Associate CET · Competency 10.0

Power Supplies

Competency 10.0 is the power supply: how equipment turns the mains or a battery into the steady DC its circuits need, and how to service a supply without getting hurt. This page follows ETA's list in order of the signal path: the hazards, transformer and line-connected supplies, rectifiers and their DC voltages, filters, regulation, switch-mode and multiple-output supplies, the integrated high-voltage transformer, batteries and charging, and fuses. Section 12 is job practice. Try the numbers in the Power Supply calculator, the Rectifier Lab and the Battery tool.

Competency
10.0
Sections
12 + self-check
Then
Drill comp 10

Shock hazards in power supplies

Competency 10.1 asks for the shock hazards of servicing power supplies. The supply is where the dangerous energy in most equipment lives, and it stays dangerous after the power switch is off. Treat each of these as live until you have measured it:

HazardWhy it is there
Line voltage on the primary sideEverything ahead of the power transformer is at mains potential: the cord, fuse and switch, and in a switch-mode supply the input rectifier, the bulk capacitor and the switching transistor. Unplug the unit before you open it.
Stored charge in filter capacitorsA filter capacitor keeps its charge after the plug is pulled. ETA's sheet gives its stored energy as W = ½ × C × V^2: a 400 µF capacitor at 200 V holds 8 J. A bleeder resistor is meant to drain it, but never rely on one: measure. OSHA's lockout rule for fixed (hard-wired) equipment requires stored electric energy that might endanger people to be released, with capacitors discharged (29 CFR 1910.333(b)(2)(ii)(C)); do the same on a cord-and-plug unit on the bench.
High voltageA switch-mode supply's bulk capacitor charges to about 170 V on a 120 V line (120 × 1.414), and to about 300–340 V where the input runs from 240 V or through a voltage doubler. A CRT set's high-voltage supply runs at 20–30 kV, and up to about 35 kV (section 9).
A live chassisIn a line-connected supply the circuit common, and often the metal chassis, is tied to the mains (section 2). Touching it while you are grounded completes a path through you.

Safety

Before you touch: unplug the unit, discharge the filter capacitors through a resistor, and confirm with a meter that they read near zero (section 12 shows how). If you must probe a live supply, clip the meter's common lead on first so you probe with one hand, and keep the other hand away from the chassis. The Safety guide covers the rest: current through the chest, lockout, and first aid.

Transformer-powered vs line-connected

Competency 10.2 asks how transformer-powered supplies differ from line-connected ones. In a transformer-powered supply the mains feeds the primary winding of a power transformer, and the circuits run from a separate secondary winding. The two windings share only a magnetic field, so the secondary circuits, and the chassis they are tied to, are isolated from the mains. A line-connected supply has no transformer between the mains and the rectifier: the AC goes straight to the rectifier (or a voltage doubler), so the circuit's common is tied to the power line. Leaving out the 60 Hz power transformer saves its size, weight and cost, which is why many CRT televisions were built this way, and every off-line switch-mode supply is line-connected on its primary side.

QuestionTransformer-poweredLine-connected (hot chassis)
Isolated from the mains?Yes, by the transformer's separate windings.No. The circuit is tied to the line.
Is the chassis safe to touch?The circuits are isolated from the line by the transformer.Treat it as live: the chassis may be electrically live with respect to earth ground.
Clip a grounded scope's ground lead to circuit common?Yes.Not without an isolation transformer: the ground lead would short that point to earth.
Where you meet itBench supplies and other linear equipment built around a power transformer.CRT TVs and monitors, and the primary side of every off-line switch-mode supply.

Why an isolation transformer matters on the bench. A bench oscilloscope's signal common is tied to the safety ground through its power cord, so every probe's ground lead is an earth connection. Clip that lead to a hot chassis and you short the line to earth through the lead; Tektronix warns that the current can heat the lead until it melts open like a fuse. A 1:1 isolation transformer feeds the unit under test from a winding that has no connection to earth, so touching one point of the circuit while you are grounded no longer completes a path, and the scope's ground lead no longer shorts the line. It does not make the circuit safe: a path from one side of its secondary to the other still shocks you. Never lift the scope's own safety ground to "float" it; isolate the unit under test, or use a differential probe.

On the exam

Exam point: a line-connected (hot-chassis) supply has no transformer between the mains and its circuit, so its chassis or common can sit at line voltage. Power it through an isolation transformer before you connect grounded test equipment.

Trap

Field trap: a variac is not an isolation transformer. Its input and output share one winding, so its output is still tied to the line. If you want to bring a repaired set up slowly, put the variac after an isolation transformer. A GFCI is no substitute either: on a line-connected set it may not protect you at many points in the circuit.

Rectifier circuits

Competency 10.10 asks you to identify and analyze rectifier circuits. A rectifier is a set of diodes that lets current through the load in one direction only, turning AC into pulsating DC. Each diode must withstand the reverse voltage across it when it is off, its peak inverse voltage (PIV), so the rating has to exceed the figure below with margin.

RectifierDiodesRipple on a 60 Hz lineReverse voltage per diode (with a filter capacitor)Notes
Half-wave160 Hz2 × the secondary's peakUses one half-cycle and wastes the other, so it is hard to filter. Fine for small currents only.
Full-wave center-tap2120 Hz2 × the peak of each half (the whole winding's peak)Needs a center-tapped secondary; each half conducts on alternate half-cycles. One diode drop.
Full-wave bridge4120 HzThe secondary's peakNo center tap. Two diodes conduct in series on each half-cycle, so two drops (about 1.4 V for silicon) are lost.

Full-wave rectifiers deliver two pulses per AC cycle, so their ripple is at twice the line frequency. Diode faults show up clearly. If one diode of a bridge opens, the bridge works as a half-wave rectifier: the DC drops and the ripple rises and falls to 60 Hz. If a diode shorts, the next half-cycle puts the secondary across a dead short and the fuse blows. Watch both cases in the Rectifier Lab.

Calculating DC voltages

Competency 10.9 asks how to calculate the DC voltages in a power supply. Every calculation starts from the transformer's secondary voltage, which a meter reads as RMS. Use the factors on ETA's formula sheet, and allow 0.7 V for each conducting silicon diode (see Semiconductor Basics).

V_peak = 1.414 × V_rms  ·  V_avg = 0.637 × V_peak (full-wave, no filter)  ·  V_DC ≈ 1.414 × V_rms − 1.4 (bridge into a capacitor, light load)
CircuitWith a filter capacitor, light loadNo filter: what a DC meter reads
Half-wave1.414 × Vrms − 0.7 V0.318 × Vpeak
Full-wave center-tap (Vrms of each half)1.414 × Vrms − 0.7 V0.637 × Vpeak
Full-wave bridge1.414 × Vrms − 1.4 V0.637 × Vpeak

ETA's sheet gives 1.414 (peak from RMS) and 0.637 (the average of one half-cycle). The half-wave figure, 0.318, is half of 0.637 because half the cycles are missing; it is not on the sheet. The 0.637 and 0.318 figures treat the diodes as ideal; real ones take a little off. A filter capacitor charges toward the peak, so the more current the load draws, the further the DC sags below that light-load figure (section 5).

Worked example

From a 12.6 V secondary: the peak is 12.6 × 1.414 ≈ 17.8 V. Through a bridge into a filter capacitor at light load, the DC is about 17.8 − 1.4 ≈ 16.4 V. With no capacitor a DC meter reads the full-wave average, 0.637 × 17.8 ≈ 11.3 V (ideal diodes); a half-wave rectifier with no capacitor averages 0.318 × 17.8 ≈ 5.7 V. Each bridge diode blocks the 17.8 V peak, so choose a PIV rating well above it.

On the exam

Exam point: a capacitor-input filter charges toward the peak, so its light-load DC is about 1.414 × the RMS voltage minus 0.7 V for each conducting silicon diode. Without a filter, a DC meter reads the average: 0.637 × the peak for full-wave.

Filtering, hum and the L, T and pi filters

Competency 10.4 asks why supplies are filtered, what hum is, and which filters are common. A rectifier's output is DC with a large ripple riding on it. The filter's job is to pass the DC and block the ripple, the AC left over from rectification. Ripple that reaches an audio stage is heard as hum: 120 Hz from a full-wave supply, 60 Hz from a half-wave one. (A 60 Hz hum in full-wave equipment points elsewhere, such as a ground loop or pickup.) A filter capacitor that has dried out, with its ESR risen and its capacitance fallen, is the classic cause of rising hum.

ripple 0 V pulses, no capacitor

Full-wave pulses (dashed) and the filter capacitor's voltage (solid). The capacitor charges to each peak, then feeds the load and sags until the next pulse tops it up. The sag is the ripple, exaggerated here so it shows.

For a capacitor filter, the ripple follows from the I = C(dV/dt) line on ETA's sheet: between peaks the capacitor alone feeds the load for about one ripple period, so

V_ripple(pp) = I_load / (f_ripple × C), with f_ripple = 2 × f_line for full-wave

This is an approximation, and it is not printed on ETA's sheet. It shows the levers: more capacitance, less load current, or a higher ripple frequency all mean less ripple.

On the exam

Exam point: full-wave ripple is at twice the line frequency (120 Hz from 60 Hz); half-wave ripple is at the line frequency. More capacitance, or less load current, means less ripple.

Worked example

Ripple and regulator headroom: a bridge supply peaks at 16.4 V and feeds a 7812 regulator delivering 1 A. With 4700 µF the ripple is 1 / (120 × 0.0047) ≈ 1.77 V p-p, so the troughs sit near 16.4 − 1.77 ≈ 14.6 V, above the 14 V the 7812 needs (12 V plus its typical 2 V dropout). Let the capacitor fade to 2200 µF and the ripple becomes 1 / (120 × 0.0022) ≈ 3.79 V p-p: the troughs fall near 12.6 V, below 14 V, so the regulator drops out on every trough and 120 Hz ripple appears on the 12 V output.

The L, T and pi filters (competencies 10.4.1–10.4.3). Adding an inductor (a choke) in series helps because its reactance opposes the ripple while the DC passes, dropped only by the choke's small winding resistance. Which element comes first decides how the filter behaves. A filter that starts with a capacitor across the rectifier is capacitor-input: the capacitor charges toward the peak. The L-section (choke-input) filter puts the choke first, so the output sits near the average of the rectified wave instead of the peak. Hammond's transformer guide gives about 0.90 × the secondary's RMS voltage for a choke-input bridge against up to 1.41 × (the peak) for a capacitor-input one at light load. The 0.90 is on ETA's sheet too: the average over a half period is 0.9 × the RMS value.

FilterOrder, from the rectifier to the loadHow it behaves and when to use it
Capacitor onlyOne capacitor across the outputCapacitor-input: DC near the peak, ripple grows with load current. Light, steady loads.
L-section (choke-input), 10.4.1Series choke, then a capacitor across the outputOutput near the average, not the peak, but it holds steadier as the load changes. The Navy's NEETS text uses it where regulation matters and the load current is high and varying: radar and transmitter supplies.
T, 10.4.2Series choke, a capacitor across, then a second series choke to the loadETA's description: an L-section with a second choke added on the output side. Still choke-input, so it behaves like the L-section, and the second choke removes more ripple. The load sees the capacitor's voltage less the second choke's small DC drop.
Pi (π), 10.4.3A capacitor across, a series choke, then a second capacitor acrossCapacitor-input: DC near the peak and the smoothest output, because most of the remaining ripple drops across the choke. NEETS lists it in radio receivers and small audio amplifier supplies. Its DC sags more with load than a choke-input filter's.
RC piCapacitor, series resistor, capacitorCheaper than a choke, but the resistor drops DC as well as ripple. Only for low, constant currents where regulation is not needed.

Choose the L-section when the load varies and regulation matters; choose the pi when you need the highest, smoothest DC at a modest, steady current. Size the filter capacitor's voltage rating for the no-load peak, and its ripple-current rating for the job: Hammond notes the RMS ripple current in a filter capacitor can be 2 to 3 times the DC load current.

Regulation

Competency 10.5 asks why supplies are regulated and which parts do it. Every source has internal resistance, so its output falls as the load draws more current. ETA's sheet states it for a battery, and the same idea applies to a transformer and rectifier:

V_out = EMF − (R_int × I_out)

An unregulated supply's output also moves with the line voltage. A regulator holds the output steady despite both; NEETS notes that regulators holding within ±0.1% are quite common. How well a supply holds up under load is its percent regulation, which compares the change from no load to full load with the full-load voltage:

% regulation = ((V_noload − V_fullload) / V_fullload) × 100

Lower is better: 0% is ideal. ETA's 2025 formula sheet does not print a percent-regulation formula; this is the textbook form, the one NEETS describes and the Power Supply calculator uses.

PartWhat it does in a regulated supply
Zener diodeBreaks down at a set reverse voltage and holds it; the reference in most simple regulators. On its own it is a shunt regulator, in parallel with the load.
Series-pass transistorIn series with the load, with its base held at the zener voltage. The output is the zener voltage minus about 0.7 V: a 15 V zener gives about 14.3 V.
Fixed three-terminal regulator (78xx)One IC with the reference, pass transistor and protection inside: fixed outputs such as 5, 12 and 15 V (the 78xx family runs from 5 to 24 V), over 1.5 A with a heat sink, thermal and short-circuit protection. Its input must stay about 2 V above the output (typical dropout).
Adjustable regulator (LM317)Two resistors set the output from 1.25 V to 37 V: V_out = 1.25 × (1 + R_2 / R_1), ignoring a small adjust-pin current. It needs up to 3 V of headroom.
Switching regulatorHolds the output by changing how long a switch stays on (section 7). Far more efficient, but noisier.

Worked example

Three regulation numbers: a supply reads 15.0 V with no load and 14.2 V at full load, so its regulation is (15.0 − 14.2) / 14.2 × 100 ≈ 5.6%. A 12.6 V battery with 0.05 Ω of internal resistance delivering 20 A gives 12.6 − (0.05 × 20) = 11.6 V at its terminals. An LM317 with R1 = 240 Ω and R2 = 390 Ω gives 1.25 × (1 + 390/240) ≈ 3.28 V.

Switch-mode supplies

Competency 10.11 asks how a switch-mode power supply (SMPS) works. Start with what it replaces. A linear regulator drops the difference between input and output across its pass transistor and turns it into heat:

P = (V_in − V_out) × I

A 7805 making 5 V from 12 V at 0.5 A delivers 2.5 W and wastes (12 − 5) × 0.5 = 3.5 W: about 42% efficient, because a linear regulator's efficiency is roughly Vout ÷ Vin.

A switching supply instead drives its transistor either fully on, where it drops almost nothing, or fully off, where it passes no current, so it wastes little either way. It switches at high frequency, and an inductor and capacitor (a low-pass filter) smooth the pulses to their average. The output is the pulse height times the duty cycle, the fraction of each period the switch is on. For a step-down (buck) converter:

V_out ≈ D × V_in, for example 12 V × 0.42 ≈ 5.04 V

A feedback loop measures the output and adjusts the switch's on-time to hold it steady: that is pulse-width modulation. Other arrangements step the voltage up (boost) or produce several outputs. An off-line supply, such as a PC or phone-charger supply, runs this chain from the mains:

  1. Rectify the line. A bridge and a large bulk capacitor make high-voltage DC straight from the mains (line-connected).
  2. Chop it. The switching transistor turns that DC on and off at high frequency.
  3. Transform it. A small high-frequency transformer isolates the output from the line, and its turns ratio sets each winding's voltage. Working at high frequency is what keeps the magnetics small and light.
  4. Rectify and filter again. Diodes and capacitors on the secondary make the DC outputs; the capacitors' ESR sets much of the output ripple.
  5. Feed back. An optocoupler carries the output-error signal back across the isolation barrier to the controller.

The trade-offs: switching supplies are smaller, lighter and cooler (TI's buck example runs 80–89% efficient), but the fast switching makes electrical noise (EMI) that must be filtered, and there are more parts to fail. The parts that wear out are mostly the electrolytic capacitors, whose ESR climbs as they dry out.

Safety

The primary side is live: everything before the switching transformer is line-connected. The bulk capacitor sits near 170 V on a 120 V line, or about 300–340 V on 240 V or behind a voltage doubler, and it can stay charged after you unplug. The secondary side is isolated by the transformer. Discharge and measure the bulk capacitor before you work on the primary side, and use an isolation transformer before you connect a grounded scope there.

Multiple-output supplies

Competency 10.7 asks how one supply can deliver more than one voltage. There are three common ways:

MethodHow it works
Several secondary windingsOne transformer carries several secondaries, and each turns ratio sets its own voltage. Each winding gets its own rectifier, filter and, if needed, regulator. The windings share the primary's power; they do not add to it.
A center-tapped split supplyA center-tapped secondary feeds a bridge used as two full-wave rectifiers, one for each polarity. With the center tap as common, you get equal positive and negative rails. From a 12-0-12 V secondary each rail is about 12 × 1.414 − 0.7 ≈ 16.3 V before its regulator.
A multi-winding switch-mode supplyThe high-frequency transformer has several secondaries. A PC power supply built to Intel's ATX12V design guide delivers +12, +5, +3.3 and −12 V, plus a +5 V standby rail that stays on while the main outputs are switched off.

A lower rail can also be made from a higher one with its own regulator, for example 5 V logic run from a 12 V rail. When you troubleshoot, check every output: one dead rail with the others present points at that rail's own rectifier, filter or regulator, not at the shared front end. And remember the standby rail: a PC supply that is "off" still has line voltage inside and 5 V on its standby output.

The integrated high-voltage transformer (IHVT)

Competency 10.6 names the "integrated high voltage transformer" supply and asks how it differs from other supplies. It is the high-voltage supply of a CRT (picture-tube) television or monitor. It is legacy technology: Panasonic ended its tube-TV production in 2009, manufacturers stopped making CRT sets during the 2010s, and today you meet one only on older equipment. ETA still lists it, so learn how it works.

A CRT needs 20–30 kV on its anode, far more than any line transformer would sensibly make. The flyback, or line output transformer (LOPT), produces it by working with the horizontal output transistor, which switches the flyback's primary at the horizontal scan rate; the pulses at its collector reach about 1,500 V. A high-turns secondary, or a lower-voltage winding feeding a diode and capacitor multiplier, raises them to the anode voltage. The set's B+ of about 120 V goes in; tens of kilovolts come out. In an integrated high-voltage transformer the high-voltage rectifier diodes (often a diode and capacitor multiplier) and the resistive divider for the focus and screen (G2) voltages are built into the flyback's sealed housing. Its outputs are the CRT's anode voltage and usually its focus and screen voltages, and other windings feed lower supplies for the rest of the set. The CRT's own capacitance filters the high voltage.

How an IHVT differsWhy it matters
Runs at the horizontal scan frequency, not 60 HzKilohertz-range pulses let a small transformer and small capacitors do the job.
Part of the horizontal output stageNo horizontal drive, no high voltage. A "no high voltage" fault can be in the drive stage, not the IHVT.
Very high voltage, very low currentTens of kilovolts at a couple of milliamps; the CRT stores the charge.
Rectifiers and divider sealed insideA shorted turn, internal arcing or a failed multiplier usually means replacing the whole IHVT.

Safety

CRT high voltage: a CRT can hold a painful charge for a long time after the set is unplugged. Discharge the anode connection through a high-value resistor (Goldwasser suggests 1 to 10 MΩ, rated 1 W or more) on an insulated stick, with its other end clipped to the metal frame that connects to the outside coating of the CRT (chassis ground, not the signal ground on the main board), before you touch it. Repeat the discharge a few times, and leave a shorting clip lead on while you work: the glass can store charge that reappears after a discharge. Don't measure at the horizontal output transistor with an ordinary meter: its collector pulses reach about 1,500 V.

Batteries and recharging

Competency 10.3 asks about battery supplies and their common uses. Primary cells are not rechargeable. Primary batteries run remote controls, toys, flashlights, wristwatches, tire-pressure gauges and pacemakers: 1.5 V alkaline cells in the everyday devices, and lithium primary cells in most implanted pacemakers. Secondary cells are rechargeable. Cells in series add their voltages: three 3.6 V lithium-ion cells make a 10.8 V pack. Capacity is rated in ampere-hours: divide it by the load current for a rough runtime, as the Battery tool does.

Rechargeable chemistryNominal cellCommon usesHow it is charged
Lead-acid2 VEngine starting, UPS and standby power, emergency lighting, wheelchairs, cell-site backup.Constant current, then a topping charge at 2.30–2.45 V per cell, then a float at 2.25–2.27 V per cell that it can stay on. Charging makes explosive hydrogen, so ventilate.
Nickel-cadmium (NiCd)1.2 VTwo-way radios, emergency medical equipment, professional video cameras, power tools.Constant current with the voltage free to rise. The charger ends the charge on a small voltage drop (negative delta V) or a fast temperature rise, then trickles at 0.05–0.1C.
Nickel-metal-hydride (NiMH)1.2 VConsumer AA and AAA rechargeables.Like NiCd, but with a charger made for NiMH: it must detect a smaller voltage dip and trickles lower, about 0.05C. A NiCd charger can overheat NiMH cells.
Lithium-ion3.6 VPhones, laptops, power tools, medical devices, electric vehicles.Constant current, then constant voltage at 4.20 V per cell (±50 mV) for the common cobalt, nickel and manganese types, until the current falls to about 3% of the rating. Lithium iron phosphate (3.2 V nominal) charges to 3.65 V and needs its own charger. No trickle, no float, no charging below freezing.

Recharging principles (competency 10.3.1). Charging drives current back into a secondary cell to reverse the reaction that discharged it. Three rules hold for every chemistry: use a charger made for that chemistry and cell count, limit the current, and stop (or drop to a safe maintenance level) when the cell is full. The chemistries differ in how "full" is detected and what is safe afterwards. Lead-acid tolerates a continuous float, and its charge voltage should drop about 3 mV per cell for each degree above 25 °C. Nickel cells signal full charge with a small voltage dip or a temperature rise. Lithium-ion cannot absorb overcharge at all: prolonged charging above 4.30 V on a 4.20 V cell plates metallic lithium on the anode, and so would a continuous trickle charge, which compromises safety.

On the exam

Exam point: lithium-ion is charged at constant current, then at a constant voltage (about 4.2 V per cell for most types, 3.65 V for lithium iron phosphate) until the current tapers off, and is never trickle- or float-charged. That is why it needs a managed charger and a protection circuit in the pack.

Fire classes are a separate question: Class D covers combustible metals such as lithium metal, and a lithium-ion pack is a different hazard (see the Safety guide). Don't try to recharge a primary cell.

Fuses and circuit breakers

Competency 10.8 asks where fuses and breakers go, roughly what size they are, and what to watch for in a house service box. The ratings themselves (current, voltage, interrupting rating, fast-acting vs time-delay) are in the Safety guide, section 8.

Where they go in equipment. The main fuse is normally the first part on the primary side, right after the line cord, so it protects everything behind it, including the transformer or the switch-mode front end. It belongs in the hot (ungrounded) line. A fuse that opened in the neutral alone would leave the whole circuit still connected to line voltage. The NEC states the same principle for building wiring: no overcurrent device in series with a grounded conductor unless it opens every conductor of the circuit (240.22). Some equipment adds fuses or fusible resistors elsewhere, such as on individual secondary rails.

How equipment fuses are sized. Pick a current rating so the normal current is no more than 75% of it; Littelfuse notes a 10 A fuse is not usually recommended above 7.5 A at 25 °C. The voltage rating must be at least the circuit voltage. A capacitor-input supply draws a large inrush at switch-on while its filter capacitor charges, and Littelfuse names that inrush as a common cause of nuisance blowing, which is why such supplies use time-delay fuses.

I_fuse ≥ I_normal / 0.75

Worked example

Sizing an equipment fuse: a 120 V unit draws 0.9 A in normal use. Its fuse must be rated at least 0.9 / 0.75 = 1.2 A, so choose the next standard rating at or above that, time-delay because the input has a filter capacitor, with a voltage rating of at least 120 V. When you replace it later, match the type, current and voltage ratings exactly.

In a house service box. Each branch-circuit breaker protects its wiring, so its rating is set by the wire size. Under NEC 240.4(D) (2017 edition), 14 AWG copper is protected at no more than 15 A, 12 AWG at 20 A, and 10 AWG at 30 A. Breakers sit in the hot conductors, never in the neutral alone. A 120 V branch circuit uses a single-pole breaker on its one hot leg. A 240 V load runs between the two hot legs, and its breaker must open all of the circuit's ungrounded conductors (NEC 240.15(B)): a two-pole, common-trip breaker. Two single-pole breakers count only with an identified handle tie, and a handle tie does not link their internal trip mechanisms. Before you work near a panel:

  1. The main breaker does not kill the whole box. The service conductors and the line side of the main breaker stay live even with the main off; only the utility can disconnect them. Leave the service-side barrier in place.
  2. Treat the panel as live until you have tested it with a CAT III meter, live-dead-live (Safety guide, section 11).
  3. Replace like with like. A replacement breaker must be compatible with the panel and listed for it, and its rating must match the wire. Never fit a larger breaker to stop nuisance trips.
  4. A breaker that trips again is telling you something. Find the overload or fault before you reset it, and never keep resetting it.

On the exam

Exam point: a blown fuse is a symptom, not the fault. Find and fix what drew the current before you replace it, then replace it only with the same type, current rating and voltage rating.

On the job: a dead supply

Job practice, not exam content. This walk-through applies the sections above to a linear supply on the bench: 120 V in, a 12.6 V secondary, a bridge, a 4700 µF filter capacitor and a 7812 regulator delivering 12 V at 1 A. Your employer's procedures and the equipment's service manual come first.

  1. Make it safe. Unplug it, discharge the filter capacitor (below), and confirm with a meter that it reads near zero.
  2. Look and smell. Burnt resistors, bulging or leaking capacitors, cracked solder joints and scorched boards often point straight at the fault.
  3. Fuse. Check it out of circuit with an ohmmeter. If it is open, don't replace it yet: measure across the rectifier output and the filter capacitor for a short, and test each bridge diode out of circuit. A diode that reads near zero ohms both ways is shorted.
  4. Power up gently. Once nothing reads shorted, fit the correct fuse and bring the supply up with its current limited, through a series lamp or a variac (fed from an isolation transformer), so a remaining fault cannot destroy more parts.
  5. Measure the checkpoints in the table below, from the line toward the output. The first wrong reading puts the fault between it and the last good one.
  6. Check the capacitors' ESR. A filter capacitor can still read close to its marked capacitance while its ESR has climbed. An ESR meter can test it in circuit (Test Equipment guide).
  7. Find the root cause and verify. Ask why the part failed (an overloaded output, a failing capacitor that overheated the regulator) and fix that too. Then run the supply at full load and confirm every checkpoint.
CheckpointExpectIf it is wrong
Transformer primaryAbout 120 V ACCord, switch or fuse.
Transformer secondaryAbout 12.6 V AC, a little higher with no loadPrimary good but no secondary: an open winding.
Filter capacitorPeaks near 16.4 V, so a DC meter reads about 15.5 V at 1 A (the peak less half the ripple); on an AC-coupled scope, about 1.8 V p-p of 120 Hz rippleLow DC with 60 Hz ripple: an open bridge diode. Large 120 Hz ripple: a weak capacitor (low capacitance or high ESR).
Regulator output12 V DC, flatInput above 14 V but output wrong: the regulator. Output good at no load but sagging under load: an overload, or the input's troughs dipping below 14 V.

Why a blown fuse is a symptom. A fuse opens because something drew too much current. In Goldwasser's switch-mode repair notes, a fuse that blows instantly means a short on the primary side: a rectifier diode, the filter capacitor, a surge suppressor, or the switching transistor, which can take other parts with it. Replacing the fuse without finding that short only blows the new fuse, or worse. Check everything, even after you find one bad part, because one failure often causes another.

Discharging filter capacitors safely. Unplug the unit and connect a resistor across the capacitor with insulated clip leads, so your hands stay clear. The resistor limits the current: Goldwasser suggests a high-wattage resistor, for example 2 kΩ at 10 W for a 200 V, 400 µF capacitor. Never short the terminals with a screwdriver: the resistor is what prevents the arc welding that a dead short causes. Hold the resistor on for at least five time constants, τ = R × C, then confirm with a meter before you touch anything.

Worked example

Discharge timing: for 400 µF at 200 V through 2 kΩ, the stored energy is ½ × 0.0004 × 200² = 8 J. The current starts at 200 / 2000 = 0.1 A, so the resistor briefly dissipates 20 W, falling quickly. The time constant is 2000 × 0.0004 = 0.8 s, and after 5τ = 4 s about 1.35 V is left. Measure before you touch.

Check yourself

Answer in your head first, then reveal.

  1. What separates a line-connected supply from a transformer-powered one?

    A line-connected supply has no transformer between the mains and the rectifier

  2. What is the ripple frequency of a full-wave rectifier on a 60 Hz line?

    full-wave ripple is at twice the line frequency (120 Hz from 60 Hz)

  3. Which filter starts with a series choke, and what does that do to the output voltage?

    The L-section (choke-input) filter puts the choke first, so the output sits near the average of the rectified wave instead of the peak.

  4. How is a lithium-ion cell charged?

    charged at constant current, then at a constant voltage (about 4.2 V per cell for most types, 3.65 V for lithium iron phosphate)

  5. Why is a blown fuse called a symptom?

    a blown fuse is a symptom, not the fault. Find and fix what drew the current before you replace it

Practice this topic: Power Supplies questions are in the Study Hub quiz bank, and the key terms are in Flashcards.