Why Does My Multimeter Resistance Reading Keep Changing?

Settling vs wandering

The way the number moves tells you most of what you need. A reading that climbs steadily in one direction and then stops is settling: something in the measurement is changing in a predictable way, usually a capacitor charging or a part warming or cooling to a steady temperature. A reading that jumps around at random, or changes every time you shift your grip, is wandering: the measurement path itself is unstable, because of poor contact, your body, electrical noise or a live circuit.

The ohms function works by driving a small, known test current through the part and measuring the voltage that appears. On high ranges that current is tiny, so anything that adds or steals even a little current, like a fingertip, a damp board or noise picked up by long leads, shifts the result. On low ranges the current is larger but the resistance being measured is so small that a few hundredths of an ohm of contact change matter. Both ends of the scale are where readings misbehave.

Causes at a glance

Why fingers matter on high ohms

Pinching a resistor lead and the probe tip together with each hand feels natural, and it's the most common reason a high-value reading won't sit still. Your body is a conductor. Hand-to-hand resistance at the low voltages an ohmmeter uses varies enormously with skin moisture, grip area and pressure, from roughly a hundred kilohms to a few megohms. That path sits directly in parallel with the part, so the meter reports the combination, and it changes every time your grip does.

The error depends on the ratio between the part and your body. Using the same illustrative 500 kΩ body path, the parallel formula (R × 500 ÷ (R + 500), in kΩ) gives the readings below. A 1 kΩ resistor reads 998 Ω, an error you'd never notice; a 1 MΩ resistor reads about 333 kΩ, a third of its real value.

Calculated reading as a percentage of the true value, assuming an illustrative 500 kΩ body path. Real body resistance varies widely, so the effect can be larger or smaller.

1Take your hands off the metal

Lay the part on a dry, non-conductive surface and touch the probes to its leads without touching any metal yourself, or clip one lead on with an alligator clip so only one hand is involved. One hand on one probe doesn't create a parallel path; two hands, one on each side, does. If the reading jumps up and holds steady the moment you let go, you've found the cause.

The same idea applies to anything that bridges the two probe points: a damp workbench, flux residue on a board, or a component lying on an anti-static mat, which is deliberately slightly conductive. Above a few hundred kilohms, measure on clean, dry plastic or wood.

2Fix the probe contact

A tip resting lightly on a tarnished lead, a solder joint coated in flux or a plated terminal makes a contact whose resistance changes as you breathe. Press firmly, twist the tip a few degrees to cut through the surface film, and try a fresh spot. Clean old component leads with fine abrasive or a fiberglass pen. Worn or bent probe tips and leads with a cracked conductor inside cause the same symptoms; if wiggling the lead near its plug moves the reading, the leads are the problem. See how to tell if multimeter test leads are bad.

Contact matters most on low values, which is where step 5 picks up. On high values it matters less than leakage and noise, covered in the range section further down.

3Wait out a charging capacitor

If the part is a capacitor, or a circuit with one across it, the meter's test current has to charge it before the reading means anything. The value starts low and climbs; for a capacitor on its own it keeps climbing to OL, and in a circuit it levels off at the resistance of whatever else is connected. Bigger capacitors and higher ohms ranges (which use smaller test currents) take longer, from a blink for a small ceramic to many seconds for a large electrolytic. Reverse the probes and the climb starts again, which is a quick way to confirm capacitance is the cause.

Discharge the capacitor safely, measure, and wait until the reading stops moving. If you want to know whether the capacitor itself is good, measure it properly with how to test a capacitor with a multimeter. Climbing readings caused by capacitors on a populated board, along with the parallel-path errors that come with in-circuit testing, are covered in why resistance values differ on a circuit board.

4Rule out temperature

Some parts are supposed to change. An NTC thermistor's resistance falls as it warms, typically by roughly 3 to 5% per °C near room temperature. As an illustration, a 10 kΩ (at 25 °C) thermistor that drops about 4% per degree reads around 8.2 kΩ after warming 5 °C between your fingertips (10 × 0.96⁵). Photoresistors change with light, so a hand shadow moves the reading too. Metal elements and copper windings move the other way: copper's resistance rises by about 0.4% per °C, so a motor winding measured straight after running reads high and falls as it cools.

Let the part sit at room temperature, handle it with clip leads, and compare the stable value with the maker's table at that temperature. Sensor-specific procedures are in how to test a thermistor and how to test a photoresistor.

5Zero the leads on low values

Below about 10 Ω, the leads and contacts are a real fraction of the answer. A typical pair of test leads plus contacts adds something like 0.1 to 0.3 Ω, and that figure moves with probe pressure. Measuring a 0.5 Ω shunt, a lead total of 0.2 Ω makes the meter show 0.7 Ω, a 40% error, and a change in pressure of 0.05 Ω shifts the reading by another 10%. That's why low readings seem to drift as you hold the probes.

Short the tips firmly, note the value, and subtract it, or press REL (relative or zero) if the meter has it so the display starts from zero. Then use the same firm pressure on the part. The UNI-T UT61E+'s manual specifically tells you to use REL for low-ohm readings. A meter's resolution sets the floor too: the Fluke 117's lowest range is 600.0 Ω with 0.1 Ω resolution, so a value near 0.2 Ω can only ever flick between a couple of digits.

6Stop the range hunting

Autoranging meters switch range when a reading crosses a limit, and a value sitting right on that limit can bounce between two ranges. On the Fluke 117, for example, the two lowest ohms ranges are 600.0 Ω and 6.000 kΩ, so a part near 600 Ω can alternate between something like 599.8 Ω and 0.601 kΩ, and each switch takes a moment to settle. Press RANGE to lock the higher range and the hunting stops.

High ranges are noisy for a different reason. Measuring tens of megohms means a very small test current, so mains hum picked up by long, looped leads, and leakage across a dusty or humid board, are a big share of the signal. Fluke's own spec reflects this: the 117 is rated ±(0.9% + 1) up to 6 MΩ but ±(5% + 2) on its 40.00 MΩ range. Keep leads short and close together, stay clear of power cords, and expect the least stability at the top of the scale.

7Confirm the part is unpowered

Any voltage across the part, from a battery still connected, a charged capacitor or an induced voltage on a long cable, adds to or fights against the meter's test voltage. The result can be a value that drifts with the stray voltage, a negative number or OL. Switch to DC and then AC volts and measure across the same points. If anything more than a few millivolts shows, disconnect and discharge before measuring ohms again.

If the voltage reading itself is what's jumping around, that is a separate problem with its own causes; see what causes unstable multimeter readings.

How much movement is normal?

Every digital meter has an accuracy spec written as a percentage of the reading plus a number of counts, and a reading wobbling inside that window is behaving. Working through the Fluke 117's published spec shows the size of the allowance on a few common values.

A last digit that flicks by one or two counts is ordinary. A reading that moves by more than the allowed spread, keeps moving in one direction, or changes when you touch something is one of the causes above. For more on what a spec line means in practice, see multimeter accuracy for electronics repair.

Meters that help

A relative-zero key, fine low-ohm resolution and a bar graph make drifting readings easier to judge. These picks come from our verified multimeter list.

Meters for steady resistance readings

Fluke 87V

19,999-count high-resolution mode and a 32-segment analog bar graph for watching a value settle. Hands-on Tested.

Check price on AmazonRead the review →

UNI-T UT61E+

REL mode to zero the leads and 0.01 Ω resolution on its 220.00 Ω range, at a budget price.

Check price on AmazonRead the review →

EEVblog Brymen BM786

60,000 counts, 0.01 Ω resolution on its 600.00 Ω range and a relative zero key.

Check price on AmazonRead the review →

As an Amazon Associate I earn from qualifying purchases. Prices and availability are shown on Amazon.