Root Mean Square Calculator — RMS of a data series

The root mean square calculator above takes a series of numbers, squares each one, averages the squares, and takes the square root of that average. Before going further it is worth clearing up a naming collision that sends a lot of people to the wrong page: this is not a square root calculator. Finding the square root of a single number — the value that multiplies by itself to give 49, say — is handled by our square root calculator. Root mean square is a three-step operation performed on a whole data set, and it produces a kind of average, not a root. The words overlap; the mathematics does not.

Arb Digital publishes this alongside its other statistics tools because RMS is the correct summary in a specific and common situation: when your values swing above and below zero and a plain average would cancel them out to nothing. An alternating voltage, a vibration signal, a set of forecast errors that are sometimes positive and sometimes negative — take the arithmetic mean of any of those and you get close to zero, which tells you nothing about their size. RMS answers the question the plain average cannot: how big are these values typically, ignoring which way they point?

What This Root Mean Square Calculator Does

Paste or type a series in the box, separated by commas, spaces or line breaks, and the calculator reports the root mean square as the headline figure. The supporting grid shows the sum of the squared values, the mean square (that sum divided by the count), the ordinary arithmetic mean for comparison, and the crest factor — the largest absolute value divided by the RMS, which tells you how spiky the series is relative to its typical magnitude.

The working panel prints the substituted formula with your actual numbers, so you see each square and the arithmetic that produced the answer rather than a bare result. That makes the page usable for homework and for checking a spreadsheet formula you do not fully trust.

The mode selector switches the divisor between n and n − 1. Use the complete data set option, which divides by n, for the standard RMS of a signal or a full population. The sample option divides by n − 1 and is what you want when your numbers are a sample drawn from a larger set and you are computing a quantity analogous to the sample standard deviation. For most physical and engineering uses, the complete data set option is correct.

How to Use It

1. Enter your values. Any mix of commas, spaces and new lines works, so you can paste a column straight out of a spreadsheet.
2. Include negatives as they are. Do not strip the minus signs first — squaring handles the sign, and removing it manually changes nothing but wastes your time.
3. Choose the divisor. Complete data set for a full series; sample for a subset of a larger population.
4. Set the decimal places to match the precision of your source measurements.
5. Compare RMS to the arithmetic mean in the grid. A large gap between them means your values vary widely in magnitude.

The Formula and How It's Calculated

The formula reads exactly as its name does, backwards: RMS = √( (x₁² + x₂² + … + xₙ²) ÷ n ). Square every value, take the mean of those squares, then take the square root to get back to the original units.

Using the default series of 3, 4, 5, 6 and 8: the squares are 9, 16, 25, 36 and 64, which sum to 150. Dividing by the count of 5 gives a mean square of 30. The square root of 30 is 5.4772. So RMS = √((9 + 16 + 25 + 36 + 64) ÷ 5) = √(150 ÷ 5) = √30 = 5.4772. The arithmetic mean of the same series is 26 ÷ 5 = 5.2, slightly lower, and the crest factor is 8 ÷ 5.4772 = 1.4606.

That gap between 5.4772 and 5.2 is not an error, and it is not optional. Squaring gives disproportionate weight to larger values, so the RMS always sits at or above the mean of the absolute values. The two coincide only when every value has identical magnitude. The wider the spread, the further RMS sits above the plain average — which is precisely why RMS is used where large excursions matter more than small ones. MathWorld's entry on the root-mean-square gives the formal definition and its continuous integral form.

Why Signals Are Measured in RMS and Not Average

Mains electricity is the clearest example. A household supply described as 230 volts is not 230 volts at every instant; it is a sine wave swinging between roughly +325 V and −325 V fifty or sixty times a second. Its arithmetic mean over a full cycle is zero, because the positive and negative halves cancel exactly. That zero is arithmetically correct and physically absurd, since the supply plainly delivers energy.

RMS solves this by squaring first, which makes every half-cycle positive, and it happens to produce the number engineers actually need: the RMS voltage of an alternating supply is the equivalent steady voltage that would deliver the same heating power to a resistor. That equivalence is why RMS is sometimes called the effective value. For a pure sine wave the RMS is the peak divided by the square root of two, so a 325 V peak corresponds to about 230 V RMS. The NIST reference on SI units covers the underlying quantities if you need to be precise about definitions.

The same logic applies well outside electrical work. Vibration amplitude, sound pressure level, sea state and wind turbulence are all quoted as RMS values for exactly this reason: the sign of the fluctuation carries no useful information, but its magnitude carries all of it.

RMS Error and How It Differs From Plain RMS

The most common place non-engineers meet RMS is in the phrase root mean square error, usually shortened to RMSE. It is the same calculation applied to a series of differences rather than to raw values: for each observation, subtract the predicted value from the actual value, then take the RMS of those residuals. The result is a single number in the original units expressing how far a model's predictions typically fall from reality.

RMSE is preferred to mean absolute error whenever big misses are disproportionately costly, because squaring penalises them harder. A forecast that is off by 10 once scores worse under RMSE than one that is off by 2 five times, even though the total error is identical. Whether that is the behaviour you want is a genuine modelling decision, not a technicality. If every unit of error costs the same regardless of size, mean absolute error is the more honest summary. If a single large miss causes real damage, RMSE is right. Our linear regression calculator fits the model whose residuals you would then measure this way.

Crest Factor: The Number the RMS Hides

RMS deliberately compresses a series into one typical magnitude, and in doing so it hides the peaks. Two very different series can share an RMS: a steady signal sitting near its RMS the whole time, and a mostly quiet signal punctuated by violent spikes. For anything that can be damaged by a peak — an amplifier, a structural component, a power supply — that distinction is the whole engineering problem.

Crest factor is the standard way to expose it: the largest absolute value divided by the RMS. A pure sine wave has a crest factor of about 1.414. A steady direct current has a crest factor of exactly 1. A spiky, impulsive signal can have a crest factor of 5 or more, meaning its peaks are five times the level its RMS suggests. The default series here gives 1.4606, close to sinusoidal. Always look at crest factor alongside RMS when peaks matter, and use our standard deviation calculator when you want the spread around the mean rather than the magnitude around zero.

How RMS Relates to Standard Deviation

These two quantities are closer than they look. Standard deviation is the RMS of the deviations from the mean; RMS is the standard deviation you would get if the mean were zero. That gives a clean identity: mean square equals the square of the mean plus the variance. For the default series, the mean is 5.2 so the mean squared is 27.04, and the population variance is 2.96, and 27.04 + 2.96 = 30, exactly the mean square computed above.

That relationship is a useful check and a useful intuition. When a series is centred on zero, RMS and standard deviation are the same number. When a series has a large mean relative to its scatter, RMS is dominated by the mean and tells you little about variability. Knowing which regime you are in stops you from quoting an RMS where a standard deviation was wanted. Our variance calculator and mean, median and mode calculator give you the other half of that picture.

Common Mistakes to Avoid

• Confusing RMS with a square root. A square root acts on one number; RMS acts on a whole series and returns an average-like value in the same units as the data.
• Averaging first and squaring after. The order is square, then mean, then root. Reversing the first two steps gives the mean, not the RMS.
• Stripping minus signs before entering data. Unnecessary, since squaring already removes the sign, and it invites transcription errors.
• Quoting RMS where a peak value is what matters. Equipment fails on peaks, not on averages. Check the crest factor alongside the RMS.
• Mixing RMS and peak figures in the same comparison. A 325 V peak and a 230 V RMS describe the same supply; comparing one against the other suggests a difference that does not exist.

Related Free Tools From Arb Digital

For the root of a single number, use the square root calculator or the cube root calculator — different maths from this page entirely. For spread around the mean, use the standard deviation calculator and the variance calculator. The mean, median and mode calculator covers the ordinary averages, and the geometric mean calculator handles growth rates where the arithmetic mean misleads. Browse the full free online tools hub for more.

Frequently Asked Questions