ResultThree readings of one dimension
| Reading | Value | What it means |
|---|
Conversion chart
| Millimetres | Decimal inches | Fraction | Fraction error |
|---|
How much precision does the job need?
Every fraction is a rounding decision. This is what each one costs in millimetres, set against the ISO 2768 general tolerance bands that sit in a drawing title block.
| Rounding you choose | One step | Worst case error | Against ISO 2768 medium, 6 to 30 mm |
|---|---|---|---|
| Nearest 1/8 in | 3.175 mm | up to 1.5875 mm | Nearly 8 times the band |
| Nearest 1/16 in | 1.5875 mm | up to 0.79375 mm | Almost 4 times the band |
| Nearest 1/32 in | 0.79375 mm | up to 0.3969 mm | Still outside the band |
| Nearest 1/64 in | 0.396875 mm | up to 0.1984 mm | Just inside the band |
| Decimal inches, 3 places | 0.001 in | up to 0.0127 mm | Machine shop territory |
| Decimal inches, 4 places | 0.0001 in | up to 0.00127 mm | Finer than most shops measure |
ISO 2768 general tolerance bands for reference: fine class allows plus or minus 0.1 mm from 6 to 30 mm, medium allows plus or minus 0.2 mm over the same span and plus or minus 0.3 mm from 30 to 120 mm, and coarse allows plus or minus 1.2 mm from 120 to 400 mm. The title block writes the two classes together, so ISO 2768-mK means medium linear and class K geometric.
Millimetres or fractional inches at the bench
The same length, two measuring habits. Which one you work in decides how the number should be rounded and which tools will fit it.
Working in millimetres
- Drawing default. On an ISO style drawing a dimension with no unit symbol is millimetres, so nothing has to be converted at all.
- Steps of 0.5 mm. Metric twist drills climb in 0.5 mm steps, and 0.1 mm steps in the small sizes, so the size you want usually exists.
- Tolerance in the same unit. ISO 2768 bands are quoted in millimetres, so the dimension and its allowance never need converting against each other.
- Thread pitch is a length. An M8 coarse thread is a 1.25 mm pitch, a distance you can measure with the same rule you used on the diameter.
- Sheet is sold by thickness. A 1.5 mm sheet is 1.5 mm in every material, with no gauge number to decode.
Working in fractional inches
- Marked in sixteenths. A tape or rule is divided in halves down to 1/16, so a decimal answer has to become a fraction before it can be marked out.
- Steps of 1/64 in. Imperial drill sets climb in 1/64 in steps, which is 0.396875 mm, so metric and imperial sets interleave and rarely coincide.
- Tolerance in thou. A machine shop works in thousandths of an inch, where one thou is 0.0254 mm, and quotes the allowance separately from the fraction.
- Thread pitch is a count. A 5/16 in UNC thread is 18 threads per inch, which is a 1.411 mm pitch once converted, so it cannot be compared to a metric pitch by eye.
- Sheet is sold by gauge. Gauge is a sheet number, not a length, and the same number is a different thickness in steel, stainless and aluminium.
The practical rule: convert once, at the point where the work changes hands, and carry the unrounded figure everywhere else. A dimension that passes through two roundings can drift further than the tolerance it was drawn with.
MM to inches for drawings, drills and sheet
The exact 25.4 rule, what each fraction costs you in millimetres, and the sizes that only look interchangeable.
A millimetre figure almost always arrives attached to a job: a dimension read off an engineering drawing, a drill you are about to pick up, a plate thickness a stockist quoted, a thread you need to match. This mm to inches converter answers all three questions that job asks at once, giving decimal inches, the nearest fraction on a rule, and the one number nobody else prints, which is how far that fraction misses the true size in millimetres.
Why 25.4 is exact, and what that buys you
One inch is exactly 25.4 millimetres. That is a definition, not a measurement. The international yard and pound agreement, signed on 1 July 1959 by six nations (the United States, the United Kingdom, Canada, Australia, New Zealand and South Africa), fixed the yard at exactly 0.9144 metres, and the inch follows from it at exactly 25.4 mm. Two engineering bodies had already jumped: ANSI adopted the 25.4 mm inch in 1933 and NACA followed in 1952, so the drawings were metric ready long before the treaty caught up.
Before 1959 the United States worked to an inch of roughly 25.4000508 mm. The difference is 0.0000508 mm per inch, invisible on a bench and significant only over survey distances, but it explains why very old American drawings and gauge blocks can disagree with a modern conversion in the last decimal place.
The practical consequence matters more than the history. Because the conversion factor is exact, the arithmetic adds no error at all. Divide by 25.4 to go from millimetres to inches, multiply by 25.4 to come back, and the answer is perfect until you round it. Every millimetre of error on this page is therefore a decision you made, not a limitation of the maths, and that is the reason the converter reports the size of it.
Decimal inches, fractional inches, and what each rounding costs
A decimal inch and a fractional inch are two different tools. A machine shop quotes thousandths of an inch, where one thou is 0.001 in, or 0.0254 mm. A tape measure or steel rule is marked in halves down to sixteenths, so a decimal answer has to be pushed onto the nearest mark before anyone can scribe it. Pushing it onto a mark throws away length, and the amount thrown away is fixed by the denominator you chose.
Every source that offers fractions warns that the answer will be slightly off. Here is what slightly means, in millimetres:
| Round to | One step is | Worst case error | Typical use |
|---|---|---|---|
| Nearest 1/8 in | 3.175 mm | up to 1.5875 mm | Rough carpentry and site marking |
| Nearest 1/16 in | 1.5875 mm | up to 0.79375 mm | Tape measure work, joinery |
| Nearest 1/32 in | 0.79375 mm | up to 0.3969 mm | Fine woodwork, sheet layout |
| Nearest 1/64 in | 0.396875 mm | up to 0.1984 mm | Imperial drill sizes, close fitting |
| Decimal, 3 places | 0.001 in | up to 0.0127 mm | Machining, one thou resolution |
| Decimal, 4 places | 0.0001 in | up to 0.00127 mm | Inspection and CNC offsets |
The worst case is always half a step, because the true length can fall exactly between two marks. Set that against the general tolerance a drawing already carries and the choice stops being a matter of taste. Under ISO 2768 medium class, a feature between 6 and 30 mm is allowed plus or minus 0.2 mm. A 1/32 in fraction can be 0.3969 mm out, so it is already outside general tolerance before anyone cuts anything, and 1/16 is nowhere near. Only 1/64, at 0.1984 mm, squeezes inside, and it uses almost the entire allowance doing it.
A worked example, with a rounding decision at every step
A drawing calls up a 38.5 mm width. The title block reads ISO 2768-m, so at that size the general tolerance is plus or minus 0.3 mm, because the medium class allows that from 30 to 120 mm. The shop works in inches. What number goes on the setup sheet?
Start with the exact division: 38.5 divided by 25.4 is 1.5157480315 in. Now look at the candidates. At four decimal places you get 1.5157 in, which converts back to 38.4988 mm and is out by about 0.0012 mm, so the rounding is invisible. At the nearest 1/64 you get 1 33/64 in, which is 1.515625 in or 38.4969 mm, out by 0.0031 mm and still comfortably inside. At the nearest 1/32 you get 1 17/32 in, which is 1.53125 in or 38.8938 mm, out by about 0.394 mm and therefore outside the band. At the nearest 1/16 you get a tidy looking 1 1/2 in, which is exactly 38.1 mm, out by 0.4 mm and also outside.
So two of the four candidates fail a tolerance that was never about the conversion at all. The setup sheet gets 1.5157 in, the marking out gets 1 33/64 in, and the neat 1 1/2 in never leaves the conversation. Notice also that the failures arrive without warning: 1 1/2 in looks like the friendliest number of the four and is the worst of them.
Millimetre to inch chart
Two bands, because small part work and panel work need different steps. Fractions are the nearest 1/64.
| Millimetres | Decimal inches | Nearest 1/64 | Fraction error |
|---|---|---|---|
| 0.5 mm | 0.0197 in | 1/64 | -0.1031 mm |
| 1 mm | 0.0394 in | 3/64 | +0.1906 mm |
| 1.5 mm | 0.0591 in | 1/16 | +0.0875 mm |
| 2 mm | 0.0787 in | 5/64 | -0.0156 mm |
| 3 mm | 0.1181 in | 1/8 | +0.1750 mm |
| 5 mm | 0.1969 in | 13/64 | +0.1594 mm |
| 6 mm | 0.2362 in | 15/64 | -0.0469 mm |
| 8 mm | 0.3150 in | 5/16 | -0.0625 mm |
| 10 mm | 0.3937 in | 25/64 | -0.0781 mm |
| 12.7 mm | 0.5000 in | 1/2 | 0.0000 mm |
| 20 mm | 0.7874 in | 25/32 | -0.1563 mm |
| 25 mm | 0.9843 in | 63/64 | +0.0031 mm |
| 25.4 mm | 1.0000 in | 1 | 0.0000 mm |
| 50 mm | 1.9685 in | 1 31/32 | +0.0063 mm |
| 100 mm | 3.9370 in | 3 15/16 | +0.0125 mm |
| 150 mm | 5.9055 in | 5 29/32 | +0.0188 mm |
| 200 mm | 7.8740 in | 7 7/8 | +0.0250 mm |
| 300 mm | 11.8110 in | 11 13/16 | +0.0375 mm |
For reference at the far end, 1000 mm is 39.3700787402 in, and the reciprocal that drives the whole table is 1 mm equals 0.03937007874 in.
Inch fractions back to millimetres
The reverse direction is the one most converters bury, and it is the one you need when a supplier quotes imperial stock and the drawing is metric. Every value here is exact.
| Inch fraction | Decimal inches | Millimetres |
|---|---|---|
| 1/64 | 0.015625 in | 0.396875 mm |
| 1/32 | 0.03125 in | 0.79375 mm |
| 1/16 | 0.0625 in | 1.5875 mm |
| 1/8 | 0.125 in | 3.175 mm |
| 3/16 | 0.1875 in | 4.7625 mm |
| 1/4 | 0.25 in | 6.35 mm |
| 5/16 | 0.3125 in | 7.9375 mm |
| 3/8 | 0.375 in | 9.525 mm |
| 1/2 | 0.5 in | 12.7 mm |
| 5/8 | 0.625 in | 15.875 mm |
| 3/4 | 0.75 in | 19.05 mm |
| 7/8 | 0.875 in | 22.225 mm |
| 1 in | 1 in | 25.4 mm |
| 1 1/2 in | 1.5 in | 38.1 mm |
Near misses: threads, drills and clearance holes
The dangerous conversions are not the ones that come out wrong. They are the ones that come out almost right, where a metric part and an imperial part sit so close in diameter that they thread together for two turns and then jam. Diameter is the part that looks fine. Pitch is the part that destroys the thread.
| Metric size | Imperial substitute | Gap in diameter | Verdict |
|---|---|---|---|
| M6 bolt, 6 mm | 1/4 in, 6.35 mm | 0.35 mm | Pitch 1.0 mm against 20 TPI, which is 1.27 mm. Never thread them together. |
| M8 bolt, 8 mm | 5/16 in, 7.9375 mm | 0.0625 mm | Diameters look identical, pitch 1.25 mm against 18 TPI, which is 1.411 mm. |
| 3 mm drill | 1/8 in, 3.175 mm | 0.175 mm | Oversize hole. Fine for clearance, wrong for a reamed or tapped hole. |
| 6 mm drill | 15/64 in, 5.953 mm | 0.047 mm | The closest pair on the list and usually an acceptable clearance swap. |
| 12 mm drill | 1/2 in, 12.7 mm | 0.7 mm | Far too loose for anything that locates a part. |
| 25 mm stock | 1 in, 25.4 mm | 0.4 mm | Looks interchangeable in a catalogue. It is not, in a bore. |
The working rule is this: a difference of 0.1 mm or less is usually acceptable for a clearance hole, and never acceptable for tapping, reaming or a press fit. It is worth remembering why the two sets so rarely coincide. Imperial twist drills step in 1/64 in, which is 0.396875 mm, while metric drills step in 0.5 mm, so the two ladders interleave and almost never land on the same rung. Metric and imperial threads share nothing but a nominal diameter. The pitch and the thread angle differ, so a matching diameter is a coincidence, not a fit.
Sheet, plate and the gauge trap
Sheet thickness is where a correct conversion still produces a wrong part, because gauge is not a unit at all. It is a sheet number inherited from wire drawing, and the same number is a different thickness in every material.
| Gauge and material | Inches | Millimetres |
|---|---|---|
| 16 gauge mild steel | 0.0598 in | 1.52 mm |
| 16 gauge stainless | 0.0625 in | 1.59 mm |
| 16 gauge aluminium | 0.05082 in | 1.29 mm |
| 18 gauge mild steel | 0.0478 in | 1.21 mm |
| 18 gauge aluminium | 0.04030 in | 1.02 mm |
One gauge number therefore covers a spread of about 0.3 mm depending on what the sheet is made of, which is larger than the whole ISO 2768 medium band at small sizes. Always convert from the stated thickness, never from the gauge number alone. Plate, once it gets thick enough to be called plate, is quoted in fractions instead and the conversion is clean: 1/8 in is 3.18 mm, 1/4 in is 6.35 mm, 3/8 in is 9.53 mm, 1/2 in is 12.70 mm, 3/4 in is 19.05 mm and 1 in is 25.40 mm.
Reading millimetres on a drawing
On an ISO style engineering drawing the millimetre is the default unit and the symbol is usually omitted, so a bare number in a dimension is already millimetres. If you see a unit written out, that is the exception being flagged, not the rule being confirmed. Getting this backwards is the quietest way to turn a 38.5 mm feature into something no supplier can make.
The style conventions differ by system too. Under ASME Y14.5 a metric dimension takes a leading zero and drops any trailing zero, so 0.5 rather than .5, and the number of decimal places in a metric dimension does not have to match the number in its tolerance, which is a rule that does apply to inch dimensions. If you are converting a whole drawing rather than one number, convert once at full precision, apply the drawing's rounding rule at the end, and never feed an already rounded value into the next calculation. Rounding compounds, and a dimension that passes through two roundings can drift further than the tolerance it was drawn with.
Not the same as a centimetre job: one centimetre is ten millimetres, so a figure that looks tiny in millimetres is often a centimetre value that slipped a decimal place. If your measurement came off a soft tape rather than a drawing, the cm to inches converter is the better fit.
Frequently Asked Questions
How many inches is 1 mm?
One millimetre is 0.03937007874 inches, which is simply 1 divided by 25.4. At four decimal places that is 0.0394 in. Pushed onto a rule it becomes 3/64 in, and that is worth seeing: 3/64 in is 1.190625 mm, so the fraction is 0.19 mm oversize on a 1 mm feature. At this scale a fraction is the wrong tool and the decimal is the only honest answer.
How many millimetres are in one inch, a half inch and a quarter inch?
One inch is 25.4 mm, a half inch is 12.7 mm and a quarter inch is 6.35 mm. Carrying on down the ladder, 1/8 in is 3.175 mm, 1/16 in is 1.5875 mm, 1/32 in is 0.79375 mm and 1/64 in is 0.396875 mm. Every one of those is exact rather than rounded, because halving an exact number keeps it exact.
Is 25.4 mm exactly one inch, or is it rounded?
Exactly. The international yard and pound agreement of 1 July 1959 fixed the yard at exactly 0.9144 metres, which fixes the inch at exactly 25.4 mm. ANSI had already adopted that value in 1933 and NACA in 1952. Before the agreement the United States used an inch of about 25.4000508 mm, a difference you will only ever meet on very old drawings or survey work.
Which fraction should I round to, 1/16, 1/32 or 1/64?
Work back from the tolerance rather than from habit. Rounding to 1/16 can be wrong by up to 0.79375 mm, 1/32 by up to 0.3969 mm and 1/64 by up to 0.1984 mm, because the worst case is always half a step. Under ISO 2768 medium class a feature from 6 to 30 mm is allowed plus or minus 0.2 mm, so only 1/64 fits, and it fits with nothing to spare. If the drawing is tighter than that, stop using fractions and work in decimal inches or thou.
Is 5 mm bigger than a quarter inch?
No, it is noticeably smaller. A quarter inch is 6.35 mm, so 5 mm falls 1.35 mm short of it. In the other direction 5 mm is 0.1969 in, which lands on 13/64 in on a rule, one sixty fourth under 7/32. If you are picking a drill, a 5 mm bit will not open a hole that a 1/4 in bolt passes through.
How thick is 3 mm in inches?
3 mm is 0.1181 in, which rounds to 1/8 in on a rule. The two are not the same thing: 1/8 in is 3.175 mm, so a 3 mm sheet is 0.175 mm thinner than a nominal 1/8 in sheet. That gap is fine for a clearance hole, and it is not fine if the material has to fit a slot or a rebate machined for 1/8 in stock.
What is 25 mm in inches?
25 mm is 0.9843 in, so it is 0.4 mm short of a full inch rather than equal to one. The nearest fraction is 63/64 in, which is accurate to about three thousandths of a millimetre. Treating 25 mm and 1 in as interchangeable is fine in a catalogue description and a real problem in a bore or a slot, where 0.4 mm of slack is visible.
Can I use a metric drill bit or bolt where an imperial one is specified?
For a clearance hole, a difference of 0.1 mm or less is usually acceptable, which makes a 6 mm drill a reasonable stand in for 15/64 in at a gap of 0.047 mm. For tapping, reaming or a press fit the answer is no. Fasteners are worse still, because the diameters can match while the threads do not: an M8 bolt and a 5/16 in bolt differ by only 0.0625 mm in major diameter, yet the M8 runs a 1.25 mm pitch against 18 threads per inch, which is 1.411 mm. They will start and then strip.
Is 16 gauge sheet the same thickness in steel and aluminium?
No, and this is the commonest way a correct conversion still yields a wrong part. Gauge is a sheet number, not a length, and it is material dependent: 16 gauge mild steel is 0.0598 in or 1.52 mm, 16 gauge stainless is 0.0625 in or 1.59 mm, and 16 gauge aluminium is 0.05082 in or 1.29 mm. One gauge number covers roughly 0.3 mm of spread, so always convert from a stated thickness rather than from the gauge number.