Six multimeter measurements that fix 90 % of electrical faults

An electrical fault nearly always gets sorted the same way: somebody replaces the battery, then the regulator, then the starter relay, and finally finds a green earth eyelet under the engine. Three parts bought, two weekends lost, when the measurement that pointed at the culprit takes forty seconds.

The problem is not the tool, it is the method. A multimeter always displays a number, even when you are holding it completely wrong, and a displayed number looks like an answer. These six measurements cover the bulk of 12 V circuit faults, on a motorcycle, on a tractor or on a generator. For each one: where to put the probes, what state the circuit has to be in, and what order of magnitude to expect.

The meter, and the three mistakes that kill it

For £25 to £70 you get everything you need: DC volts, resistance, a continuity buzzer, and a fused 10 A current input. Auto-ranging removes one more source of error, and a pair of crocodile clips gives you a hand back, which changes everything when you have to crank a starter at the same time.

• Measuring resistance on a live circuit. The ohmmeter injects its own current. If it meets another one coming the other way, the reading means nothing and the meter may not survive it. Resistance and continuity are measured with the battery disconnected.
• Leaving the leads in the 10 A socket to measure a voltage. That socket is very nearly zero resistance: put it across a battery and you have made a dead short. Move the red lead back to the voltage socket as soon as you have finished a current measurement.
• Trusting the continuity buzzer on a power circuit. A wire with one strand left out of nineteen buzzes perfectly well: it will light a bulb and collapse the moment a starter pulls on it. The buzzer says there is a path, not that current can get through it.

1. Resting voltage: state of charge, not state of health

Circuit: ignition off, at least two hours after a charge or a ride. Selector: DC volts, 20 V range. Probes: black on the negative post, red on the positive, directly on the posts and not on the terminal clamps.

If you have just unplugged a charger, surface charge skews the reading upwards: switch the lights on for thirty seconds, then measure. With a lithium iron phosphate battery this table does not apply: resting voltage there sits closer to 13.2 to 13.4 V when full, on a curve so flat it tells you almost nothing about the charge left.

Remember what this measurement cannot do: it gives a state of charge, not a state of health. A battery at the end of its life reads 12.7 V at rest and collapses the moment you ask it for current.

2. Cranking voltage: the real test

Circuit: under load, engine turned over on the starter, ignition killed with the engine stop switch so it does not fire up. Probes: on the battery posts, crocodile clips fitted. What you read: the lowest value during the first two seconds.

On a healthy lead-acid battery at around 20 °C, the voltage must not drop below roughly 9.6 V while cranking. Between 9.6 and 10.5 V you are in normal territory. Below 9.5 V the battery no longer holds a charge, whatever it was showing at rest. In hard frost the limit falls and the current demanded rises: that is why a marginal battery survives the autumn and dies in January.

The opposite case is the most instructive: the voltage stays at 12.3 V, the starter does not turn, it only clicks. The battery is not to blame, the current is not reaching the starter, and measurement 4 will find out where it stops. The other causes are reviewed in diagnosing a motorcycle that will not start.

3. Charging voltage: alternator and regulator

Circuit: engine running, warm, at mid revs, reckon 3000 to 4000 rpm on a motorcycle engine. Probes: on the battery posts.

Correct charging sits between 13.8 and 14.6 V on a circuit designed for lead-acid. Three abnormal readings, three diagnoses:

• Below 13.2 V at the test revs: the battery is not being recharged enough. Blame the stator, the regulator, or more often a burnt power connector between the two. Plenty of machines barely charge at idle: that is normal, take the reading with the revs up.
• Above 15.0 V: the regulator is no longer regulating. Stop the engine before you boil the battery and cook bulbs and control units. This is the fault to deal with immediately, it is expensive in collateral damage.
• Voltage that climbs then falls back when you switch on the lights and the fan: not enough charging capacity, often a stator that has lost a phase.

With lithium, check the charging window the pack accepts: it is not always the same as the one intended for lead-acid.

4. Voltage drop: the measurement nobody makes

This is the most powerful of the six, and the least practised, because it is counter-intuitive: the meter does not go at the end of the circuit but in parallel across the suspect section, while current is flowing.

Circuit: under load, with the consumer working (starter cranking, headlight on, fan running). Selector: DC volts, 20 V range, or 2 V if the meter allows. Probes: one at each end of the section being tested. For a cable, one at each end; for an earth, one on the battery negative post and the other on the metal part concerned; for a connector, one either side — back-probing the terminals is the cleanest way to do it without unplugging anything.

What you read is the voltage lost in that piece of circuit. The usual tolerances:

• A supply cable: less than 0.2 V.
• An earth connection: less than 0.1 V. The strictest figure, and the one that causes the most trouble in practice.
• A connector or a switch: less than 0.1 V across it.
• The complete starter circuit while cranking: less than 0.5 V on the positive side, less than 0.2 V on the earth side.

An earth losing 0.8 V means a yellow headlight, a lazy starter and a regulator compensating by pushing its voltage up. The brightness of a filament bulb falls far faster than the voltage feeding it, hence the immediate effect through the loom. And a parasitic resistance dissipates heat: the connector losing 0.8 V is the one that will melt in six months. Work along the circuit section by section, from the battery towards the consumer, until you find the one swallowing the volts: you will never have to strip a loom.

5. Continuity and resistance: dead circuit, isolated component

Circuit: dead, battery disconnected. Component: unplugged from the rest of the loom, otherwise a parallel path falsifies the reading.

Before any low-value measurement, touch the probes together and note the display: 0.1 to 0.3 Ω typically, that is the resistance of your leads. Subtract it, otherwise measuring a 0.5 Ω winding is meaningless.

• An ignition coil: primary of the order of a few tenths to a few ohms, secondary a few thousand to a few tens of thousands of ohms. These orders of magnitude only tell you whether you are in the right universe; the figure that decides is the one in the manual, measured between the terminals it names. A coil can be good cold and cut out hot: repeat the measurement with the engine hot if the symptom appears hot. The rest of the circuit is in testing an ignition harness without an oscilloscope.
• A temperature sensor: its resistance depends on temperature, so a value with no temperature attached means nothing. Measure at a known temperature, warm water and a thermometer, and compare with the manufacturer's table.
• A stator: measure between phases, then each phase to earth. Between phases, the three values must be close to one another. To earth, the meter must show an open circuit: the slightest continuity to earth condemns the stator.

6. Parasitic draw: the battery that goes flat in the garage

Circuit: everything off, key out, the meter in series in the earth line. Disconnect the negative terminal, put the black lead on the negative battery post and the red lead on the disconnected terminal: the vehicle's current now runs through the meter. Selector: DC current, red lead in the 10 A socket, moving down to the milliamp range once you know the value is small.

Two absolute rules: never crank the starter with the meter connected like this, and wait for the control units to go to sleep, 10 to 30 minutes depending on the machine, otherwise you are measuring a transient standby state.

The orders of magnitude: an older motorcycle with no permanent electronics should stay under a milliamp. A recent machine with a control unit, a clock and an immobiliser routinely draws a few milliamps, and an alarm takes the whole thing up towards 10 to 20 mA. Beyond thirty milliamps or so, you flatten a 10 Ah battery in one to two weeks standing: that is exactly the symptom "it is always flat when I want to go out".

To find the culprit, pull the fuses one at a time while watching the display: the one that makes the value drop identifies the circuit. If the draw disappears when you unplug the regulator, it is a leaking diode; if it goes when you unplug an added accessory, you have found your Sunday-afternoon installer.

The order to follow when nothing starts

Do not go off in all directions. This order eliminates causes in the direction of current flow, and each step takes a minute.

With no meter to hand there are rougher methods that give an indication: they are in diagnosing a motorcycle battery without a multimeter. But buy the multimeter, it pays for itself with the first part you do not replace for nothing.

For the figures specific to your machine, coil resistance, sensor resistance table, charging window and earth point locations, L'Atelier's technical assistant pulls the manufacturer data from the make, the model and the year: you avoid comparing your reading with a number found on a forum for a different generation.