Cómo usar un multímetro
A meter does not tell you what is wrong. It answers exactly the question you asked it, and most bad calls are a good meter answering the wrong question. There are only four measurements you will take with any regularity, and each one has a way of lying to you if you take it carelessly. Learn the four, learn the two lies, and the meter becomes the tool that ends arguments instead of starting them.
The four measurements
- Voltage — is there pressure between these two points? The one you will reach for far more than the other three, and the only one taken on a live circuit by choice.
- Continuity — is there a path between these two points? A cold check, on a circuit you have proved dead, with a beep for yes.
- Resistance — how much does that path oppose current? Continuity with a number on it: a good coil versus an open one.
- Current — how much is actually flowing? Almost always with a clamp around one conductor, because the alternative is breaking the circuit and putting the meter in it.
Three of those four are done with the circuit dead. That is not caution for its own sake — a dead circuit gives cleaner answers, because nothing is feeding the path you are trying to measure.
Voltage: across, never along
Voltage is a difference, so it only exists between two points. That single fact decides where the leads go: on the two points whose difference you want. Across the lamp. Across the open switch. From the terminal to the bond. There is no such thing as the voltage "at" a wire on its own — there is only the voltage between that wire and something else, and the something else is the half of the measurement people forget to say out loud.
Pick the reference deliberately. Hot to neutral on a 120 V circuit tells you what the load will see. Hot to bond tells you whether the bond is actually home. Neutral to bond should be nothing, and when it is not, the return is carrying something it should not. Those are three different questions and three different lead positions, and a reading only means anything once you can say which one you took.
On AC circuits the dial goes to the wavy V. On control circuits fed from a battery, a power supply or a PLC output, it goes to the straight-and-dashed V. Modern meters autorange, so you are not picking a scale — but you are still picking AC or DC, and an AC reading on a DC circuit reads low or reads nothing at all. If a circuit you believe is live reads zero, check the dial before you check the circuit.
This is the reading that teaches the idea, and it is worth doing on a bench once. With the switch open, the meter across the open switch shows the whole supply: the conductor on one side is live, the one on the other is not, and the difference is everything there is.
Move the same two leads onto the lamp and you get nothing, because both of the lamp's terminals are now at the same potential — the return side. Close the switch and the two readings swap over exactly: the lamp sees everything, the closed switch sees nothing at all.
That swap is the whole of fault-finding on a series string. Whatever is open has the supply across it. Walk the meter along the rung and the place where the reading appears is the place that is broken — it is the method the motor control troubleshooting lesson is built around.
Continuity: the cold check
Continuity puts a tiny current out of the meter's own battery and listens for it coming back. If it does, you get a beep. That is all it means: a path exists. It does not mean the path you were thinking of, and it does not mean a path good enough to carry load current.
Two rules make it trustworthy. First, the circuit has to be dead and proved dead — a continuity test on a live circuit gives a nonsense reading at best and blows the meter's input at worst. Second, whatever you are testing has to be isolated. Lift one end of it. Otherwise the beep may be travelling through a lamp, a winding, a pilot light or the next device along the run, all of which are simply resistance as far as a meter is concerned.
Used properly it is the fastest tool you own. Ring out a 3-way switch in your hand and the common shows continuity to one traveller screw with the handle up and the other with it down, while the two travellers never read to each other — the device tells you which screw is which in ten seconds. Ring out a length of cable from one end with the far end shorted and you have proved the whole run without walking it twice.
Resistance: continuity with a number on it
Switch to ohms and the same test comes back with a figure. That figure turns a yes-or-no into a judgement. A relay coil reads whatever that coil is built to read — tens of ohms on one, thousands on another, depending on its coil voltage — and a burnt-out one reads open. A heating element reads a sensible handful of ohms; a cracked one reads open. A motor winding reads low and — this is the useful part — the three windings of a three-phase motor read close to each other. One winding well away from the other two is the fault, without the motor ever having to run.
What counts as a sensible number is set by the device, not by a rule of thumb: it is on the manufacturer's sheet, on the device itself, or in the figures somebody recorded when it was new. This page cannot give it to you and neither can anybody else's. What it can give you is the habit of writing the reading down while the machine is healthy, so that the day it is not you have something to compare against.
Current: the clamp, nearly always
Current is the one measurement taken along a circuit rather than across it, which is why the in-line method means breaking the circuit, moving the red lead to the current jack, and putting the meter into the path. It is slow, it puts the meter in series with a live load, and getting the leads wrong is how meters get destroyed — a meter left in the current jacks and then placed across a live circuit is a short across the supply through the meter's own fuse.
So in practice you use a clamp. A clamp meter reads the magnetic field around one conductor without breaking anything, with the circuit running and the cover on. Clamp one conductor at a time: put the jaw around a cable with the hot and the neutral both inside it and the two fields cancel, and you read nothing — which, incidentally, is exactly how a GFCI decides something has gone wrong.
Current is the measurement that answers "is it working hard?". A motor pulled from its nameplate figures, three legs measured and compared, a circuit checked before somebody adds to it. What figure is acceptable comes from the equipment's nameplate and from your code book — the meter supplies the reading, the book supplies the limit.
The two readings that lie
1. Ghost voltage
Symptom: a conductor you know is disconnected reads a stubborn fraction of line voltage, and will not go away.
A digital meter draws almost nothing to take a reading, which is what makes it accurate. It also makes it sensitive to voltage that has no current behind it: a long conductor running beside a live one picks up a coupled voltage through the capacitance between them, and your meter shows it because nothing is there to drag it down. Put a real load across it — a LoZ or low-impedance setting if the meter has one, or a solenoid tester — and the reading collapses to nothing. The collapse is the proof. A reading that does not collapse is a live conductor, and you treat it as one.
2. A beep through something else
Symptom: continuity between two points that are definitely not joined — sometimes between a conductor and a completely different circuit.
The meter found a path. It cannot tell you the path went through a lamp filament, a transformer winding, a pilot light, or a device three boxes along. Every one of those is a resistor to a continuity tester, and a continuity tester beeps at resistance under its threshold without comment. Lift one end of whatever you are testing so the only possible path is the one you are asking about. If you cannot lift it, switch to ohms and read the number — a path through a winding does not read like a piece of wire.
En la obra
The meter is the one tool you will trust with your hands. Everything below is about making it earn that.
Common mistakes
- Not proving the meter. Live-dead-live, on a known live source, before and after the test. A reading of zero means either the circuit is dead or the meter is — a flat battery, a blown lead fuse, a lead that has failed inside, the dial one click off. Proving it is the only way to know which.
- The red lead left in the current jack. Then used across a live circuit, which is a short through the meter's fuse. Red goes back to VΩ the moment you finish a current reading.
- Continuity on a live circuit. Prove it dead first, every time, no exceptions.
- Resistance measured in circuit. Anything in parallel reads in parallel, and the meter gives no hint that it happened.
- Not naming the reference. "It reads 120" is not a measurement. "Hot to neutral reads 120, hot to bond reads 120, neutral to bond reads nothing" is.
- Trusting a cracked lead. The leads carry their own rating and it is printed on them. Replace a damaged lead; do not tape it. See meter safety and CAT ratings.
What a good habit looks like
- Black lead lives in COM and never moves. Red is the only lead that changes jacks, and only for current.
- One hand where you can manage it, and the body never part of the path.
- The meter matched to where you are standing — the category for the location and a working voltage at or above what you are measuring, leads included.
- Readings written down while the machine is healthy, so there is a baseline the day it is not.
- Say the measurement out loud before you take it: which two points, and what you expect. A surprise then means something, instead of being explained away.
The walk-through, in order
- Decide the question. Which two points, and what would a healthy answer be?
- Set the dial for that question — AC volts, DC volts, ohms, continuity — and check it is what you meant.
- Prove the meter on a source you know is live.
- Take the reading with the leads on the two points you named, not on whatever is convenient.
- Prove the meter again on that known live source. Only now does a reading of zero mean dead.
- For anything cold — continuity, resistance — isolate the piece first by lifting one end of it.
- Write down what you found and against what you compared it. That note is what makes the next visit short.
Preguntas frecuentes
Why does my meter read voltage on a wire I know is disconnected?
That is ghost voltage, and it is the meter doing its job too well. A digital multimeter has a very high input impedance, so it draws almost no current to take a reading. A disconnected conductor running for a long way beside a live one picks up a small capacitively coupled voltage, and a meter that loads the circuit with almost nothing will happily show it. Put any real load across it — a low-impedance (LoZ) setting if the meter has one, or a solenoid tester — and the reading collapses to nothing. That collapse is the proof the conductor is dead. A reading that does not collapse is a live conductor.
Why did my meter beep on continuity through a circuit that is not connected?
Because something else completed the path. A lamp filament, a motor winding, a transformer, a pilot light, even another device further along the run will all read as a path to a continuity tester, because to the meter they are just resistance. Continuity does not tell you what the path is, only that one exists. Isolate the piece you are testing — lift one end of it — before you trust the beep.
Do I measure voltage across something or along it?
Across. Voltage is a difference between two points, so the two leads go on the two points you want the difference between — across the lamp, across the open switch, from the terminal to the bond. Current is the one you measure along a circuit, which is why an in-line current measurement means breaking the circuit and putting the meter in it, and why almost everybody uses a clamp instead.
Ver también
- Meter safety and CAT ratings — whether the meter belongs where you are standing at all.
- Lockout / tagout, walked through — proving dead is half of this page and all of that one.
- Motor control troubleshooting — walking the meter along a rung until the voltage disappears.
- Hot, neutral, ground — the three references every voltage reading is taken against.
- All the on-the-job pages
¿Ve algo incorrecto en esta página? Repórtelo
Gracias — su reporte fue enviado.
No se pudo enviar. Vuelva a intentarlo más tarde.
Avisos de herramientas que le sirven — por correo, opcional
Deje su correo y marque la casilla para recibir un aviso breve, como máximo cada dos semanas, sobre herramientas, ejercicios y diagramas que encajan con lo que ha estado estudiando — aquí y en bentright.com y fieldcalc.net. Es la misma lista para los tres, se sale con un clic y no se envía nada hasta que confirme desde su bandeja de entrada.