump followed by an equal stretch of zero, its average over the whole cycle is half of that:

where: $I_m = V_m/(R_f + R_L)$ is the peak current (source peak voltage divided by the diode's forward resistance plus the load), $I_{dc}$ is the average or DC current the load actually receives, and $\pi \approx 3.1416$. The physical reading: the diode is "on" only half the time and even then delivers a hump whose mean is $2/\pi$ of the peak, so the steady equivalent is barely a third of the peak current. Correspondingly the DC output voltage is $V_{dc} = I_{dc}R_L = V_m/\pi$ for an ideal diode. The other number we need is the root-mean-square current, which sets the heating and the total power. Averaging the square of the waveform (the mean of $\sin^2$ over the hump is $\tfrac12$, and again the diode is on only half the time) gives

where: $I_{rms}$ is the effective (heating) current. Notice it exceeds $I_{dc}$: the waveform carries more "wobble" than "steady," which is exactly the ripple problem. One more feature matters for the smoothing stage: the half-wave output delivers one hump per input cycle, so its fundamental ripple frequency equals the mains frequency $f$ — $50\ \text{Hz}$ on a $50\ \text{Hz}$ supply.

The half-wave rectifier's great sin is that it discards a perfectly good half of the input. The full-wave rectifier keeps it — not by passing the negative half through unchanged (that would just reverse the current) but by flipping it up so both halves push current through the load the same way. There are two classic ways to arrange this. The centre-tapped design uses a transformer with a grounded centre tap and two diodes. On one half-cycle the top half of the winding forward-biases the upper diode; on the next half-cycle the bottom half forward-biases the lower diode; either way current runs through the load in the same direction. It is elegant but needs a special transformer, and each diode must withstand a peak inverse voltage of $2V_m$ (the whole winding appears across the off diode). The bridge design — the diamond of four diodes in your charger — needs no centre tap. On each half-cycle two diagonally opposite diodes conduct and steer whichever polarity arrives down through the load the same way; the other two block. It costs two extra diodes but uses an ordinary transformer and each diode only sees a peak inverse voltage of $V_m$, which is why the bridge is the standard power-supply front end today.

Whichever arrangement you use, the load now sees a sine hump every half-cycle, back to back with no gaps. Averaging that string of humps — each with mean height $2/\pi$ of the peak, and now on all the time — gives twice the DC of the half-wave case:

where: the symbols are as before, and the factor $2/\pi \approx 0.637$ is simply the mean of one sine hump — the same $0.637$ that also happens to be twice the half-wave coefficient. The DC output voltage is $V_{dc} = 2V_m/\pi$ for ideal diodes. The rms current now equals that of a plain sine, because folding the negative humps up does not change any squared value:

where: $I_{rms}$ is again the heating-equivalent current. And because there are now two humps per input cycle, the fundamental ripple frequency is $2f$ — $100\ \text{Hz}$ on a $50\ \text{Hz}$ supply. That doubling is a quiet gift: a higher ripple frequency is easier for a capacitor to smooth, which is the last piece of the story.

Two figures of merit decide which rectifier you reach for. The first is the ripple factor $r$ — the ratio of the leftover AC wobble to the useful DC. Since the total rms splits into a DC part and an AC part by $I_{rms}^2 = I_{dc}^2 + I_{ac}^2$, the ripple factor is

where: $I_{ac} = \sqrt{I_{rms}^2 - I_{dc}^2}$ is the rms of the wobble alone. Plug in the half-wave numbers ($I_{rms}/I_{dc} = (I_m/2)/(I_m/\pi) = \pi/2 = 1.571$): $r = \sqrt{(\pi/2)^2 - 1} = \sqrt{1.467} = 1.21$. The AC content actually exceeds the DC — a lousy supply. For full-wave ($I_{rms}/I_{dc} = (I_m/\sqrt2)/(2I_m/\pi) = \pi/(2\sqrt2) = 1.111$): $r = \sqrt{(1.111)^2 - 1} = \sqrt{0.234} = 0.48$. Less than half the wobble, before any filtering at all. The second figure of merit is the rectification efficiency $\eta$ — the fraction of the AC input power that comes out as useful DC power. For ideal diodes it is $\eta = (I_{dc}/I_{rms})^2$, which works out to

where: $\eta$ compares DC output power $I_{dc}^2R_L$ with the AC input power $I_{rms}^2R_L$; the maximum full-wave value $81.2\%$ is exactly twice the half-wave $40.6\%$. (Real diodes drop the figure below these ceilings by the factor $R_L/(R_L + R_f)$.) Between them, equations (5) and (6) are the whole case for full-wave: half the ripple, double the efficiency, for the price of one or three extra diodes. Even $0.48$ of ripple is far too much for a phone or a radio. The cure is a filter — most simply a large capacitor in parallel with the load. The capacitor charges up to near the peak voltage during each hump, then holds the output up by discharging slowly through the load during the valleys, so the output sags only a little before the next peak tops it up. For a capacitor-filtered full-wave supply the ripple factor collapses to a quoted engineering result,

where: $f_r$ is the ripple frequency ($2f$ for full-wave), $C$ the filter capacitance and $R_L$ the load — so a bigger capacitor, a bigger load resistance, or the higher $2f$ ripple frequency all press the ripple down. This is why full-wave's $100\ \text{Hz}$ ripple is doubly welcome: it starts smaller and filters better. A modest $100\ \mu\text{F}$ across a kilohm load can knock the ripple from $48\%$ to a few percent — the worked examples make that concrete.

Board level — a half-wave rectifier's DC and rms currents (CBSE and ICSE pattern)

A sinusoidal source of peak voltage $V_m = 10\ \text{V}$ and frequency $50\ \text{Hz}$ drives a half-wave rectifier feeding a load $R_L = 100\ \Omega$. Treat the diode as ideal. Find the peak current, the DC (average) current, the DC output voltage, the rms current, the ripple factor and the frequency of the output.

Step 1. Peak current: with an ideal diode $I_m = V_m/R_L = 10/100 = 0.1\ \text{A} = \boxed{100\ \text{mA}}$.

Step 2. DC current from equation (1): $I_{dc} = I_m/\pi = 100/\pi = \boxed{31.8\ \text{mA}}$.

Step 3. DC output voltage: $V_{dc} = I_{dc}R_L = 31.8\ \text{mA} \times 100\ \Omega = 3.18\ \text{V}$ (equivalently $V_m/\pi = 10/\pi = \boxed{3.18\ \text{V}}$).

Step 4. rms current from equation (2): $I_{rms} = I_m/2 = \boxed{50\ \text{mA}}$. Ripple factor: $r = \sqrt{(50/31.8)^2 - 1} = \boxed{1.21}$. Output frequency: a half-wave rectifier gives one hump per input cycle, so the ripple frequency equals the input, $\boxed{50\ \text{Hz}}$.

Board level — a full-wave bridge with an ideal transformer (CBSE and A-Level pattern)

A bridge rectifier is fed from a transformer whose secondary has peak voltage $V_m = 12\ \text{V}$ at $50\ \text{Hz}$, into a load $R_L = 120\ \Omega$ (ideal diodes). Find the peak current, the DC current, the DC output voltage, the ripple frequency and the ideal efficiency.

Step 1. Peak current: $I_m = V_m/R_L = 12/120 = 0.1\ \text{A} = \boxed{100\ \text{mA}}$.

Step 2. DC current from equation (3): $I_{dc} = 2I_m/\pi = 2 \times 100/\pi = \boxed{63.7\ \text{mA}}$ — exactly double what a half-wave rectifier would give from the same peak.

Step 3. DC output voltage: $V_{dc} = I_{dc}R_L = 63.7\ \text{mA} \times 120\ \Omega = 7.64\ \text{V}$ ($= 2V_m/\pi = 24/\pi = \boxed{7.64\ \text{V}}$).

Step 4. Ripple frequency: two humps per input cycle gives $2f = \boxed{100\ \text{Hz}}$. Ideal efficiency from equation (6): $\eta = (2\sqrt2/\pi)^2 = \boxed{81.2\%}$. Note that a bridge diode need only withstand a peak inverse voltage of $V_m = 12\ \text{V}$, whereas a centre-tapped design would demand $2V_m = 24\ \text{V}$.

Competitive level — efficiency of a real centre-tapped full-wave rectifier (JEE Main and JEE Advanced pattern)

A centre-tapped full-wave rectifier uses diodes of forward resistance $R_f = 25\ \Omega$ each into a load $R_L = 1000\ \Omega$, driven so that each half of the secondary has peak voltage $V_m = 100\ \text{V}$. Find the peak current, the DC current, the DC output voltage and the actual rectification efficiency.

Step 1. Peak current with the diode resistance included: $I_m = V_m/(R_f + R_L) = 100/1025 = 0.09756\ \text{A} = \boxed{97.6\ \text{mA}}$.

Step 2. DC current: $I_{dc} = 2I_m/\pi = 2 \times 97.56/\pi = \boxed{62.1\ \text{mA}}$.

Step 3. DC output voltage across the load: $V_{dc} = I_{dc}R_L = 62.1\ \text{mA} \times 1000\ \Omega = \boxed{62.1\ \text{V}}$.

Step 4. Efficiency is the ideal ceiling scaled by the resistance-division factor: $\eta = \dfrac{8}{\pi^2}\cdot\dfrac{R_L}{R_L + R_f} = 0.811 \times \dfrac{1000}{1025} = 0.811 \times 0.976 = \boxed{79.1\%}$ — close to the $81.2\%$ ceiling because the diode resistance is small next to the load. Timing note: about ninety seconds. The one trap that eats time is forgetting $R_f$ in Step 1 — it belongs in the peak-current denominator but the DC output voltage uses only $R_L$. Carry $2/\pi = 0.637$ and $8/\pi^2 = 0.811$ as memorised constants and the whole item is two divisions and a multiply.

Competitive level — how much a capacitor filter cuts the ripple (JEE Advanced pattern)

The full-wave supply above ($50\ \text{Hz}$ mains, so ripple frequency $100\ \text{Hz}$) is filtered by a capacitor $C = 100\ \mu\text{F}$ across a load $R_L = 1000\ \Omega$. Using the standard filtered-ripple formula $r \approx 1/(2\sqrt3\,f_r C R_L)$, find the ripple factor and compare it with the unfiltered value.

Step 1. Identify the ripple frequency: for a full-wave rectifier on $50\ \text{Hz}$ mains, $f_r = 2f = 100\ \text{Hz}$.

Step 2. Substitute: $r = \dfrac{1}{2\sqrt3 \times 100 \times (100\times10^{-6}) \times 1000}$. The denominator is $2\times1.732\times100\times0.1 = 34.6$.

Step 3. So $r = 1/34.6 = 0.0289 \approx \boxed{2.9\%}$.

Step 4. Compare: the unfiltered full-wave ripple was $0.48$, i.e. $48\%$. The capacitor has cut the ripple by a factor of about $0.48/0.0289 \approx 17$, from $48\%$ down to under $3\%$ — turning visibly lumpy DC into a supply flat enough for most electronics. Timing note: under a minute once you recognise $f_r = 2f = 100\ \text{Hz}$ for full-wave (a favourite trap — using $50\ \text{Hz}$ here doubles your ripple answer). Keep $2\sqrt3 = 3.46$ memorised and the arithmetic is a single reciprocal.

Calling the rectified output "AC" or "pure DC." It is neither — it is pulsating DC: one-directional but rippled. Say "pulsating DC" or "DC with ripple." Writing "the output is AC" denies the diode's whole purpose; writing "the output is steady DC" forgets the filter.

Using the wrong output frequency. Half-wave output ripples at the mains frequency $f$; full-wave ripples at $2f$. On a $50\ \text{Hz}$ supply that is $50\ \text{Hz}$ versus $100\ \text{Hz}$. Every capacitor-filter question hinges on this — put $2f$ into equation (7) for full-wave or your ripple comes out double.

Confusing the two full-wave layouts and their peak inverse voltage. A centre-tapped rectifier needs two diodes each rated for $\text{PIV} = 2V_m$; a bridge needs four diodes each rated for only $V_m$. Mixing up which needs the higher-voltage diodes is a stock JEE distractor.

Dropping the diode's forward resistance where it matters. In real-diode problems $R_f$ belongs in the peak-current denominator $I_m = V_m/(R_f+R_L)$ and in the efficiency factor $R_L/(R_L+R_f)$ — but the DC output voltage across the load is $I_{dc}R_L$. Sprinkling $R_f$ everywhere (or nowhere) loses marks.

Quoting $I_{dc} = 0.318\,I_m$ for a full-wave rectifier. That coefficient is the half-wave value; full-wave is $0.637\,I_m$. A fast sanity check: full-wave must give more DC than half-wave, never less.

Board scope. Rectifiers — half-wave, full-wave, bridge, ripple, efficiency and the capacitor filter — are core on CBSE, ICSE (ISC), NEET, JEE Main and JEE Advanced, and appear in the electronics content of older A-Level (CIE 9702) syllabuses, though current 9702 no longer carries the rectifier calculations. They are not part of AP Physics 2, which carries no semiconductor electronics at all, so an AP-2 student meets none of this. IGCSE 0625 teaches the diode and its one-way action qualitatively but does not require the $I_m/\pi$, ripple-factor or efficiency calculations — give an IGCSE answer in words. IB Physics has moved most electronics out of its core, so do not rely on it appearing there.

Texts. NCERT Physics Class 12, Chapter 14 ("Semiconductor Electronics"), the rectifier and filter sections · H. C. Verma, Concepts of Physics Vol. 2, "Semiconductors and Semiconductor Devices" · Boylestad and Nashelsky, Electronic Devices and Circuit Theory, for the rectifier and filter analysis in engineering depth · Sedra and Smith, Microelectronic Circuits, for the power-supply front end.

Provenance. Standard rectifier relations. Recompute: mean of a sine hump $2/\pi = 0.637$; half-wave $I_{dc} = I_m/\pi = 0.318\,I_m$, $I_{rms} = I_m/2 = 0.5\,I_m$; full-wave $I_{dc} = 2I_m/\pi = 0.637\,I_m$, $I_{rms} = I_m/\sqrt2 = 0.707\,I_m$; ripple $r = \sqrt{(I_{rms}/I_{dc})^2-1}$ giving $\pi/2 = 1.571 \to r = 1.21$ (half) and $\pi/(2\sqrt2) = 1.111 \to r = 0.48$ (full); efficiencies $(2/\pi)^2 = 40.6\%$ and $(2\sqrt2/\pi)^2 = 81.2\%$, ratio exactly $2$; output frequency $f$ (half) and $2f$ (full), i.e. $50$ and $100\ \text{Hz}$; PIV $V_m$ (bridge/half) and $2V_m$ (centre-tap). Worked values: half-wave $V_m=10\ \text{V}$, $R_L=100\ \Omega \Rightarrow I_m=100\ \text{mA}$, $I_{dc}=31.8\ \text{mA}$, $V_{dc}=3.18\ \text{V}$, $I_{rms}=50\ \text{mA}$, $r=1.21$; bridge $V_m=12\ \text{V}$, $R_L=120\ \Omega \Rightarrow I_m=100\ \text{mA}$, $I_{dc}=63.7\ \text{mA}$, $V_{dc}=7.64\ \text{V}$, $\eta=81.2\%$; real centre-tap $V_m=100\ \text{V}$, $R_f=25\ \Omega$, $R_L=1000\ \Omega \Rightarrow I_m=97.6\ \text{mA}$, $I_{dc}=62.1\ \text{mA}$, $V_{dc}=62.1\ \text{V}$, $\eta=79.1\%$; capacitor filter $C=100\ \mu\text{F}$, $R_L=1000\ \Omega$, $f_r=100\ \text{Hz} \Rightarrow r = 1/(2\sqrt3 f_r C R_L) = 0.029$, a $\sim\!17\times$ cut from $0.48$. Diode knee Si $0.7\ \text{V}$, Ge $0.3\ \text{V}$. Schottky-diode forward drop $\sim\!0.2$–$0.3\ \text{V}$.

Prerequisites. The Junction Diode and its I–V Curve — the one-way valve action (conducts past the knee, blocks in reverse) that every rectifier here is built on. Its knee voltages and reverse-breakdown limit are exactly what set a rectifier's output offset and PIV rating. Leads to. Special Diodes: LED, Photodiode and Solar Cell, where the diode stops merely conducting and starts emitting or absorbing light. For the metal–semiconductor Schottky barrier and fast rectification in real depth, the university seams on semiconductor diode physics and photodetectors treat the junction quantitatively (go-deeper only, never an exam dependency here).