ISO 286 Fit Calculator

Hole limits, shaft limits and the real clearance or interference range for any ISO 286 pair — H7/g6, H7/n6, H8/f7, H11/c11 and 700 more — with the class of fit, the hole-basis and shaft-basis equivalence, and every band computed from the standard’s own rules rather than copied from a table.

Last updated 24 Sep 2026

ISO 286 limits and fits

Size and two classes → four limits and the fit

Nominal size (mm)The basic size both parts are made to. ISO 286 covers 1 mm to 500 mm in thirteen size steps, and the tolerance for a step is computed at the geometric mean of its two limits — so every size inside a step gets the same band.

Hole tolerance classThe capital letter fixes where the band sits relative to the nominal size; the number is the IT grade, which fixes how wide it is. H puts the band’s bottom exactly on the nominal size, which is why hole basis is the usual choice.

Shaft tolerance classSmall letters for the shaft. a to h sit below the nominal size, js straddles it, k to u sit above it. h puts the band’s top exactly on the nominal size.

Where this hole actually came out (% of its band)0% is the hole at its smallest, 100% at its largest. This is what makes a transition fit visible: move both sliders and watch the actual fit change sign.

Where this shaft actually came out (% of its band)0% is the shaft at its smallest, 100% at its largest.

Not a circuit: the two tolerance zones drawn on a common deviation axis in micrometres, with the nominal size as the horizontal line through the middle. The left column is the hole's band and the right column is the shaft's, each placed by its own letter and sized by its own IT grade, and the vertical scale is shared so the two can be compared directly. Below them the whole clearance range is shaded on its own scale with zero marked: a band entirely to the right of zero is a clearance fit, entirely to the left is an interference fit, and a band that STRADDLES the line is a transition fit — which is the point of the drawing, because that is the case the designation does not tell you about. The single line inside the shaded range is where the particular pair you set with the two sliders lands. Both scales are symmetric about zero and rescale themselves to whatever fit you choose, so read the numbers rather than the lengths.

41.0µmExample

A Ø25 mm H7 hole and a g6 shaft — the sliding fit, with both parts assumed to come out in the middle of their bands

Two bands, four limits, and a subtraction

hole limits = D + ES and D + EI · shaft limits = D + es and D + ei · maximum clearance = ES − ei · minimum clearance = EI − es · ES − EI = es − ei = IT(grade)

D: nominal size. Both parts are dimensioned from the same number; the letters and grades say where each band sits relative to it
ES, EI: the hole’s upper and lower deviations from the nominal size, in micrometres. Capitals for a hole, always
es, ei: the shaft’s upper and lower deviations. Small letters for a shaft
IT(grade): the standard tolerance for that grade at that size step — the WIDTH of the band. It is the same number for a hole and a shaft of the same grade, and it is the check that catches a mis-copied table: ES − EI must equal it
Δ: the correction ISO 286 applies to hole letters J to ZC, equal to IT(n) − IT(n−1) above 3 mm and zero at and below it. It is what makes hole basis and shaft basis give the same fit
i: the standard tolerance factor, 0.45·∛D + 0.001·D micrometres, evaluated at the geometric mean of the size step. Every IT grade is a fixed multiple of it

Worked example

A Ø25 mm H7 hole and a g6 shaft — the sliding fit, with both parts assumed to come out in the middle of their bands
Find the size step. 25 mm falls in “over 18 to 30”, whose geometric mean is √(18 × 30) = 23.238 mm. Every diameter in that step gets the same tolerance, so a Ø18.5 shaft and a Ø30 shaft share these numbers
The standard tolerance factor there is i = 0.45·∛23.24 + 0.001 × 23.24 = 1.307 µm, so IT7 = 16i = 20.9 and IT6 = 10i = 13.1. The standard rounds those to 21 and 13 µm, and those are the two band widths
The hole is H, so its LOWER deviation is zero by definition: EI = 0 and ES = +IT7 = +21 µm. The hole runs 25.0000 to 25.0210 mm
The shaft is g, whose fundamental deviation is its UPPER one: es = −2.5·D^0.34 = -7.29, which the standard rounds to −7 µm. Then ei = es − IT6 = −20 µm. The shaft runs 24.9800 to 24.9930 mm
Now subtract. The biggest hole meets the smallest shaft: maximum clearance = ES − ei = 21 − (−20) = 41 µm. The smallest hole meets the biggest shaft: minimum clearance = EI − es = 0 − (−7) = 7 µm
Both ends are positive, so this is a CLEARANCE fit and it will always come apart. The spread is 34 µm, which is just IT7 + IT6 — the two grades added. That is always true and it is the cheapest sanity check there is: if your loosest and tightest clearances do not differ by exactly the sum of the two grades, one of the four deviations is wrong
The loosest clearance is 5.9 times the tightest, and that ratio is the real difficulty with fits. A bearing designed for 7 µm of running clearance gets up to 41, and the oil film, the running temperature and the noise are all different at the two ends. Tighten it by going to H6/g5 and the spread falls to 22 µm, for a real cost in machining
Change nothing but the shaft letter and watch the character change. With n6 the range is -28 to 6 µm — interference at one end and clearance at the other, which is what TRANSITION means. With p6 it is -35 to -1 — interference at both ends, but only just

The ten preferred fits, with the clearance each gives at Ø25 mm in micrometres

Read this table by the last column, because that is how people actually arrive: knowing the job and not the code. Every clearance figure is computed from the ISO 286 construction rather than copied from a table of limits, and the three published worked fits that could be checked against a source all reproduce exactly — H8/f7 at Ø50 gives 25 to 89 µm, H7/k6 gives 18 µm of interference to 23 µm of clearance, and H7/p6 gives 1 to 42 µm of interference. Two things in the table are worth a second look. H7/k6 and H7/n6 are both called locational TRANSITION fits and both really do go either way; they differ in where the transition sits, not in whether there is one. And H7/p6, which the standard calls a locational INTERFERENCE fit, has a minimum interference of one micrometre at Ø25 — it locates, and it will not carry torque without a key or a pin. The second column is the same fit expressed shaft basis; see the next table for why those two columns are genuinely equal and when they are not. These dimensions come from a published standard’s table, not from a formula. The standard itself is cited below and the printed values are attributed to the catalogue they were taken from; a different publisher may round differently in the last digit.
Hole basis	Shaft basis	What it is called	Clearance at Ø25 mm (µm, negative is interference)	Class of fit	What it is for
H11/c11	C11/h11	Loose running	110 to 370	clearance	Wide commercial tolerances or allowances on external members: pivots, latches, parts affected by corrosion, heat or contamination.
H9/d9	D9/h9	Free running	65 to 169	clearance	Not for use where accuracy is essential, but good for large temperature variations, high running speeds or heavy journal pressures.
H8/f7	F8/h7	Close running	20 to 74	clearance	Running on accurate machines and accurate location at moderate speeds and journal pressures.
H7/g6	G7/h6	Sliding	7 to 41	clearance	Not intended to run freely, but to move and turn freely and locate accurately.
H7/h6	H7/h6	Locational clearance	0 to 34	clearance	A snug fit for locating stationary parts that can still be freely assembled and disassembled.
H7/k6	K7/h6	Locational transition	-15 to 19	transition	Accurate location, a compromise between clearance and interference.
H7/n6	N7/h6	Locational transition	-28 to 6	transition	More accurate location where greater interference is permissible.
H7/p6	P7/h6	Locational interference	-35 to -1	interference	Rigidity and alignment with prime accuracy of location, but without special bore pressure requirements.
H7/s6	S7/h6	Medium drive	-48 to -14	interference	Ordinary steel parts, or shrink fits on light sections; the tightest fit usable with cast iron.
H7/u6	U7/h6	Force	-61 to -27	interference	Parts that can be highly stressed, or shrink fits where the heavy pressing forces required are impractical.

Hole basis against shaft basis at Ø25 mm — where the identity holds, and where it quietly does not

The first four columns are the identity everybody quotes, and it is exact — not approximately, exactly, at every one of the thirteen size steps and for all seven interference letters, which is 120 pairs checked here. But it is not a fact about the LETTERS. It holds because ISO 286’s Δ correction for a hole letter is exactly IT(n) − IT(n−1), and because the preferred pairs put grade n on the hole and grade n−1 on the shaft. Break either condition and it stops. The last two columns show what happens when both members are at grade 7: H7/n7 allows up to 6 µm of clearance and N7/h7 allows 14, which are different fits with the same letters. And below 3 mm the identity fails even for the preferred pairs, because the standard sets Δ to zero on the first size step — at Ø2 mm, H7/n6 and N7/h6 differ by IT7 − IT6 = 4 µm. Use hole basis for general machining, where a reamer or a plug gauge fixes the hole and the shaft is turned to suit. Use shaft basis where the shaft is the fixed thing: drawn bar, a ground shaft bought to size, a bearing journal, a standard spindle.
Hole basis	Its clearance range (µm)	Shaft basis	Its clearance range (µm)	Agreement	H7 with the shaft at grade 7 (µm)	The hole letter at grade 7 with h7 (µm)
H7/k6	-15 to 19	K7/h6	-15 to 19	identical	-23 to 19	-15 to 27
H7/m6	-21 to 13	M7/h6	-21 to 13	identical	-29 to 13	-21 to 21
H7/n6	-28 to 6	N7/h6	-28 to 6	identical	-36 to 6	-28 to 14
H7/p6	-35 to -1	P7/h6	-35 to -1	identical	-43 to -1	-35 to 7
H7/r6	-41 to -7	R7/h6	-41 to -7	identical	-49 to -7	-41 to 1
H7/s6	-48 to -14	S7/h6	-48 to -14	identical	-56 to -14	-48 to -6
H7/u6	-61 to -27	U7/h6	-61 to -27	identical	-69 to -27	-61 to -19

Seven fits across all thirteen ISO 286 size steps, in micrometres of clearance

Every cell computed, none copied. The second column is the number that does the work: ISO 286 evaluates its formulas at the GEOMETRIC mean of each size step, not the arithmetic one, which is why the 1 to 3 mm step is evaluated at 1.73 mm and not at 2. Three patterns are worth reading off. H7/h6 always starts at exactly zero at the tight end, because H’s lower deviation and h’s upper deviation are both zero by definition — it is the only fit in the preferred set that is guaranteed to go together and guaranteed not to be loose. H7/p6’s tight end is 1 µm of interference at Ø25 and never more than 5 µm anywhere, because p’s fundamental deviation is IT7 to within five micrometres and H7’s upper deviation IS IT7, so the two cancel by construction. And H11/c11 is two orders of magnitude looser than H7/h6 at the same size, which is the whole reason the letter-and-grade system exists: one code covers a hinge pin and a bearing seat.
Nominal size step (mm)	Geometric mean D (mm)	H7/g6	H7/h6	H7/n6	H7/p6	H7/s6	H8/f7	H11/c11
over 1 to 3	1.73	2 to 18	0 to 16	-10 to 6	-12 to 4	-20 to -4	6 to 30	60 to 180
over 3 to 6	4.24	4 to 24	0 to 20	-16 to 4	-20 to 0	-27 to -7	10 to 40	70 to 220
over 6 to 10	7.75	5 to 29	0 to 24	-19 to 5	-24 to 0	-32 to -8	13 to 50	80 to 260
over 10 to 18	13.42	6 to 35	0 to 29	-23 to 6	-29 to 0	-39 to -10	16 to 61	95 to 315
over 18 to 30	23.24	7 to 41	0 to 34	-28 to 6	-35 to -1	-48 to -14	20 to 74	110 to 370
over 30 to 50	38.73	9 to 50	0 to 41	-33 to 8	-42 to -1	-59 to -18	25 to 89	120 to 440
over 50 to 80	63.25	10 to 59	0 to 49	-39 to 10	-51 to -2	-72 to -23	30 to 106	140 to 520
over 80 to 120	97.98	12 to 69	0 to 57	-45 to 12	-59 to -2	-93 to -36	36 to 125	170 to 610
over 120 to 180	146.97	14 to 79	0 to 65	-52 to 13	-68 to -3	-125 to -60	43 to 146	210 to 710
over 180 to 250	212.13	15 to 90	0 to 75	-60 to 15	-79 to -4	-159 to -84	50 to 168	260 to 840
over 250 to 315	280.62	17 to 101	0 to 84	-66 to 18	-88 to -4	-202 to -118	56 to 189	330 to 970
over 315 to 400	354.96	18 to 111	0 to 93	-73 to 20	-98 to -5	-226 to -133	62 to 208	360 to 1,080
over 400 to 500	447.21	20 to 123	0 to 103	-80 to 23	-108 to -5	-272 to -169	68 to 228	440 to 1,240

The inch system’s fit classes, and why they are not a translation of these

This page does not carry the inch classes’ numeric limits and that is deliberate rather than lazy. ANSI/ASME B4.1’s limits are tabulated in ten-thousandths of an inch and are not generated by any rule this page could derive and check, so carrying them would mean copying about a thousand numbers out of a summarised fetch — and in this batch alone, four such fetches shifted a column. What IS worth having is the boundary itself, stated in the standard’s own words above: the inch families are built to hold a running performance or a bore pressure constant across the size range, and the ISO letters are built to hold a deviation formula constant. There is no one-to-one mapping, there cannot be one, and a table claiming otherwise is wrong in the third digit even where it looks right in the first. The same designation can mean different dimensions in different standards families — ANSI against ISO, inch against metric, one national standard against another. The family used here is named beside every figure; check which one your part was made to.
Family	What ANSI/ASME B4.1 says it is for	Does it map onto an ISO class?
RC1 to RC9, running and sliding	Intended to provide a similar running performance, with suitable lubrication allowance, throughout the range of sizes	No, and that sentence is the reason. An RC class holds a PERFORMANCE constant across the size range; an ISO letter holds a DEVIATION RULE constant. Those cannot both be true of the same ladder
LC1 to LC11, locational clearance	Intended for parts which are normally stationary, but which can be freely assembled or disassembled	Loosely comparable to the H/h and H/g families in intent, and not equal to any of them in numbers
LT1 to LT6, locational transition	A compromise between clearance and interference fits, for application where accuracy of location is important, but either a small amount of clearance or interference is permissible	The same idea as H7/js6, H7/k6 and H7/n6 — and note that the inch system needs six grades where the ISO preferred set uses two
LN1 to LN3, locational interference	Used where accuracy of location is of prime importance and for parts requiring rigidity and alignment with no special requirements for bore pressure	The same idea as H7/p6. The phrase to notice is “no special requirements for bore pressure”, which is the inch standard saying the same thing ISO says by making p’s minimum interference zero
FN1 to FN5, force and shrink	Force or shrink fits constitute a special type of interference fit, normally characterized by maintenance of constant bore pressures throughout the range of sizes	No. A constant bore pressure across the size range is a different construction from a constant deviation rule, and it is why an FN2 at half an inch and an FN2 at four inches are not the same ISO class

The letter moves the band, the number widens it, and transition means transition

A fit is a range, and the code on the drawing is a rule for generating it. H7/g6 on a Ø25 feature is not “a bit of clearance”. It is 7 to 41 µm, because the hole may be anywhere in a 21 µm band and the shaft anywhere in a 13 µm one, and the two ends of that range are what the design has to work at. The capital letter and the small letter say where each band sits relative to the nominal size; the numbers are IT grades and say how wide each band is. Those are two independent decisions and people routinely conflate them — a tighter grade does not move the fit, it only narrows it, and if the clearance is wrong the letter is what needs changing.

ISO 286 is the one standard on this site whose values can be derived rather than copied, and this page derives them. The standard tolerance factor is i = 0.45·∛D + 0.001·D micrometres, evaluated at the geometric mean of the size step; IT7 is 16i, IT6 is 10i, and IT(n+5) is exactly ten times IT(n) from IT7 upward. Each letter’s fundamental deviation has its own rule: g is −2.5·D^0.34, f is −5.5·D^0.41, m is IT7 − IT6 exactly, n is 5·D^0.34, and a hole letter’s is the shaft letter’s with the sign reversed plus a correction Δ. Every one of those was checked here against published tables before it was trusted, and the check is the trivial one that no shifted column survives: ES − EI must equal the IT grade. It rejected four separate fetches in the making of this page, including ISO 286-2’s own published preview. The dowel pin and hole calculator on this site uses the same machinery for the one fit a dowel pin is allowed.

Clearance, transition, interference — and transition really is both. If the smallest hole is still bigger than the biggest shaft, the fit is a clearance fit and it always comes apart. If the biggest hole is still smaller than the smallest shaft, it is an interference fit and it never goes together by hand. In between, the bands overlap and the SAME designation gives interference on one pair of parts and clearance on the next. H7/k6 and H7/n6 are both like this, and it is the most misread thing in the standard: you cannot promise an assembler that an H7/n6 will press together and you cannot promise it will not. This page reports the split as a percentage — computed exactly, as a ratio of areas over the two bands — and lets you put each part anywhere in its own band to see the actual clearance change sign.

Hole basis and shaft basis, and the identity that is narrower than people think. H7/n6 and N7/h6 give the same fit. That is exact — it holds at every size step and for every interference letter, which is 120 pairs checked here — but it is not a fact about the letters. It holds because Δ is exactly IT(n) − IT(n−1) and because the preferred pairs put grade n on the hole and grade n−1 on the shaft. Put both members at grade 7 and H7/n7 allows 6 µm of clearance where N7/h7 allows 14. Go below 3 mm, where the standard sets Δ to zero, and the identity fails even for the preferred pairs. Which basis to use is a manufacturing question: hole basis for general machining, because a reamer and a plug gauge fix the hole and the shaft is easy to turn to suit; shaft basis where the shaft is already fixed, by drawn bar, a bought ground shaft, a bearing journal or a standard spindle.

What this page is not. It gives limits, not consequences. For the contact pressure, torque capacity, hub stress and assembly temperature of an interference fit, the interference fit and shrink fit calculator does the Lamé solution and subtracts the surface roughness the faces lose on assembly. For what happens when the assembly is not at 20 °C, the thermal effect on fit calculator moves both bands. For the grade you can actually hold on a given machine, and what a grade costs, the IT grade tolerance calculator. And for the plug and ring gauges that accept or reject the parts, the go / no-go gauge tolerance calculator, which is where you find out how much of your band the gauges themselves eat.

Frequently asked questions

What clearance does an H7/g6 fit give?

A range, not a number. At Ø25 mm it is 7 to 41 µm of diametral clearance: the hole is 25.0000 to 25.0210 and the shaft is 24.9800 to 24.9930. The tight end is the smallest hole against the biggest shaft and the loose end is the other way round. The spread between them, 34 µm, is exactly IT7 + IT6 — the two grades added — and that identity is the quickest check that all four limits are right.

Is H7/n6 a press fit?

No. It is a transition fit, and at Ø25 it runs from 28 µm of interference to 6 µm of clearance. About nine in ten pairs will interfere and about one in ten will drop together, so it locates well and it carries no guaranteed torque. If you need torque from the fit itself you want s6 or u6, and you want the Lamé calculation that goes with them. If you only need the parts concentric, n6 is an excellent choice and the occasional loose one does no harm.

Does H7/n6 really give the same fit as N7/h6?

Yes, exactly, at every size from 3 mm to 500 mm — and it is worth knowing why, because the reason tells you when it stops. The hole letter’s fundamental deviation is the shaft letter’s with the sign reversed, plus a correction Δ equal to IT(n) − IT(n−1). Work the algebra through and the two clearance ranges coincide provided the hole is one grade coarser than the shaft, which every preferred pair is. Make both grade 7 and they part company: H7/n7 allows 6 µm of clearance and N7/h7 allows 14. Below 3 mm they part company too, because Δ is zero there.

Which basis should I use, hole or shaft?

Hole basis for general machining, and it is the default for a good manufacturing reason: a hole is made by a reamer, a boring bar or a broach and checked with a plug gauge, all of which come in fixed sizes, so it is much cheaper to keep the hole at H and vary the shaft. Shaft basis where the shaft is the fixed thing — drawn bar used as bought, a ground shaft to a standard diameter, a bearing journal, a commercial spindle — because then it is the bore that is easy to adjust. Do not mix bases within an assembly if you can help it; the tooling savings are the whole point and they disappear.

Can I get the inch RC, LC, LT, LN and FN classes here?

Their meanings, yes; their numbers, deliberately not. The five inch families and what ANSI/ASME B4.1 says each is for are in a table on this page, and so is the reason they do not map onto the ISO letters: B4.1’s own words are that an RC class is “intended to provide a similar running performance … throughout the range of sizes” and an FN class is “characterized by maintenance of constant bore pressures throughout the range of sizes”. Those hold a PERFORMANCE constant across the size range, where an ISO letter holds a deviation formula constant. No single ISO class tracks an RC or an FN class across the range, so any table claiming a one-to-one equivalence is wrong somewhere in the middle.

Why does a Ø30 and a Ø30.5 shaft get different limits for the same class?

Because ISO 286 quantises size into thirteen steps and 30 mm is the last size in one of them. The tolerance for a step is computed once, at the geometric mean of the step’s two limits, and applied to every size inside it. So h6 is 0/−13 µm up to and including Ø30 and 0/−16 µm from just above it. The series on this page plots the staircase. It is also a real trap on a drawing: a feature nominally at a step boundary can have its band change if somebody rounds the nominal size.

The two parts are made to the same nominal size — how can there be clearance at all?

Because the nominal size is a reference, not a target. Neither part is required to be 25.000 mm and neither will be. The letters say which side of 25 each band sits on and the grades say how wide each band is, and the fit is whatever the arithmetic gives. H/h is the interesting case: H’s lower deviation and h’s upper deviation are both exactly zero, so the tightest H7/h6 pair is metal-to-metal at precisely the nominal size and every other pair has clearance. It is the only fit in the preferred set that is guaranteed to assemble and guaranteed not to be loose.

Related calculators

References

ISO 286-1:2010, Geometrical product specifications (GPS) — ISO code system for tolerances on linear sizes — Part 1: Basis of tolerances, deviations and fits. Cited by number; the standard is copyrighted and its tables are not reproduced here. What this page uses from it is its STRUCTURE, which is not a table: the standard tolerance factor i = 0.45·∛D + 0.001·D evaluated at the geometric mean of each nominal size step, the grade multipliers (IT5 = 7i, IT6 = 10i, IT7 = 16i, IT8 = 25i, IT9 = 40i, IT10 = 64i, IT11 = 100i and so on), the rule that IT(n+5) is ten times IT(n) from IT7 upward, the formulas for each letter’s fundamental deviation, and the Δ correction for hole letters J to ZC. Every one of those is checked here against a published table rather than trusted.
ISO 286-2:2010, … Part 2: Tables of standard tolerance classes and limit deviations for holes and shafts. Cited, and its freely published preview was fetched and then DISCARDED. The extract returned an upper-deviation column belonging to a different letter and a lower-deviation column mixed from several size steps — g6 as −10/−68 where it is −6/−17, h6 as 0/−58 where it is 0/−11, js6 as ±29 where it is ±5.5. The identity ES − EI = IT rejected every row. Recorded because this is the second batch on this site to have the same document shift a column on it, and because it is the strongest possible argument for deriving the values instead.
Machining Doctor. Tolerance charts by fundamental deviation letter (one page per letter, hole and shaft, stated as ISO 286-1:2010). The source for the fundamental deviations of c, d, e, f, g, k, m, n, p, r, s and u on this page, fetched one letter at a time so a shifted column could not hide in a wide table. Every row of every letter satisfies es − ei = IT(grade) exactly — 220 independent checks — and its N7 hole column reproduces the Δ = IT(n) − IT(n−1) rule at all thirteen size steps. Its u6 column also agrees with the interference-fit page already on this site at all sixteen steps that page covers, which cross-validates both.
RoyMech. ISO 286-2 shaft tolerances and ISO hole limits and fits tables. The second source for the hole column: its whole 18–30 mm row (E6 to R7, twenty-four classes) is reproduced here exactly from the Δ rule, which is what proves Δ is IT(n) − IT(n−1) and not IT(n) − IT6. Its SHAFT table was fetched too and partly rejected: it returned a11’s deviations in the c11 row (−290/−470 against c11’s −95/−205 at 10–18 mm) and IT9 in the lower column of the d9 and e8 rows. Its upper column is right at every row and is used as the independent confirmation of d, e, f and g.
AmesWeb. Preferred fits and tolerances charts (ISO and ANSI metric standards), attributing its descriptions to ISO 286-1 and ANSI B4.2-1978. The source for the ten preferred hole-basis fits, their shaft-basis twins and the description of each — loose running, free running, close running, sliding, locational clearance, two locational transitions, locational interference, medium drive and force. Cross-read against Wikipedia’s Engineering fit tables, whose wording is different and whose meaning agrees at all ten rows.
MetricMech. ISO 286 fits chart: H7/g6 to H7/s6 tolerances. The independent worked check at Ø25 mm: H7 +21/0, g6 −7/−20, h6 0/−13, k6 +2/+15, n6 +15/+28, p6 +22/+35, s6 +35/+48, and the resulting clearance and interference ranges for all six fits. Every one is reproduced here from the derivation, which is the check that the derivation is the standard’s and not merely self-consistent.
Wikipedia, Engineering fit. Used for two things and checked on both. Its c11 and H11 columns are the second source for the loose-running fit, and one row of them is wrong: it repeats the 10–18 mm lower deviation of c11 (−205) on the 6–10 mm row, where ES − EI = IT11 gives −170. Its worked clearances at Ø50 for H8/f7 (+0.025 to +0.089), H7/k6 (+0.023 to −0.018) and H7/p6 (−0.001 to −0.042) are reproduced here exactly and are three of the page’s assertions.
Firat Bearing. ISO 286 — IT tolerance grades reference. The source for the IT01 to IT4 and IT12 to IT16 rows. The IT12 to IT16 rows turned out to be EXACTLY ten times IT7 to IT11 at all sixty-five cells, which is ISO 286-1’s own rule, so IT17 and IT18 are derived from it here rather than copied. The IT2, IT3 and IT4 rows satisfy the standard’s other construction rule — a geometric progression from IT1 to IT5 — at twelve of the thirteen size steps.
ANSI/ASME B4.1-1967 (R1994), Preferred Limits and Fits for Cylindrical Parts, and ANSI B4.2-1978 for its metric twin. The source for the five inch fit families and, crucially, for their own statements of intent: the RC classes are “intended to provide a similar running performance, with suitable lubrication allowance, throughout the range of sizes” and the FN classes are “normally characterized by maintenance of constant bore pressures throughout the range of sizes”. Those two sentences are why an inch class cannot map onto an ISO letter: the inch classes hold a PERFORMANCE constant across the size range and the ISO letters hold a DEVIATION RULE constant, and the two cannot both be true of the same ladder. Descriptions quoted via AmesWeb’s published ANSI preferred-fits charts; the numeric limit tables were not reproduced.