Electrical Fundamentals

Three quantities describe everything happening in an electrical circuit, and you will spend the rest of your career measuring them.

Voltage is electrical pressure — a difference in electrical potential between two points, measured in volts. Current is the flow that pressure causes, measured in amperes. Resistance is what opposes that flow, measured in ohms.

The most important word in that paragraph is difference. Voltage is never a property of a single wire. It is always a measurement between two points, which is why every voltage reading needs two probes and a decision about where the second one goes. More bad diagnoses come from a badly chosen reference point than from a faulty meter.

What you measure, and how

| Quantity | Unit | Meter goes | Circuit must be | | --- | --- | --- | --- | | Voltage | volts (V) | across the device | live | | Current | amps (A) | through the circuit, or a clamp around one conductor | live | | Resistance | ohms (Ω) | across the device | dead and isolated |

Resistance is the odd one out. Your meter measures it by pushing its own tiny test current through the device and seeing what happens. If the circuit is already live, that measurement is meaningless — and on anything above control voltage it can destroy the meter or injure you.

On a running plant, roughly four out of five electrical diagnoses are made with voltage measurements alone, and most of the rest come from a clamp meter. Resistance is the confirmation step, not the opening move — because getting to a resistance measurement means isolating the circuit, and isolating the circuit means you have already stopped production.

Ohm's law, and what it actually buys you

The three quantities are locked together:

V = I × R — and therefore I = V / R and R = V / I.

Textbooks stop there. On the floor, the value of Ohm's law is that it lets you predict a number before you measure it, so the measurement either confirms your model of the circuit or tells you the model is wrong.

Take a 24 VDC relay coil with a winding resistance of 600 Ω. Ohm's law says it should draw 24 / 600 = 40 mA. So:

• 40 mA — the coil is fine.
• 0 mA with 24 V present at its terminals — the winding is open.
• Noticeably more than 40 mA, coil running hot — shorted turns.

You have diagnosed a coil without removing it, because you knew what the number should have been.

The one idea that finds most electrical faults

In a series circuit the supply voltage divides across the resistances in proportion to their size. Control circuits are almost entirely series circuits, and their components sit at two extremes: a closed contact or a good wire is near enough 0 Ω, and an open contact or a broken conductor is near enough infinite.

Put those two facts together and you get the single most useful rule in industrial fault-finding:

The full supply voltage appears across the break.

The corollary is what catches people out. Everything downstream of the break reads 0 V even though it is in perfect condition. A coil measuring 0 V across its terminals is not a faulty coil. It is a coil with nothing arriving at it. Technicians replace healthy components on the strength of that reading every single day.

So the method is: work along the string, measuring across each device in turn. Every healthy closed device reads about 0 V. The one that reads full supply voltage is your fault.

The fault that does not fully break

Not every bad connection is an open circuit. A screw terminal that has relaxed, a crimp that was never properly made, a corroded ferrule — these are not infinite resistance. They are some resistance, and they steal a share of the supply in proportion.

This is why you measure across connections and not merely at them. A terminal block that should be a straight-through conductor and is instead dropping 3 V is a fault, even while the machine is still limping along. In a low-current control circuit a healthy connection drops tens of millivolts at most. Anything you can read as a whole volt is a joint that has started turning into a heater, and heating makes the resistance worse, which makes the heating worse.

This measurement only means anything with the circuit energised and current actually flowing through the joint. A resistance only reveals itself as a voltage drop when there is current to drop it. Measure across the same connection with the circuit dead and you read nothing; measure across it while the circuit is energised but the load is off, or while the joint is on the open side of a switched circuit, and you can read the full supply voltage instead — which looks alarming and means something completely different. Get the machine running, or at least get the circuit loaded, before you read anything into a number taken across a connection.

Series and parallel behaviour in control circuits

Series is AND. Every contact must be closed. This is how interlocks and permissives are built: guard closed AND overload healthy AND run request present, then the coil pulls in. The same current flows through every element, and the supply voltage divides between them. One open contact kills the entire string — and the string will not tell you which one until you measure across each in turn.

Parallel is OR. Any complete path energises the load. This is how start buttons, multiple operator stations, and seal-in circuits work. Every branch sees the same voltage, and the current divides between them.

The two failure characters are different, and the difference matters:

A series fault stops the machine. A parallel fault stops the machine from stopping. A welded contact in a parallel branch keeps a circuit energised when it should have dropped out. Nothing appears broken — production carries on — and the failure only reveals itself at the moment somebody needs the machine to stop. That is the dangerous kind, and it never announces itself.

AC and DC on the plant floor

Both are in the same panel, usually within a few centimetres of each other.

Three-phase AC feeds the motors, the drives, and the heaters. On a 400 V system you measure 400 V between any two phases and about 230 V from a phase to neutral; on a 480 V system, 480 V phase-to-phase and 277 V phase-to-neutral. AC voltages are quoted as RMS by convention, and the peak is RMS × 1.414 — so "230 V" is swinging to roughly ±325 V, which is what the insulation and the component ratings actually have to survive.

Control voltage is 120 or 230 VAC through a control transformer on older machines, and 24 VDC on nearly everything built in the last two decades. PLC I/O, sensors, safety relays and HMI power are almost always 24 VDC.

DC has polarity and AC does not, which changes how you work:

• On DC, reversed probes give you a negative reading — informative, not a fault.
• On DC, polarity is part of the circuit's correctness. Sensor wiring, PNP versus NPN, and diode-protected coils all depend on it (Chapter 7).
• On AC there is no polarity, but there is phase. Two conductors can each be live to earth and still have 400 V between them.

Reading real values without lying to yourself

Your meter is an instrument with characteristics, not an oracle.

Ghost voltage. A digital meter has a very high input impedance — typically 10 MΩ — so it barely loads the circuit. That is normally a virtue. But a long run of cable sitting in a tray beside live conductors picks up voltage capacitively, and a high-impedance meter will happily show you 40 or 80 V on a conductor that is genuinely dead. There is no current behind it. Switch to your meter's low-impedance (LoZ) mode, or use a two-pole tester: loading the circuit collapses the ghost and leaves the real voltage standing.

True-RMS versus average-responding. A cheap meter measures the average of the rectified waveform and multiplies by a fixed factor that is only correct for a pure sine wave. On anything distorted — a VFD output, a switch-mode supply, a phase-controlled heater — it can be wrong by tens of percent. Use a true-RMS meter, and know which one is in your hand.

VFD outputs are a special case. The output of a drive is not AC at all in the usual sense; it is a train of high-voltage pulses whose width is modulated to approximate a sine wave. Most meters read it wrongly — an average-responding meter completely, and even a true-RMS meter can be thrown by the carrier frequency riding on top of the waveform.

So the order is: the drive's own display first. Its readout of output voltage, current and frequency is the trustworthy first reference, and it costs you nothing to look at.

Know what that display cannot tell you, though. It reports what the drive believes it is delivering at its own terminals, so it will not see a fault downstream of them — a damaged motor cable, a loose connection in the motor terminal box, a failing winding. That is where measurement earns its place: use a meter with a low-pass filter setting (often marked VFD or LPF) if you must read the drive output directly, or better, put a clamp meter on the motor current, where a genuine imbalance between phases will show up plainly. Go to a plain meter only for the incoming supply.

Resolution and range. On a 600 V range, the last digit is not a measurement you can rely on. Know what range you are on before you attribute meaning to a small number.

CAT rating. A meter's CAT II/III/IV rating describes the transient overvoltage it can survive at that point in the installation — not the steady voltage it can read. Chapter 4 covers instrument selection properly.

Electrical faults on production equipment are not evenly distributed. A small number of failure modes account for most of the calls, and knowing the distribution is what makes the difference between a technician who searches and one who checks.

The order matters more than the individual measurements. This sequence is built so that each step either eliminates a large part of the circuit or tells you which half to look in next.

1. Establish the safe condition first

   Safety

   Identify every energy source feeding the equipment — electrical, pneumatic, hydraulic, stored, gravitational — and apply your site's lockout/tagout procedure to the extent the work requires. If you must work on it live because the fault only exists when it is running, that decision needs the correct authorisation, the correct PPE and, wherever possible, a second person. See Chapter 2.

2. Prove your meter: live, dead, live

   Safety

   Test the meter on a known live source. Test the circuit you are about to work on. Test the known live source again. A meter that failed between the first and last check has just been caught, instead of telling you a live circuit was dead. This takes twenty seconds and is the step people skip.

3. Define the symptom precisely

   Not "it's not working". What stopped, when, and under what conditions? Does it fail from cold or only when warm? At one position or everywhere? Every time or one shift in three? Ask the operator — they have watched this machine for longer than you have, and the conditions surrounding the failure usually narrow the search more than the first three measurements would.

4. Check the supply before you check the circuit

   Confirm the control transformer or 24 V supply is at its rated voltage under load. There is no point tracing a control string that is being fed 18 V. This step costs one measurement and eliminates an entire class of fault.

5. Split the circuit in half

   Do not walk the string end to end. Measure at a point roughly in the middle. If the voltage is present there, the fault is downstream; if it is absent, the fault is upstream. Each measurement halves what is left. On a twenty-device string that is five measurements instead of twenty.

6. Follow the voltage across each device

   Within the half you have isolated, measure across each device in turn. A healthy closed contact or conductor reads about 0 V. The device holding the full supply voltage across it is the break. Remember that everything past the break reads 0 V and is probably innocent.

7. Confirm with resistance — after isolating

   Safety

   Once you believe you have found it, isolate the circuit, prove it dead, and confirm with a resistance measurement. A suspect contact should read near 0 Ω when made and open when released. Never take a resistance reading on a live circuit.

8. Find out why it failed

   A blown fuse, a tripped overload and a burnt contact are all consequences. Ask what caused them. Replacing the component that failed without answering that question buys you a shift, and the callout comes back on somebody else's night.

9. Verify under real operating conditions

   Restore the guarding and the isolation, and run the machine through the conditions that produced the fault — the same cycle, the same position, the same load. A repair confirmed only at standstill is not confirmed. Then write down what you found, so the next person inherits your evidence instead of repeating it.

Every one of these is made by competent, experienced technicians. They are worth studying precisely because none of them come from carelessness — each has a reason behind it that makes sense in the moment.

• Measuring resistance on a live circuit

  Why

  Bench habits transfer badly. On a workbench everything is dead by default, so reaching for the ohms range is automatic. In a panel the default is the opposite. The meter's own test current cannot compete with an energised supply, so the reading is meaningless — and above control voltage you will damage the meter or hurt yourself.

• Trusting a reading without proving the meter

  Why

  The meter works virtually every time, so checking it feels like ceremony. That is exactly the trap: the failure is rare, silent, and lands on the one occasion you needed it — when you declare a live circuit dead and then put your hands on it. Live-dead-live costs twenty seconds against a consequence that is permanent.

• Chasing ghost voltage

  Why

  A high-impedance meter reading 60 V on a disconnected conductor looks exactly like a live wire, and the instinct is to go and find where it is being fed from. It is capacitive coupling from cables running alongside, with no current behind it. Half an hour disappears before somebody suggests LoZ mode.

• Measuring against the wrong reference

  Why

  Voltage is a difference, so a reading is only as good as where the black probe is. Clipping the common onto a convenient piece of chassis rather than the circuit's actual reference gives readings that look plausible and mean nothing. This one is quiet — it produces numbers, and the numbers are wrong.

• Replacing what failed without asking why it failed

  Why

  Fitting a new fuse makes the symptom disappear immediately, and immediate feedback feels like a fix. Under production pressure, with a supervisor waiting, that feeling is very hard to argue with. But a fuse does not fail of old age — something took more current than it should have, and it is still there.

• Using an averaging meter on a drive output

  Why

  Nothing warns you. The meter shows a confident, stable, plausible number, and it can be wrong by tens of percent on a PWM waveform. The technician then reasons carefully from a false measurement, which is worse than having no measurement, because now there is evidence pointing the wrong way.

Five questions. Each one is answerable by reasoning from the readings rather than by remembering a definition — which is the point of the whole course.