Measure, Test, Prove

A meter does not tell you the truth. It answers a very specific question, and most bad measurements are the right answer to the wrong question.

Why this class is here#

The single most valuable professional habit in this entire course is the difference between these two sentences:

"It should be fine."

"I measured it and it reads 231 volts between live and earth at the socket."

Everybody can say the first. The second is what you are paid for. A technical director who measures is worth several who assume, and the gap is entirely a matter of habit rather than talent.

This class is about building that habit and, just as importantly, about understanding what the instrument is actually doing so you know when to disbelieve it.

What each meter mode really does#

A multimeter is not one instrument. It is four or five different instruments sharing a display, and each connects to the circuit in a fundamentally different way. Confusing them is how meters die.

Voltage: a very high resistance placed across two points#

In voltage mode the meter is a large resistance, typically 10 MΩ, that you connect in parallel across the thing you want to know about. It draws a tiny current and reports the pressure difference. Because it is high resistance, it disturbs most circuits very little.

Most of the time. Not always, and the exception is Block C at the bench today.

Current: a very low resistance placed in the path#

In current mode the meter becomes a near short circuit, a fraction of an ohm, that you must insert in series by breaking the circuit and letting the current flow through the meter.

This is the mode that kills meters, blows fuses and occasionally makes a bang, because a near-zero-resistance instrument placed across a supply is a deliberate short circuit.

Learn this as a physical habit, not a fact: when you move the red lead to the current jack, you have changed the instrument into something that must never touch two points at different potentials. Move it back the moment you are done. Many professionals simply never leave a meter in current mode.

Resistance: the meter supplies the current itself#

In resistance mode the meter pushes a small known current through the component and measures the resulting voltage. Two consequences follow immediately, and both are examinable:

1. The circuit must be dead. Any voltage present adds to or fights the meter's own test current, and the reading is meaningless. It can also damage the meter.
2. You are measuring every path between the probes, not the component. A resistor soldered onto a board is in parallel with everything else on that board. This is why you lift one leg, and why in-circuit resistance readings that look wrong usually are not the component's fault.

Continuity: resistance, with a threshold and a buzzer#

Continuity is resistance mode with an opinion. Below some threshold, often 30 to 50 Ω, it beeps.

The beep is a convenience and a trap. It tells you a path exists. It does not tell you the path is good enough to carry current, and it does not tell you the path will still be there under mechanical load or when warm.

A cable that beeps is not a proven cable. A cable that beeps while you flex it along its whole length and at both connectors is much closer to proven. That flex test is the single most useful five seconds in cable maintenance and almost nobody does it.

The four rules of measuring#

These are the rules the bench work today is built on, and they are the ones the capstone will assess you on.

1. Say what you expect before you look#

Write the prediction down. A measurement you have not predicted is a number; a measurement that contradicts a prediction is information. This is not a study technique, it is how you catch a mis-set meter: if you expected 12 V and see 0.012, you notice the range immediately.

2. Record both probes, every time#

"230 V" is not a measurement. "230 V between live and neutral at socket 4, house lights off" is. Six weeks later, in a fault report, only one of those is worth anything.

3. Prove the meter before you trust the result#

Especially before you declare something dead. The procedure is live, dead, live: prove the meter reads correctly on a known live source, take your reading on the circuit under test, then prove the meter again on the known live source. A meter with a flat battery or a broken lead reads zero volts on everything, and zero volts is exactly the answer you were hoping for.

This is not optional professional caution. It is the standard method, and skipping it is how people are killed by circuits they had "tested".

4. A measurement without a written record did not happen#

Your notebook is the deliverable. In the capstone, a correct diagnosis with no measurement log scores below a partial diagnosis with a clear one, because only one of those two is repeatable by somebody else.

When the meter lies#

Three effects that will each cost you an hour, once, if you have not met them.

Ghost voltage#

A 10 MΩ meter connected to a disconnected wire that runs alongside a live one will read a substantial voltage — sometimes 40, 60, over 100 volts. It is real, and it is capacitively coupled through the insulation between the two cables. There is essentially no current behind it.

This matters in two directions. It makes people think a dead circuit is live, which wastes time. And it makes people learn to ignore meter readings, which is far worse. The correct response is to use a low-impedance mode if your meter has one, or a proving unit, and never to develop the habit of dismissing a reading you do not like.

The loading effect#

Ten megohms is a lot, but it is not infinite. Across a high-impedance source — a divider made of 1 MΩ resistors, a piezo sensor, an unbuffered analogue input — the meter itself changes the reading by drawing current. You measure 4 V on a divider that is genuinely sitting at 5 V, and the difference is your instrument.

The rule of thumb: if the source impedance is more than about a hundredth of the meter's input impedance, expect to be reading low. This is exactly the situation in Class 13 with sensor dividers, which is why it is worth meeting now.

Averaging versus true RMS#

A cheap meter measures the average of a rectified AC waveform and multiplies by 1.11, which is the correct factor for a pure sine wave and the wrong factor for everything else. Put it on the output of a phase-controlled dimmer, a switch-mode supply or an LED driver, and it will be wrong, often by twenty to forty per cent, always in a direction you cannot predict.

A true-RMS meter computes the actual heating value regardless of shape. In a theatre, where almost nothing draws a clean sine any more, this is worth the price difference and it is on the kit list for that reason.

Testing a cable properly#

This is the one procedure everybody in this room will use every week for the rest of their working life, so it is worth doing correctly rather than habitually.

For a two-conductor or three-conductor cable, with the cable disconnected from everything:

1. Continuity, pin to pin. Pin 1 to pin 1, pin 2 to pin 2, pin 3 to pin 3. Expect under 1 Ω for anything short, and record the actual number rather than the beep.
2. Isolation, pin to pin. Pin 1 to pin 2, pin 1 to pin 3, pin 2 to pin 3. Expect open circuit. This is the test that catches a whisker of stray strand bridging two pins, which is the fault that makes an audio cable hum and a DMX line drop out.
3. Isolation to shell. Every pin to the connector body. Expect open, unless the design bonds the shield to the shell, which for XLR you should know either way and label.
4. The flex test. Repeat step 1 while flexing the cable at both connector strain reliefs and along its length. This catches the cracked conductor that is intermittently making contact, which is the single most common cable fault and the one that never shows up on a static test.
5. Label it. A cable that has passed and is not labelled will be tested again next week by somebody else. Date and initial.

Steps 2 and 4 are the ones that get skipped, and they are the two that find the faults that waste a technical rehearsal.

Measuring current without breaking the circuit#

Everything above assumes you can open the circuit and insert the meter. On a mains circuit, in a rack, during a show, you cannot.

A clamp meter measures the magnetic field around a conductor and reports the current in it, with no electrical connection and nothing disconnected. It is the only practical way to answer "how much is that circuit actually drawing" on a live installation, and for a technical director it is often the more useful instrument of the two.

Four things to know before you trust one:

1. Clamp one conductor only. Around a whole flex, the live and neutral currents are equal and opposite and cancel: the meter reads zero. That is not a fault, it is the same physics an RCD uses, and it catches everybody once.
2. AC clamps are common, DC clamps are not. A cheap clamp meter reads AC only, because it works by transformer action and a steady field induces nothing. A Hall-effect clamp reads both and costs more.
3. Low currents are where they are worst. Below an amp or so, a general-purpose clamp is guessing. Wind ten turns of the conductor through the jaw and divide the reading by ten: an old trick, and it genuinely works.
4. True RMS matters more here than anywhere, because what you are usually clamping is a dimmer or a switch-mode load, which is exactly the waveform an averaging meter gets wrong.

Proving insulation, which is a different question#

Continuity asks "is there a path". Insulation testing asks the opposite: "is there definitely no path", at a voltage high enough to matter.

An insulation tester applies 250, 500 or 1000 V DC between conductors, or between a conductor and earth, and measures the resistance. Your multimeter applies a couple of volts, and insulation that looks like ten megohms at 2 V can break down completely at 500 V. That gap is the whole reason the instrument exists.

What the numbers mean, roughly:

Two warnings. It applies a real voltage, so nothing sensitive may be connected: semiconductors, surge protection and any electronics will be damaged or will make the reading meaningless. And the cable holds a charge afterwards, so a proper tester discharges it and you check that it has.

You will watch this in Class 6 on your own driver board, and again in Class 15 on a cable.

At the bench#

Nearly three hours. Low voltage throughout, except one demonstration you will watch rather than do.

Block A — Calibrate yourself against your meter (30 min)#

Before you measure anything unknown, find out what your instrument does on things you know.

Measure a fresh AA cell, a known 10 kΩ resistor, a piece of wire, and your own body resistance between two fingers, dry, then with damp fingers. Write all of it down. That last measurement is the one worth keeping: put the two numbers in your notebook and remember them for Class 15.

Block B — Prediction discipline (45 min)#

Ten small circuits on the bench sheet. For each: predict, then measure, then record the gap.

The mark is not for accuracy. It is for whether your prediction was written before your measurement, which is why the sheet has the two columns in that order and the technician will initial the prediction column before you start measuring.

Block C — Make the meter lie to you, deliberately (45 min)#

Three set pieces, and you should reproduce all three yourself:

1. Loading. A 1 MΩ / 1 MΩ divider on 10 V. Predict 5 V. Measure. Explain the gap using the meter's input impedance, with the arithmetic written out.
2. Ghost voltage. A disconnected conductor run alongside an energised one. Measure it on high impedance, then load it with a 100 kΩ resistor and measure again. Watch it collapse.
3. In-circuit resistance. A resistor on a populated board. Measure in circuit. Lift one leg. Measure again. Explain the difference in terms of parallel paths.

Block D — Fifteen cables (30 min)#

Fifteen cables on the bench. Some are good. Some have exactly one of the faults in the list above. Test all fifteen with the five-step procedure, log them, and label them.

Then swap benches and test somebody else's fifteen. The pairs of results will not agree, and the disagreements are the discussion that ends the class.

Watch, do not do — proving dead on a live circuit (15 min)#

The technician demonstrates live-dead-live on a real circuit with a proving unit. You watch. You will do this yourself in Class 15, after the safety systems class, and not before.

Numbers from this class#

Self test with model answers

Answer it yourself first, out loud or on paper, then open the answer. Reading the answer without attempting the question teaches you almost nothing.

1. 1Explain, in terms of what the instrument becomes, why a meter in current mode across a supply is dangerous.
2. 2You measure a divider of two 1 MΩ resistors on 10 V and read 4.0 V rather than 5.0 V. Is the divider faulty?
3. 3Why is "live, dead, live" a procedure rather than a suggestion?
4. 4A cable beeps on continuity. Name two faults it could still have.
5. 5Why will an averaging meter be wrong on a dimmer output, and by how much?

Note. These are model answers, not the only correct ones. If yours differs and you can defend it with the arithmetic and a measurement, that is worth more than matching the wording.