Diagnose a dead or misbehaving board
Answer each question with your meter in hand. The flow follows the same order a repair tech uses: eyes first, then power in, then fuse, then rails, then heat, then signals. Ninety percent of faults fall out before the last step.
How to measure each part
Every card shows the meter mode, where the probes go, what a healthy part reads, and how the common failures look on the display. Unless a card says otherwise: power off, capacitors discharged, and for a trustworthy reading, lift one leg of the part out of circuit.
▸Resistor
Ω resistance
- Read the marked value first — color bands or SMD code (use the Calculators tab).
- Set the meter to Ω. Autorange handles the rest; on a manual meter pick the range just above the marked value.
- Touch one probe to each end. Polarity doesn't matter.
- Don't pinch both probe tips with your fingers — your body is a ~1 MΩ resistor in parallel and will skew high-value readings.
| Display | Verdict |
|---|---|
| 9.87 kΩ on a 10 kΩ part | GOOD within tolerance (±1% or ±5% typically) |
| OL | OPEN — the classic resistor failure, often after overheating |
| Way above marked value | DRIFTED — old carbon resistors drift high; replace |
| Below marked value in circuit | NORMAL — parallel paths through the circuit pull the reading down. Lift one leg to confirm. |
▸Diode & LED
⏄ diode mode
- Set the meter to diode mode (⏄ symbol). It pushes a small test current and displays the forward voltage drop.
- Forward: red probe on the anode, black on the cathode (the end with the stripe). Expect a drop.
- Reverse: swap the probes. Expect OL — no conduction.
- LEDs test the same way and will glow faintly in forward. Their drop is higher: red ≈1.8 V, blue/white ≈2.7–3.3 V.
| Forward / Reverse | Verdict |
|---|---|
| 0.5–0.7 V / OL | GOOD silicon diode (1N4148, 1N4007, rectifiers) |
| 0.15–0.45 V / OL | GOOD Schottky diode — the low drop is normal |
| ~0.00 V both directions | SHORTED — the most common diode failure |
| OL both directions | OPEN |
| Low reading both directions, in circuit | RECHECK — something in parallel is conducting; lift one leg |
▸Capacitor
F capacitancedischarge first
- Capacitance mode (F): best done out of circuit. Reading should be within tolerance of the marked value — electrolytics are typically ±20%.
- Quick Ω check (electrolytics, in a pinch): on the Ω range the meter charges the cap — a healthy one shows a climbing number, then OL. A stuck-low reading means shorted.
- Electrolytics also fail by ESR rising — capacitance can still measure fine. A cheap ESR meter tests this, even in circuit. Suspect ESR when a power supply whines, sags, or has ripple.
- Polarity matters on electrolytics and tantalums: the stripe marks negative on aluminum cans, positive on tantalums.
| Test | Verdict |
|---|---|
| 467 µF on a 470 µF part | GOOD |
| Ω mode stays near 0 Ω | SHORTED — common killer of power rails |
| Capacitance far below marked | DRIED OUT — electrolyte gone; replace |
| Bulged top, leaking base | REPLACE — no measurement needed (see Visual Inspection) |
▸Transistor (BJT)
⏄ diode mode
A bipolar transistor tests like two diodes joined at the base. Find the pinout on the datasheet first — packages vary.
- NPN: red probe on the base. Black to the emitter → ~0.6 V. Black to the collector → ~0.6 V.
- PNP: same two tests but with the black probe on the base.
- Reverse each junction → OL.
- Collector-to-emitter must read OL in both directions (a small reading may appear on Darlingtons or parts with built-in resistors — check the datasheet).
| Reading | Verdict |
|---|---|
| B–E and B–C ≈ 0.6 V, reverse OL, C–E OL | GOOD |
| C–E ≈ 0.00 V either direction | SHORTED — the classic dead-output-stage failure |
| A junction reads OL both ways | OPEN |
▸MOSFET
⏄ diode mode
Power MOSFETs (N-channel shown) are the most-blown part in power supplies, motor drivers, and laptop boards. Three quick checks, part out of circuit or at least isolated:
- Discharge the gate first: briefly short gate to source with your fingers or a resistor.
- Body diode: red on source, black on drain → ~0.4–0.6 V. Swapped → OL. This diode is built in and is normal.
- Gate isolation: gate-to-source and gate-to-drain → OL both ways. Any reading = blown gate oxide.
- Turn-on trick: touch red to gate for a second (charges the gate), then measure drain-to-source (red on drain) — it now conducts and reads low. Short gate to source and it goes back to OL. That proves the FET actually switches.
| Reading | Verdict |
|---|---|
| Body diode one way, gate OL, turn-on trick works | GOOD |
| D–S ≈ 0.00 V both ways | SHORTED — usually takes the fuse or the driver with it |
| Any G–S / G–D conduction | GATE BLOWN |
▸Fuses, traces & connectors
⋓ continuity
- Set the meter to continuity (beep symbol). It beeps below roughly 30–50 Ω depending on the meter.
- Fuses: probe both ends. Beep + 0.1 Ω = good. OL = blown. Glass fuses can look intact and still be blown — always measure.
- Traces: probe each end of the suspect trace (component pad to component pad). Hairline cracks near connectors and mounting holes are common — flex the board gently while listening for the beep to cut out.
- Connectors & cables: probe pin-to-pin through the mated cable while wiggling it. An intermittent beep is a broken crimp or fatigued wire.
- A blown fuse is a symptom. Before fitting a new one, measure resistance from the fuse's load side to ground — a few ohms or less means the short that killed it is still there.
▸Inductor / transformer winding
Ω resistance
- Set the meter to Ω. A healthy inductor is just a coil of wire: expect a very low reading — from under 1 Ω (power inductors) to a few hundred ohms (small transformer primaries).
- OL = open winding = dead. This is the failure a DMM can catch.
- What a DMM can't catch: shorted turns. The winding still reads low ohms but the inductance collapses. If a transformer runs hot, buzzes, or its supply hiccups, suspect shorted turns and substitute or use an LCR meter.
- Transformers: also check primary-to-secondary — must be OL (isolated).
| Reading | Verdict |
|---|---|
| 0.4 Ω power inductor | GOOD |
| OL | OPEN WINDING |
| Primary-to-secondary continuity | INSULATION FAILURE — replace, safety hazard |
| Low ohms but runs hot / buzzes | SUSPECT SHORTED TURNS — needs LCR meter or substitution |
▸Voltage regulator
V⎓ DC voltspowered test
Linear regulators (78xx, AMS1117, LDOs) are tested live — this is one of the few powered measurements in routine troubleshooting.
- Find the pinout (e.g. 7805 TO-220: IN–GND–OUT left to right, tab = GND; AMS1117: GND–OUT–IN, tab = OUT).
- Black probe on ground. Power the board.
- Measure the input pin: it needs headroom — a 7805 wants ≥7 V in; an LDO needs the output voltage plus ~0.3–1 V.
- Measure the output pin: should be within ~4% of nominal (5 V → 4.8–5.2 V).
| Symptom | Likely cause |
|---|---|
| Input good, output 0 V | Dead regulator — or a short on the output rail dragging it down. Power off, measure rail-to-ground Ω before condemning the regulator. |
| Output low and regulator hot | Overloaded rail (partial short downstream) or not enough input headroom |
| Output ≈ input (no regulation) | Regulator shorted through — this overvolts everything downstream; check what it may have killed |
| Switching regulator, no output | Check its enable pin (often needs a logic high) and the inductor before blaming the chip |
▸Integrated circuits
V⎓ + Ω
You can't fully test an IC with a multimeter, but you can catch the failures that matter most:
- Supply-pin short check (power off): Ω between the IC's VCC and GND pins. Compare against a known-good board if you have one. Dead short (<5 Ω) usually means the IC — or a decoupling cap next to it — is shorted internally.
- Power check (power on): DC volts on each supply pin, black probe on ground. No volts = trace/rail problem, not the chip.
- Heat check: a chip too hot to hold a finger on within 2–3 seconds, doing nothing, is almost certainly shorted internally.
- Isolating a shorted rail: if a decoupling cap and the IC share the short, feed 1 A at under 1 V from a bench supply into the rail and feel/IR for the warm part — or lift components one at a time.
- Logic activity, crystals, and data lines need an oscilloscope or logic probe — a DMM only shows averages.
| Check | Healthy | Suspect |
|---|---|---|
| VCC–GND resistance (off) | kΩ range, matches donor board | <5 Ω |
| Supply pin voltage (on) | Nominal rail ±5% | Missing or sagging |
| Temperature at idle | Warm at most | Untouchable in seconds |
| Output pin stuck at rail or 0 | — | Driver stage blown |
Know your multimeter
Every diagnosis in this app comes down to five meter modes and a handful of habits. Learn the dial once and the rest of the app reads like a checklist.
The golden rules
- Voltage is measured in parallel, powered on. Black probe on ground, red probe on the point of interest. You never break the circuit.
- Resistance, continuity, and diode tests: power off, caps discharged. Voltage on the probes gives garbage readings and can damage the meter.
- Current is measured in series — you break the circuit and route it through the meter. Rarely needed for board repair; when you do, mind the meter's fused current jack and never leave the leads in the A jack afterward (measuring "volts" with leads in the amp jack is a dead short).
- Black lead lives in COM. Red moves between VΩ and A jacks depending on the job.
- In-circuit ohm readings are lower bounds. Everything in parallel with the part conducts too. A reading at or above the marked value in circuit is trustworthy; a low one means lift a leg and re-measure.
- One hand in the pocket when probing anything mains-connected — no current path across your chest.
Reading the display
| Display | Means |
|---|---|
| OL / 1 . | Overload / open — beyond range, or no conduction at all. On continuity and diode tests, OL means "no path." |
| 0.00 on Ω | Dead short — or your probes are touching. Short the probes together first to learn your lead resistance (typically 0.1–0.5 Ω) and subtract it mentally. |
| mV / k / M unit prefixes | Autorange moved the decimal. 4.70 kΩ and 4700 Ω are the same reading — check the unit before you judge the number. |
| Jumpy last digit | Normal noise. Jumpy first digit means a bad probe contact or an intermittent joint — which may itself be the fault you're hunting. |
Sinking vs. sourcing
Every output pin, sensor, and driver moves current in one of two directions — it either sources current (pushes it out of the pin, through the load, to ground) or sinks it (the load hangs from the positive supply and the pin pulls current in, to ground inside the device). Get the direction wrong and a perfectly healthy circuit does nothing.
Quick reference
| Sourcing | Sinking | |
|---|---|---|
| Current at the pin | flows out | flows in |
| Load connects between | pin and ground | V+ and pin |
| Pin voltage when active | ≈ V+ | ≈ 0 V |
| Also called | high-side, PNP-type | low-side, NPN-type, open-collector/drain |
| Typical examples | PNP proximity sensors, high-side load switches | ULN2003/2803 drivers, low-side MOSFETs, NPN sensors, most relay/LED driver stages |
Identify it with your meter
- Powered, load connected: DC volts from the output pin to ground, then activate the output. Swings up to ≈ V+ when active → sourcing. Drops to ≈ 0 V when active → sinking.
- Open-collector / open-drain outputs float when off. With no pull-up you'll read garbage or a slowly drifting voltage. Clip a 10 kΩ resistor from the pin to V+ first: now it reads V+ when idle and ~0 V when active. A "dead" output that works fine once pulled up isn't dead — it just only sinks.
- Voltage across the load is the universal check: full supply voltage across the load = it's being driven, whichever side does the switching.
- Dead-output autopsy: a sinking driver that failed usually reads shorted pin-to-ground (diode-test it like a transistor — see Measure Parts); a sourcing driver that failed often puts V+ on the load permanently.
Matching sensors to inputs
The rule that untangles every PLC and controller hookup: a sourcing device feeds a sinking input, and a sinking device feeds a sourcing input. Current has to leave one side and enter the other — two sources or two sinks face each other and nothing flows.
| Sensor / output | Needs an input that… |
|---|---|
| PNP sensor (sourcing) | sinks — input common wired to 0 V. The default in European/IEC equipment. |
| NPN sensor (sinking) | sources — input common wired to +24 V. Common in Asian-market equipment. |
| Open-collector output → MCU | a GPIO with a pull-up (internal or external) — the output only ever pulls down |
Sink and source in power supplies
The same words describe power gear, and the distinction matters at the bench:
- A normal bench supply only sources — it pushes current out. It cannot absorb current pushed back at it: connect it across a charged battery set lower than the battery voltage and current flows the "wrong way" with nothing regulating it.
- An electronic load only sinks — it's the test-bench partner for exercising supplies and batteries.
- Two-quadrant supplies (battery simulators, SMU/source-measure units) do both, which is why they — not a bench supply — are the right tool for battery-charging circuits.
- On a board, this is why a rail can measure fine unloaded but sag in service: the regulator sources happily, but a motor braking or an inductive load dumping energy needs somewhere to sink it — that's the job of flyback diodes and brake resistors. A missing or open flyback diode kills the driver transistor.
Component code calculators
Read the part before you measure it — a reading only means something compared to what the part claims to be.
Resistor color code
Capacitor code
Ceramics and films are marked in picofarads: two digits + a multiplier. 104 = 10 × 10⁴ pF. 4R7 = 4.7 pF. Electrolytics print the real value.
SMD resistor code
3-digit: two digits × 10ⁿ. 4-digit: three digits × 10ⁿ. R is a decimal point (R047 = 0.047 Ω). A single 0 is a jumper. (Tiny parts marked with two characters use the EIA-96 table — look those up.)
Common values cheat sheet
| Marking | Value | Typically found |
|---|---|---|
| 102 | 1 nF | Snubbers, filters |
| 104 | 100 nF (0.1 µF) | Decoupling caps — everywhere, next to every IC |
| 473 | 47 nF | Timing, filtering |
| 225 | 2.2 µF | LDO input/output caps |
| 331 (SMD R) | 330 Ω | LED current limiting |
| 472 (SMD R) | 4.7 kΩ | Pull-ups on I²C and reset lines |
| 103 (SMD R) | 10 kΩ | Pull-ups/downs — the most common resistor on any board |
| 000 / 0 | 0 Ω jumper | Configuration links, current-measurement points |
Reading quiz
A reading only helps if you can call the verdict in the moment. Each question hands you a display reading or a bench situation — pick the interpretation. Code questions are freshly generated every time, so the drill never runs dry.
Fault finder
Match what the board is doing to the usual suspects. Start with the first check listed — they're ordered by likelihood and effort.
| Symptom | Likely causes | Check first |
|---|---|---|
| Completely dead — no LED, no sign of life | No input power · blown fuse · shorted input protection · dead regulator | DC volts at the input connector, then continuity across the fuse, then rail voltages. Follow the Diagnose flow. |
| Fuse blows instantly, every time | Hard short after the fuse: shorted rectifier, MOSFET, or capacitor · shorted TVS/protection diode doing its job | Power OFF: Ω from the fuse's load side to ground. Diode-test the bridge rectifier and any power FETs. Never fit a bigger fuse. |
| Powers on, then shuts down after seconds or minutes | Overcurrent protection tripping · thermal shutdown · failing electrolytic that worsens as it warms | IR gun or finger-hover for anything heating fast. Watch a rail with the meter as it dies — a sagging rail points at its load. |
| Works intermittently, or when the board is flexed or tapped | Cracked solder joint · fractured trace · bad connector crimp · cracked SMD part near a screw hole | Gently flex the board and tap components with an insulated stick while it runs. Magnify joints on connectors and large/heavy parts — reflow suspects. |
| One component runs hot | The part itself is shorted · or its load downstream is shorted and it's supplying the fault current | Power off, Ω across the hot part and from its output to ground. Hot regulators usually mean an overloaded rail, not a bad regulator. |
| Audible whine or buzzing | Dried-out electrolytics in a switching supply making it unstable · magnetics under abnormal load | Inspect for bulged caps; check ESR. Whine that changes with load points at the switching supply feedback loop. |
| Rails all correct, still doesn't work | Held in reset · dead crystal/clock · corrupted firmware · dead MCU/IC · unseated cable | DC volts on the reset line (should be released, usually high). Reseat every connector and ribbon. Beyond this you want a scope on the crystal. |
| Erratic, glitchy behavior | Ripple from dried filter caps · bad ground connection · corrosion · marginal solder on a busy IC | Meter on AC millivolts across the DC rail: more than a few tens of mV of ripple means filter caps. Inspect grounds and mounting screws. |
| Burnt smell or visible scorch mark | A part exceeded its rating and let the smoke out — the interesting question is why | Identify the charred part from a schematic or donor board, then test everything electrically adjacent. Replacing it without finding the root cause usually burns the new part. |
| Liquid spill or corrosion | Electrolytic corrosion eating traces and vias · conductive residue bridging pins | Disconnect power immediately. Scrub with isopropyl alcohol and a soft brush, dry fully, then continuity-test traces in the affected area — green/white crusty traces are often open underneath. |
| Rail voltage low or sagging | Partial short loading the rail · weak regulator · insufficient input supply | Compare rail-to-ground Ω against a good board. Disconnect loads (pull connectors) one at a time and watch the rail recover. |
| Ripple / hum in audio or video output | Dried-out filter capacitors · ground loop | AC mV across each electrolytic in the supply path — the one with volts of AC across it has quit filtering. Replace in kind (capacitance, voltage, temperature rating, low-ESR if it was). |
Visual inspection reference
Half of all board faults are visible before you ever pick up the meter. Work under bright light with magnification, smell the board, and check both sides. Here's what failure looks like.
Electrolytic capacitors
The scored top is a pressure vent. Any doming, splitting, brown crust at the vent, or residue at the rubber base means the cap is done — bulged caps are the single most common repairable fault in consumer electronics.
Fix: replace with same capacitance and voltage, same or better temperature rating (105 °C), low-ESR for switching supplies. Mind the polarity stripe.
Solder joints
A good joint is shiny (with leaded solder) and concave, wetted onto both pad and lead. Cold joints look dull, grainy, and balled-up. Ring cracks — a hairline circle around the lead — form on heavy parts and connectors from vibration, and cause "works when I press on it" faults.
Fix: reflow with flux and fresh solder. On connectors and big caps, reflow every pin, not just the visibly bad one.
Burns and scorch marks
Browned board material, charred parts, or bubbled solder mask mean a part ran far beyond its rating. You'll often smell it before you see it — burnt phenolic is unmistakable. The burned part is the victim; the fault that overloaded it may be elsewhere.
Fix: identify the part's value from a schematic, donor board, or the readable half of the markings. Test its neighbors — especially semiconductors — before powering up with the replacement.
Corrosion & liquid damage
Green or white fuzz around pins, vias, and battery holders is electrolytic corrosion — and the copper underneath is often eaten through even when the trace looks intact. Leaked electrolyte and battery alkali keep corroding until removed.
Fix: scrub with 90%+ isopropyl alcohol (vinegar first for alkaline battery leaks), dry thoroughly, then continuity-test every trace through the affected zone. Bridge eaten traces with thin wire.
SMD faults under magnification
Sweep the board with a loupe or phone macro lens: tombstoned parts standing on one end, hairline cracks across ceramic capacitors (especially near screw holes and board edges), solder bridges between fine-pitch IC pins, and parts knocked off entirely.
Fix: tombstones and bridges reflow easily with flux. Cracked ceramic caps must be replaced — they fail shorted and can burn.
The full inspection ritual
- Smell it. Burnt board smell survives for years and leads you to the fault area.
- Bright light + magnification, both sides of the board, methodically — row by row like reading a page.
- Look at every electrolytic top and base for doming and leakage.
- Rake light across solder joints at a low angle — cracks catch shadows. Focus on connectors, transformers, heavy caps, and anything that gets hot or flexed.
- Check the mechanical stress points: around mounting screws, connectors, and board edges — that's where ceramics crack and traces fracture.
- Wiggle test: gently push components and flex the board while powered (low-voltage boards only) and watch for the fault to come and go.
- Compare with photos of a known-good board when available — missing and cooked parts jump out.