Why the test works
A multimeter set to volts reads the potential difference between its two probes. Put both probes on the same conductor at different points, and what you read is the voltage lost in the length of conductor between them.
A perfect conductor would read zero. A real one reads I × R for the section between the probes. So the meter is measuring resistance indirectly, but at the current the circuit actually carries, rather than at the trickle a resistance range uses.
This is why the test finds problems that other methods miss. A loose lug, a corroded crimp or a burnt switch contact might be a fraction of an ohm. Across the whole run that fraction is invisible next to the conductor’s own resistance. But under load, that fraction of an ohm produces a localised drop you can put a probe on either side of.
Why not just measure resistance
Three reasons, and each of them is enough on its own.
| Resistance range | Voltage drop test | |
|---|---|---|
| Test current | Under 1 mA | The circuit’s actual load current |
| Circuit state | Must be de-energised and often disconnected | Energised and loaded, tested as installed |
| Resolution | Typically 0.1 Ω: the same order as a whole run | Millivolts, which resolves a single joint |
| Finds a marginal joint | Often not: it looks fine at 1 mA | Yes: that is what it is for |
The middle row is the practical one. A connection that measures 0.05 Ω looks perfect on an ohmmeter. At 30 A it drops 1.5 V and dissipates 45 W in a space the size of a fingernail. That is a connection on its way to becoming a fire, and only the loaded test shows it.
The procedure
- Establish what current the circuit will draw. Use a clamp meter on one conductor. Without knowing the current you cannot compare the result with a calculation, and you cannot tell a heavily loaded good circuit from a lightly loaded bad one.
- Set the meter to volts, AC or DC to match the circuit, on a range that resolves tenths of a volt or better.
- Turn the load on and let it settle. Motors draw inrush current for the first moments; a heater’s element changes resistance as it warms. Read once the current has stabilised.
- Place the probes at the two ends of the section under test, on the same conductor. For a whole run that means the source terminal and the load terminal. Test the supply conductor and the return conductor separately. A problem on the neutral or the negative is just as real and is easy to miss.
- Read the meter while the load is running. That number is the voltage drop across that section, in volts.
- Convert to a percentage by dividing by the source voltage and multiplying by 100, so it can be compared with a design target.
Reading the result
| What you measured | Reading | What it suggests |
|---|---|---|
| Whole circuit under load | Close to calculated | Conductors and connections are behaving as designed |
| Whole circuit under load | Well above calculated | Something localised: start halving the run to find it |
| Across one connection | Under 0.1 V | Normal for a sound mains connection |
| Across one connection | Several tenths of a volt | Loose, corroded or undersized termination |
| Across a switch or breaker | Above about 0.2 V | Contact degradation: the device is heating itself |
Compare against a calculation rather than against a feeling. Run the circuit through the calculator first, then measure. If the measurement is a little above the calculation, that is the terminations doing what terminations do. If it is double, there is a fault to find.
Isolating the bad section
When the whole-run reading is high, binary search finds the culprit faster than guessing. Put one probe at the source and move the other to the midpoint of the run. If most of the drop is in that half, the problem is behind you; if the reading is small, the problem is ahead. Halve again and repeat.
Each measurement point is usually a junction box, a device or a splice, which is convenient, because those are also the places faults live. A high reading across a 12-inch length of conductor with a splice in the middle of it is not a conductor problem.
The automotive version
Vehicle electrical work leans on this test more than almost any other, because a 12 V system has no headroom. A 0.5 V drop is 4% of the supply, and a starter circuit carrying 200 A turns a milliohm of connection resistance into 0.2 V and 40 W.
The working figures technicians use are tighter than mains practice, because the supply is small:
| Section | Working maximum |
|---|---|
| One connection or terminal | 0.05 V |
| A length of wire | 0.20 V |
| A switch or relay contact | 0.30 V |
| A ground path | 0.10 V |
| Whole circuit, load side | 0.50 V |
These are widely used shop guidelines rather than a published standard. Where the vehicle manufacturer gives its own figures for a circuit, those take precedence.
The ground path is the one most often skipped and most often at fault. A vehicle chassis is a return conductor with bolted joints, paint and corrosion in it, and it ages worse than the wiring does.
Questions
How do you test for voltage drop with a multimeter?
Set the meter to DC or AC volts as appropriate, put one probe at each end of the section you want to test (same conductor, same polarity), and turn the load on. The meter reads the voltage lost across that section. The load must be running: with no current flowing there is no drop and the meter reads near zero regardless of how bad the connection is.
Why measure voltage drop instead of resistance?
Because a multimeter measures resistance using a tiny test current, typically under a milliamp, and a corroded or loose connection can look fine at that current while failing badly at 20 A. A voltage drop test loads the connection with the current it actually has to carry, which is the condition where a bad joint reveals itself. It also requires no disconnection, so the circuit is tested as installed.
What is an acceptable voltage drop reading on a test?
It depends what you are measuring. Across a whole branch circuit under full load, 3% of the supply voltage is the usual design target and 5% the usual outer limit. Across a single connection (a terminal, a splice, a switch contact), anything above about 0.1 V on a mains circuit or 0.05 V in automotive work points at a problem, because a good connection should drop almost nothing.
Why is my measured voltage drop higher than my calculation?
A calculation models an ideal conductor and nothing else. The real circuit has resistance at every termination, splice, breaker, switch and device, none of which appear in the formula. It may also be running hotter than the 75 °C the reference data assumes, and carrying a different current from the one you assumed. Somewhat higher is normal. Far higher, twice the calculated figure or more, usually means a specific bad connection rather than an undersized conductor.
Can I measure voltage drop without a load?
No. Voltage drop is proportional to current, so with no load there is no drop and the far end reads full supply voltage. This is exactly why an outlet can measure a perfect 120 V with nothing plugged in and still sag badly under load. The load is not incidental to the test. It is the test.
Before you measure, calculate what you expect: the voltage drop calculator gives you the figure to compare against, and how to calculate voltage drop walks through the method.