Calibrate an Online Ruler With a Bank Card

The calibration, step by step

Five steps, and the mono line under each one is exactly what the ruler's own state line will read when you get there — these are the engine's strings, not a paraphrase of them.

1. Start from the assumption, and see it named

Open the ruler with nothing stored and it draws itself from the CSS specification's fixed inch: 96 pixels, whatever your panel is. It says so rather than looking finished.

UNCALIBRATED — assumed 96 CSS ppi
2. Find a flat card

Any bank card, ID card or driving licence. A card that has been sat on and bowed is no longer a straight edge, and nothing on the page can see that it is bent — that failure is yours to catch, which is why it is written here.

Reference: bank card, ID card or driving licence
3. Hold it on the outline and drag until the edges agree

The slider moves the outline in steps chosen so that one step is never worth more than 0.35% of the object you are matching — on a card at a typical density that is one pixel, or 0.31%. Its travel is re-derived every time you change edge, because the plausible range is a property of the reference: 129 to 1369 px across the long edge, 81 to 863 across the short one. Match the long edge if you can: the same 0.7 mm of hand slop is 1.59 times smaller a share of 85.60 mm than of 53.98, and the band that comes out says so.

On a phone there is a third option and it is the one to take. The long edge does not fit across a portrait screen — an iPhone 15 is about 64 mm of glass against the card's 85.60 — so stand the card upright and match it down the screen instead. The reference is still the full 85.60 mm and the band is still ±1.2%. The short edge is there for when neither way up fits, and it is the more expensive choice at ±1.7%, not a free one.

324 px wide · steps of 1 px
4. Press the button, and read the band you earned

The readout turns green for this and for nothing else on the site. The number after the comma is not decoration: it is the object's own dimensional spread plus your alignment, added as bounds, and every measurement you take afterwards carries it.

CALIBRATED — bank card, ID card or driving licence long edge, ±1.2%
5. Come back, and let it re-check itself

The stored quantity is device pixels per millimetre, so page zoom divides back out of it on Chrome, Edge and Firefox, and an operating-system display-scaling change divides back out on Windows and on GNOME — the calibration survives those, and the correction is reported rather than hidden. It is not corrected on WebKit, which does not report page zoom through the pixel ratio, and not on a macOS scaled resolution, where the factor lands in the reported pixel count instead; each is caught as a change and re-offered. What survives nothing is a display that reports something new, a pinch held live, or six months of silence. Each of those drops the readout out of green into a neutral state that names what happened.

CALIBRATION MAY BE STALE — your display changed, re-check

The reference objects that are offered

An object becomes selectable only when its dimensional tolerance has been read out of a published standard and recorded here with at least two independent sources and a verification date. The engine applies that floor itself at load: a row flipped on without the evidence is demoted into the withheld list with the missing item named, so this table cannot be widened by editing a flag. Everything below is rendered from that data file, never transcribed.

For a bank card that clause is the ID-1 row of ISO/IEC 7810:2019, Figure 1, under the heading for personalized and returned cards: 85.37 to 85.90 mm across, 53.82 to 54.18 mm tall. That is the row that governs a card in a wallet, because the standard's own definition of an unused card excludes anything that has been through a personalization operation, and every bank card, ID card and licence has been printed, embossed or encoded. The envelope is not symmetric about the nominal size, so the figure quoted is the larger excursion rather than the average of the two — a bound, which is what a ± sign reads as.

For a sheet of A4 it is clause 7.1 of ISO 216:2007, and the band that gets quoted around the web is the wrong one. The ±1.5 mm figure governs dimensions of 150 mm and under; both A4 edges are larger than that, so both fall in the ±2 mm band, and 2 mm is 0.673% of the 297 mm edge and 0.952% of the 210 mm one. One number cannot describe both edges of the same sheet, which is why two are recorded. The 297 mm edge is the one to use: it earns ±1.0%, the tightest band on this page.

A sheet is only ever the second opinion, and that is enforced rather than advised — ask the tool to calibrate from paper and it refuses by name and tells you to match a card first. Paper bows away from a vertical screen, and unlike tolerance and hand slop that bias is systematic and is not inside the ±1.0%. What the sheet contributes is length, and the check is on the next section.

Reference and edge	Nominal	Object tolerance	Alignment slop	Error band	Verified on
A4 sheet, long edge	297.00 mm	0.673%	0.236%	±1.0%	2026-08-09
A4 sheet, short edge	210.00 mm	0.952%	0.333%	±1.3%	2026-08-09
bank card, ID card or driving licence, long edge	85.60 mm	0.351%	0.818%	±1.2%	2026-08-09
bank card, ID card or driving licence, short edge	53.98 mm	0.371%	1.297%	±1.7%	2026-08-09

A4 sheet, 210 × 297 mm (ISO 216 trimmed size)

Offered · verified 2026-08-09 · 3 sources · tier crosscheck

Where the tolerance comes from. ISO 216:2007(E), clause 7 "Tolerances", 7.1 "Permissible tolerances", quoted: "Unless closer tolerances are specified at the time of ordering, the permissible tolerances on the dimensions are as follows: a) for dimensions <= 150 mm: upper limit +1,5 mm / lower limit -1,5 mm; b) for dimensions > 150 mm but <= 600 mm: upper limit +2 mm / lower limit -2 mm; c) for dimensions > 600 mm: upper limit +3 mm / lower limit -3 mm." A4's nominal 210 × 297 mm is Table 1 of the same standard. BOTH A4 edges fall in band (b), because 210 mm and 297 mm are each greater than 150 mm and not greater than 600 mm; band (a) governs A5's 148 mm edge and smaller, not A4. The envelope is symmetric, so the bound is the excursion: long edge 2/297 = 0.67340% -> 0.673%; short edge 2/210 = 0.95238% -> 0.952%. Clause 7.1 calls this the range outside which a sheet cannot be regarded as being of a given size, and says a process tolerance "is likely to be more stringent" — so it bounds the object from above, which is the only safe direction. Clause 7.2 conditions the measurement on the ISO 187 standard atmosphere.

What the band does not mean. What the ±2 mm does NOT mean: clause 7.1 says in its own words that this is not a manufacturing tolerance — it is the range outside which a sheet stops counting as A4. A real ream is almost certainly tighter than this, unknowably so, because mill tolerances are agreed between buyer and supplier and are not published. So this bounds what ISO 216 permits, and it is not a statement about the particular sheet in your hand. Two effects sit OUTSIDE the band as well. Clause 7.2 measures under a controlled atmosphere and paper is hygroscopic: a sheet in an ordinary room moves with the humidity, more across the grain than along it, which is a further tenth to half a percent and is deliberately not folded into the figure. And a sheet held against a vertical screen bows away from the glass, which is why paper is offered here as a second opinion and never as the calibration itself.

Bank/payment card, ID card, driving licence (ISO/IEC 7810 ID-1)

Offered · verified 2026-08-09 · 5 sources · tier primary

Where the tolerance comes from. ISO/IEC 7810:2019(E), Figure 1 "Card size dimensions", block headed "Personalized card and returned card requirements", row ID-1: dimension a (width) max 85,90 / min 85,37; dimension b (height) max 54,18 / min 53,82 mm. Clause 5.2 requires the whole outline to fall between two concentric rectangles at those limits; clause 5.1 conditions them at 23 C ± 3 C and 40-60 % RH. The envelope is asymmetric about the nominal, so the honest bound is the LARGER excursion. Long edge: 85,90 - 85,60 = +0.30 mm against 85,60 - 85,37 = -0.23 mm, so 0.30/85.60 = 0.3505% -> 0.351%. Short edge: 54,18 - 53,98 = +0.20 mm against 53,98 - 53,82 = -0.16 mm, so 0.20/53.98 = 0.3705% -> 0.371%.

What the band does not mean. The band is a CONFORMANCE requirement — what a card must still measure if it were returned for testing — not a measured distribution of worn cards. A card bowed past 85.90 mm is non-conforming, not impossible, so this bounds conformance rather than reality, and it is not a claim about every card in circulation. The limits are also defined only at 23 C ± 3 C and 40-60 % RH, so a card measured in a hot car or in dry winter air is outside the conditions the envelope itself assumes. And clause 5.2 is a two-concentric-rectangles zone the whole outline must sit inside, not a per-edge ± on one measurement, which is exactly why the full endpoint range is the right thing to bound and why the asymmetry must be bounded rather than averaged.

Verified, and deliberately not offered

These two were measured against the same standard, and their numbers are sound. They are withheld anyway, because a reference can fail in a way no error band covers: being confused with a neighbouring object. A tolerance failure is bounded by the figure printed next to it. An identification failure is bounded by nothing, and it arrives wearing the same tight band as a correct pick.

ISO/IEC 7810 ID-000 (mini-SIM / 2FF plug-in card)

Verified 2026-08-09 · withheld · would print ±3.6% on the long edge, ±5.8% on the short edge

Why it is not offered. Verified, deliberately not offered. The band this object earns is ±3.6% on the long edge and ±5.8% on the short — not a calibration. The object a user holds is almost certainly a nano-SIM (12.3 × 8.8 mm, an ETSI form factor with no published ISO tolerance), so a mis-pick is 103% wrong. And the mandated corner bevel means one edge is not a clean straight edge to butt against.

Where its tolerance comes from. ISO/IEC 7810:2019(E) Figure 1, "Personalized card and returned card requirements", row ID-000: a max 25,20 / min 24,80; b max 15,15 / min 14,84 mm. Bound = larger excursion from nominal: long ±0.20 -> 0.20/25 = 0.800%; short +0.15 against -0.16 -> 0.16/15 = 1.0667% -> 1.067%. Clause 5.2 additionally mandates a 3 ± 0,1 mm 45 degree ± 2 degree bevel on one corner.

What is still missing. ETSI TS 102 221 would be the proper second numeric source for SIM geometry. It was NOT read — etsi.org returned HTTP 403 — so nothing here rests on it. Recorded so the gap is visible rather than papered over.

ISO/IEC 7810 ID-2 card (older-style national ID card, visa sticker)

Verified 2026-08-09 · withheld · would print ±1.0% on the long edge, ±1.4% on the short edge

Why it is not offered. Verified, deliberately not offered. ID-1 and ID-2 are both "a card from my wallet", so a mis-pick calibrates 22.7% wrong with every on-screen signal identical — an unbounded identification error dressed in a ±1.0% band. Legacy-only in issue, and the commonest ID-2 object is a visa sticker in a passport, which cannot be laid flat against a screen.

Where its tolerance comes from. ISO/IEC 7810:2019(E) Figure 1, "Personalized card and returned card requirements", row ID-2: a max 105,30 / min 104,80; b max 74,30 / min 73,70 mm. Same clause 5.2 concentric-rectangles rule and clause 5.1 environment as ID-1. Bound = larger excursion from nominal: long +0.30 against -0.20 -> 0.30/105 = 0.2857% -> 0.286%; short ±0.30 -> 0.30/74 = 0.4054% -> 0.406%.

What is still missing. Availability claims (Germany 2010, Romania 2021, Romanian ID-2 valid for travel to 2031) currently rest on a single encyclopaedia source and would need a government registry before any of that reasoning goes on-page.

Carried, dimensioned, and not verified

The nominal sizes of these are not in dispute and are printed below, because they are what a later verification pass will attach a tolerance to. The tolerance column is empty because it is empty in the data: an earlier working figure existed for several of these rows and was deliberately not carried forward, since a small assumed tolerance understates the band while looking exactly like a measured one. Understating a band is the one direction this site treats as a lie rather than an approximation.

Object carried	Nominal	Verified tolerance	Why it is not offered
1 euro coin	23.25 mm	none recorded	Tolerance unverified — needs an EU coin specification plus a second source.
2 euro coin	25.75 mm	none recorded	Tolerance unverified — needs an EU coin specification plus a second source.
10 euro note	127.00 × 67.00 mm	none recorded	Tolerance unverified — needs an ECB technical specification plus a second source.
20 euro note	133.00 × 72.00 mm	none recorded	Tolerance unverified — needs an ECB technical specification plus a second source.
5 euro note	120.00 × 62.00 mm	none recorded	Tolerance unverified — needs an ECB technical specification plus a second source. Euro notes differ by denomination, so there is one row per denomination and never a family row.
50 euro note	140.00 × 77.00 mm	none recorded	Tolerance unverified — needs an ECB technical specification plus a second source.
US Letter sheet	279.40 × 215.90 mm	none recorded	Tolerance unverified, and the honest finding is that no size-defining standard states one. ASME Y14.1 defines the 8.5 × 11 in sheet and is silent on tolerance; ASTM D3460, the cut-size office paper specification, was withdrawn in 2014 with no replacement; and the only live clause found is a government procurement trim tolerance that does not govern retail paper and has no independent corroboration. A4 is offered instead, and it is a better cross-check anyway at 297 mm.
US dime	17.91 mm	none recorded	Tolerance unverified — needs a US Mint specification plus a second source.
US nickel	21.21 mm	none recorded	Tolerance unverified — needs a US Mint specification plus a second source.
US penny	19.05 mm	none recorded	Tolerance unverified — needs a US Mint specification plus a second source.
US quarter	24.26 mm	none recorded	Tolerance unverified — needs a US Mint specification plus a second source. Also a last-resort reference by design: at 24.26 mm the alignment slop alone is 2.9%.
US banknote (any denomination)	155.96 × 66.29 mm	none recorded	Tolerance unverified. The spike entered 0.10% as a placeholder because §6.4 says "tight; worn notes bias" and gives no number; the addendum blocked the row on exactly this. A small assumed figure UNDERSTATES the band, which §4.1 forbids, so the number is not carried forward at all. Needs a BEP or Federal Reserve dimensional specification, and a second source, before it can ship.

The cross-check, and what a disagreement does not mean

Matching a second, independent object and finding that the two agree is the only evidence on this site strong enough to earn the word verified in the readout. Two measurements go in, the difference as a percentage of their mean comes out, agreement within 1% earns the cross-checked state, and a wider disagreement is reported as a number while the first measurement stands. It runs on the ruler, under the card controls: calibrate against a card, then match a sheet of A4 and press the second button. Two measurements of the same object are refused rather than scored, because agreeing with yourself is not evidence of anything.

Here is the tension it ships with, printed rather than smoothed over. A card at ±1.2% and an A4 long edge at ±1.0% can legitimately disagree by 2.1% with both readings correct and inside their own stated bands, against a gate set at 1%. So a disagreement above 1% means the second object did not confirm the first. It never means the first one was wrong, and this page will not say that it does. The gate stays at 1% because a fixed threshold is one you can check and a per-pair threshold is one you have to trust.

The screen-size shortcut, and why it never earns a badge

You probably know your screen's diagonal, and the arithmetic from there looks convincing: multiply the reported resolution by the pixel ratio, take the hypotenuse, divide by the inches, and out comes a pixel density. Marketed diagonals are usually within about half a percent of the truth, so the answer often lands within a percent or two of a card match. The ruler accepts that number and uses it for exactly one thing — moving the outline to roughly where it belongs, so there is less dragging.

It is not a calibration and the tool will not let it become one. The arithmetic assumes the pixel count your browser reports is the panel's native one, and that is false under macOS scaled resolutions, where the framebuffer is not the panel; false in Safari under page zoom, which changes the size of a CSS pixel without reporting it; and false wherever a fingerprinting-resistance mode is rewriting those values on purpose. Once a card has been matched, the two figures are printed side by side — your card says 4.02 px/mm; the screen size you gave implies 3.96 — 1.5% apart — as a sanity check that changes nothing. Try it on a 13.3-inch screen, a 15.6-inch one or a 27-inch monitor.

What is measured, how it is kept, and what it cannot tell you

What is measured. Two quantities, from two different places. The object's dimensions and its permitted spread are read out of a published standard and recorded with their clause and their arithmetic, so the derivation can be checked without fetching the document again. Your alignment is not measured at all — it is bounded, at 0.7 mm across the two edges you line up, one figure applied to every reference so that the differences between references come only from their lengths and their tolerances rather than from a per-object fudge. The band printed on a reading is those two bounds added.

How it is kept. The result of a match is stored as device pixels per millimetre together with a snapshot of what the display was reporting at the time: pixel ratio, native pixel count, browser chrome height, layout scale, pinch scale. On the next visit those are compared against the display as it is then. A pixel-ratio change is arithmetic the stored quantity already absorbs, so it is corrected and the correction is reported. A native pixel count that has changed is not, so it drops out of green and asks for twenty seconds. Nothing is repaired quietly, and a stored record that fails its own schema or bounds is discarded and named rather than trusted.

What it cannot tell you. Whether you picked up the right object. That is the whole argument for the withheld list above, and it is also why a single reference is offered instead of a generous menu. Nor can it tell you that you are looking at the same monitor as last time, if the new one reports the same resolution: that swap is undetectable and can be a third wrong, which is why a calibration is re-offered after about six months even when every signal still agrees. And it cannot see your card bowing, your hand angling the card off the glass, or a curved panel — none of those are inside the band, and inflating the band until they were would leave the number meaning something other than what it says.

Questions

Which card should I hold against the screen?

Any bank card, credit card, ID card or driving licence. They are all the same object to the standard that governs them, ISO/IEC 7810 format ID-1, and they are all nominally 85.60 by 53.98 mm. Use one that is flat: a card that has been sat on and bowed no longer presents a straight edge, and the tool has no way to see that it is bent.

Why are only two reference objects offered?

Because two are all that have been through a two-source verification pass so far: a bank card and an A4 sheet. A reference is offered only when its dimensional tolerance has been read out of a published standard, recorded with at least two independent sources and a date. Banknotes, US Letter sheets and coins are carried with their nominal dimensions and no tolerance, and the tool refuses them by name rather than quietly calibrating from a guess.

An older-style ID card has a better tolerance. Why is it refused?

Its own band would be about 1.0 percent against a bank card's 1.2, so on tolerance alone it wins. It is withheld because a bank card and an older-style ID card are both a card from your wallet, and picking the wrong one calibrates 22.7 percent wrong with every signal on the page identical and no second reference to catch it. A tolerance error is bounded by the number printed next to it; an identification error is not bounded by anything.

Where does the plus or minus 1.2 percent come from?

Two bounds added together. The card may sit up to 0.30 mm above its nominal 85.60 mm long edge under the personalized and returned card row of ISO/IEC 7810:2019, which is 0.351 percent, and lining two edges up by eye is taken as 0.7 mm across the pair, which is 0.818 percent of 85.60 mm. They are added rather than combined in quadrature because a plus-or-minus sign reads as a bound, and a bound is what is being stated.

How often does a calibration need redoing?

Not on a schedule. The page re-checks the display on every visit and on a timer, and it asks you to match again when the display reports something it did not report before, or after about six months without one. What it cannot detect is a swap to a different monitor of the same resolution, so if you have moved the same profile to another screen, calibrate again even though nothing on the page has asked you to.

Go and measure something

This page is the method, and the instruments are in the navigation above. What it carries here is the four ways into the ruler with a choice already made, so the walkthrough above can be followed straight through. Open the ruler as it comes, or start it on the long edge laid flat, on the long edge stood upright — the one a phone in portrait needs — or on the short edge, which is wider by a stated 0.5 of a percentage point. Every one of those links states its reference, edge and orientation, and the ruler refuses any of the three by name rather than falling back to a default it did not announce.

Where the reasoning behind this page lives

Everything above is the procedure. The three pages below are the argument underneath it — what the site claims and where those claims stop, why a browser cannot do this without you, and what each object is actually worth as a reference.

Calibrating an on-screen ruler with a bank card