What the Resistor Tolerance Checker does
This checker takes a nominal resistance, a tolerance and one or more measured values, and tells you the permitted band, how far each reading is from nominal in ohms and per cent, how much of the tolerance it uses, and whether it passes. It also names the preferred value the measurement is actually closest to, which is often more useful than the verdict.
Values can be written the way they are printed on the part or in the schematic: 4700, 4.7k, 4k7, 470R and 1M2 all mean the same thing here, and tolerance can be a percentage or an IEC 60062 letter code.
It is arithmetic on the numbers you enter, done in your browser. It does not know your meter, so a marginal result is worth checking against the meter's own accuracy before anyone touches the board.
How to use it
- Type the nominal value in whatever notation is to hand - 4k7 and 4700 are the same to this page.
- Enter the tolerance as a percentage, or as the IEC 60062 letter printed on the part: F is 1%, G is 2%, J is 5%, K is 10%, M is 20%.
- Paste the measured values, one per line. Readings that cannot be parsed are reported individually rather than silently dropped.
- Choose a preferred-value series to compare against. E24 is the 5% series, E96 the 1% one - a measurement can sit outside its own band but very close to a neighbouring standard value, which usually means the wrong part was fitted rather than a faulty one.
- Open the temperature section to see what the part's own coefficient is worth. On a 1% part a 100 ppm/K coefficient over a 30 degree rise eats a third of the tolerance before anything has gone wrong.
Reading the results
Tolerance used is the figure to watch. A reading at 40% of tolerance is comfortable; one at 95% is inside the band but will not stay there once temperature, ageing and your meter's own error are added.
A marginal result on a low-value part is usually the meter. Below about 10 ohm the test leads themselves are a meaningful fraction of the reading, so null the leads or use a four-wire measurement before believing it.
If every reading from a reel is off in the same direction by a similar amount, that is a nominal-value problem or a meter calibration problem, not a batch of bad parts. Random scatter around nominal is what a healthy batch looks like.
The nearest preferred value is a sanity check on the schematic. If a part marked 4k7 measures closest to 4k75, it is an E96 part in an E24 position - functional, usually, but not what the drawing says.
Worked example: three readings from a 4k7 5% reel
A 4k7 resistor with a 5% tolerance is permitted 4700 x 0.95 = 4465 ohm to 4700 x 1.05 = 4935 ohm, a band 470 ohm wide.
Reading one is 4812 ohm: 112 ohm high, which is +2.383% of nominal and uses 2.383 / 5 = 48% of the tolerance. A clean pass.
Reading two is 4690 ohm: 10 ohm low, -0.213%, 4% of the tolerance. Ideal.
Reading three is 4455 ohm: 245 ohm low, -5.213%, which is outside the band by 10 ohm. It fails - but only just, so before condemning it, check the meter. A meter quoting 0.5% of reading contributes 22 ohm of uncertainty at this value, which is twice the margin by which the part failed.
The mean of the three is 4652 ohm. The nearest E24 value is still 4.7k, 1.0% away; in E96 the nearest is 4.64k. If the whole reel came out near 4650 the parts are probably 4k64 E96 parts, not out-of-tolerance 4k7 ones.
Formulas and scoring rules
- Permitted band
low = nominal x (1 - tolerance/100); high = nominal x (1 + tolerance/100)- Deviation
deviation = measured - nominal; deviationPercent = deviation / nominal x 100- Tolerance used
used = |deviationPercent| / toleranceAbove 0.8 the page calls the reading marginal; above 1.0 it fails.- Preferred values
E48, E96, E192 = round(10^(n/m)) to 3 significant figuresE192 has one official exception at n = 185, where IEC 60063 gives 9.20 rather than the calculated 9.19. E3, E6, E12 and E24 are the published list, not the formula: 2.7, 3.0, 3.3, 3.6, 3.9, 4.3, 4.7 and 8.2 were fixed historically.- Temperature coefficient
shift = nominal x ppm x 1e-6 x (t - 20 C)100 ppm/K on 10 kohm over 30 K is 30 ohm, or 0.3%.
Why E24 is not a rounded logarithm
The preferred-value series are meant to be equally spaced on a log scale: twenty-four steps per decade for E24, each about 10% apart. Round 10^(n/24) to two significant figures and you get 1.0, 1.1, 1.2, 1.3, 1.5, 1.6, 1.8, 2.0, 2.2, 2.4 - and then 2.6, where every resistor ever sold says 2.7.
Eight of the twenty-four values differ from the calculation: 2.7, 3.0, 3.3, 3.6, 3.9, 4.3, 4.7 and 8.2. They were already established before the standards work of the late 1940s, and IEC 63 (now IEC 60063) kept them rather than obsolete the world's stock. Any tool that generates E24 from the formula will quietly disagree with reality at a third of its values, so this page uses the published list.
E48, E96 and E192 came later and do match the formula to three significant figures, with the single exception of E192's 9.20.
Tolerance is not accuracy
The tolerance printed on a resistor is a manufacturing limit at a reference temperature, usually 20 or 25 C, at the moment it left the factory. It says nothing about drift with temperature, ageing, humidity, soldering stress or self-heating, all of which are specified separately and all of which are additional.
That is why the temperature section here matters. A 1% part with a 100 ppm/K coefficient is already permitted to be 1% out before it warms up; running 30 degrees above the reference adds another 0.3%. Precision work uses parts specified in ppm/K precisely because the tolerance alone does not describe them.
Limitations: what the result does not prove
- It compares numbers. It cannot see the part, the board or the meter, so it cannot tell a genuinely out-of-tolerance resistor from a measurement error, a parallel path on the board or unnulled test leads.
- Your meter's own accuracy is not included. On a 1% part a 0.5% meter uses half the tolerance before the resistor does anything; add its specification to any marginal verdict.
- In-circuit measurements are almost always wrong, because other components sit in parallel with the part. Lift one leg or measure out of circuit.
- The temperature figure uses a simple linear coefficient. Real parts have a curve, and the specified ppm/K is usually a worst case over a range rather than a slope at a point.
Privacy: where your data goes
Everything you paste, type or drop is processed in this browser tab. It is not uploaded, logged, stored or sent to analytics. Session recording and tag-manager scripts are switched off on this page.
Standards and sources
- NIST Technical Note 1297 - Evaluating and expressing measurement uncertainty - checked 19 Sep 2026
- IEC 60063 - Preferred number series for resistors and capacitors
- IEC 60062 - Marking codes for resistors and capacitors
Frequently asked questions
Is a 4812 ohm reading within tolerance for a 4k7 5% resistor?
Yes. The band is 4465 to 4935 ohm, so 4812 ohm is 112 ohm or 2.38% high and uses about 48% of the permitted tolerance. It is a comfortable pass rather than a marginal one.
What do the letters after a resistor value mean?
They are IEC 60062 tolerance codes: B is 0.1%, C is 0.25%, D is 0.5%, F is 1%, G is 2%, J is 5%, K is 10% and M is 20%. The letters R, K and M inside a value mean something different - they mark the decimal point, so 4k7 is 4.7 kilohms and 0R1 is 0.1 ohm.
Why does my resistor measure differently in circuit?
Because you are measuring it in parallel with everything else connected to those two nodes, which can only read lower than the part itself. Lift one leg or measure out of circuit; an in-circuit reading that is lower than nominal is usually the board, not the resistor.
What is the nearest standard resistor value?
That depends on the series. For a measured 4812 ohm the nearest E24 value is 4.7k and the nearest E96 value is 4.87k. This page shows both the nearest value and how far away it is, so you can tell a mis-stuffed part from a drifted one.
Does temperature change a resistor's value enough to matter?
On a precision part, yes. A 100 ppm/K coefficient means 0.01% per degree, so a 30 degree rise is 0.3% - a third of the tolerance of a 1% part, and three times the whole tolerance of a 0.1% one. General-purpose thick-film parts can be 200 to 250 ppm/K.
Why are 5% resistors only made in E24 values?
Because the series steps are sized to the tolerance. E24's steps are about 10% apart, so consecutive 5% bands just touch: every possible resistance is covered exactly once. A 1% part needs the finer E96 steps for the same reason, which is why E96 values look so arbitrary next to E24 ones.
How do I measure a resistance below 1 ohm accurately?
Not with a two-terminal handheld meter, whose test leads are typically 0.2 to 0.5 ohm on their own. Null the leads if the meter offers it, or use a four-wire (Kelvin) measurement, where separate current and sense pairs keep the lead resistance out of the reading entirely.
The whole batch reads high. Are they all faulty?
Almost certainly not. Parts from one reel come from one production lot and drift together, so a consistent offset points at the meter, the nominal value you entered, or a different part number than the drawing says. Scatter in both directions is what a normal batch looks like.
Last reviewed by the A2Z.Tools team against the sources listed above.
