Parallel Key and Keyway Dimensions Calculator

Key width, height and both keyway depths to DIN 6885-1 and ISO 773 from the shaft diameter, with BOTH depth conventions printed side by side, the real clearance or interference each fit class delivers computed from the ISO 286 bands, Woodruff keys as a separate family, and the hub wall left over the keyway.

Last updated 24 Sep 2026

Parallel key and keyway dimensions

Shaft diameter → key size, both keyway depths, the fit

Shaft diameter (mm)The whole key size comes off this one number, as a step function. Shafts at the bottom of a band get the same key as shafts at the top, so they lose more section.

Fit classThe key is h9 in every class; the keyways change. Watch the clearance figures below — none of the three is a press fit.

Key length (mm)The engaged length, not the hub length. Published guidance is 1 to 1.5 shaft diameters, and 1.5 is quoted as both a floor and a ceiling by different publishers.

Hub outside diameter (mm)Used only for the hub wall left over the keyway floor — the dimension a hub actually cracks at, and the one no key standard knows about.

Not a circuit: the shaft and hub in cross-section, with the keyway drawn to scale in its proportion to the shaft. The key width and both keyway depths are drawn from the standard's table for the diameter you entered, quantised to fiftieths of the diameter so the figure stays light. The two dimension lines on the shaft are the point: the short one on the right is t₁, measured from the shaft surface down to the keyway floor, which is what DIN 6885-1 and ISO 773 tabulate; the long one on the left is d − t₁, measured from the floor across to the opposite surface, which is what ANSI B17.1 tabulates as its S dimension. They are the SAME keyway. The stack on the right prints both, plus the third reading — what an ordinary micrometre actually gives, which is neither. The sketch at the bottom is why: the anvil is wider than the keyway, so it bridges the slot and rests on the two shoulders either side, never reaching the floor at all. It measures the shaft and reports the full diameter. There is no partial reading and no correction to make — a contact that does reach the flat floor reads d − t₁ exactly, because the floor is a plane and the opposite contact is a tangent plane. Use a narrow-footed depth gauge, or a pin.

14mmExample

A 50 mm shaft, the normal fit, a 70 mm key and a 90 mm outside diameter hub

A table lookup, an ISO 286 calculation, and one piece of geometry

b, h, t₁, t₂ = f(shaft Ø) from DIN 6885-1 · t₁ + t₂ − h = 0.2, 0.3 or 0.4 mm by key width · clearance = keyway width − key width, from the ISO 286 bands

b, h: key width and height. A step function of the shaft diameter, not a formula — the standard’s table is the definition
t₁: keyway depth in the shaft, measured from the shaft surface to the keyway floor on the centreline
t₂: keyway depth in the hub, measured from the bore surface outwards
d − t₁: the same keyway expressed as the distance from the floor to the opposite side of the shaft. ANSI B17.1’s S dimension
IT9: the ISO 286 grade every keyway width tolerance on this page is built from, computed from the standard tolerance factor i = 0.45·D^(1/3) + 0.001·D at the geometric mean of the size band
h9, N9, P9, H9, D10, Js9: the classes. h9 on the key in every fit; the keyways change class with the fit

Worked example

A 50 mm shaft, the normal fit, a 70 mm key and a 90 mm outside diameter hub
50 mm falls in the 44 to 50 band, so the key is 14 × 9. Both published sources agree on the whole b × h ladder and the width is 0.280 of the shaft diameter, which is the quarter-of-the-shaft rule of thumb arriving from a table
Shaft keyway t₁ = 5.5 mm, hub keyway t₂ = 3.8 mm. Check them: t₁ + t₂ − h = 5.5 + 3.8 − 9 = 0.3 mm, and the clearance ladder for a 14 mm key says 0.3. The table is self-consistent
The SAME keyway, on the other convention, is d − t₁ = 50 − 5.5 = 44.5 mm across the shaft, and the hub's is d + t₂ = 53.8 mm. If your drawing says 44.5 it is not asking for a 44.5 mm deep keyway
A word on measuring it. Put an ordinary micrometre across the shaft and it will read 50.00 mm — the full diameter — because the anvil is wider than the 14 mm keyway and bridges it, resting on the shoulders either side. Those shoulders stand 1.00 mm proud of a chord across the keyway, so there is nothing in between to read. Use a narrow-footed depth micrometre or a pin in the keyway
Normal fit: the key is h9, so 14.000 to 13.957 mm. The shaft keyway is N9, -55 to -12 µm on the width, so the joint runs from 55 µm of interference at the tight end to 31 µm of clearance at the loose end. The hub is Js9 and runs 21.5 µm interference to 64.5 µm clearance
The close fit does NOT fix that. P9 against h9 gives 61 µm of interference at worst and 25 µm of CLEARANCE at best. Every keyway fit in the standard is a transition fit
Key length: 70 mm is 1.40 × the shaft diameter, inside the 1 to 1.5 d band the publishers give — and note that they give the same 1.5 figure as a floor and as a ceiling, for different reasons
Hub web over the keyway: (90 − 50)/2 − 3.8 = 16.2 mm, which is 4.3 times the keyway depth. That is the dimension a hub cracks at, and it is the only thing on this page that deserves the word web

The DIN 6885-1 / ISO 773 ladder, with both depth conventions printed side by side

The seventh column is the check that this table is not a mis-transcription, and it is worth understanding. t₁ + t₂ − h is the clearance left over the top of the key once it is seated, and it is a three-step ladder keyed on the key WIDTH: 0.2 mm up to b = 3, 0.3 mm from b = 4 to 16, 0.4 mm from b = 18. It holds at all nineteen rows exactly. One published extract of this table returned the hub depth in the column headed “h” and the shaft depth in a column headed “t4”, which would make the key for a 22–30 mm shaft 3.3 mm tall instead of 7 mm; the identity caught it on the first row. The last column is the other sanity check: a key is between a quarter and a third of the shaft diameter wide at the BOTTOM of each band and between 0.225 and 0.333 at the top, which is where “about a quarter of the shaft” comes from and also why it is only ever about. 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.
Shaft Ø (mm)	Key b × h	t₁ shaft (mm)	t₂ hub (mm)	d − t₁ at the top of the band (mm)	d + t₂ at the top of the band (mm)	t₁ + t₂ − h (mm)	IT9 on the widths (µm)	b ÷ d at the top of the band
over 6 to 8	2 × 2	1.2	1.0	6.8	9.0	0.2	25	0.250
over 8 to 10	3 × 3	1.8	1.4	8.2	11.4	0.2	25	0.300
over 10 to 12	4 × 4	2.5	1.8	9.5	13.8	0.3	30	0.333
over 12 to 17	5 × 5	3.0	2.3	14.0	19.3	0.3	30	0.294
over 17 to 22	6 × 6	3.5	2.8	18.5	24.8	0.3	30	0.273
over 22 to 30	8 × 7	4.0	3.3	26.0	33.3	0.3	36	0.267
over 30 to 38	10 × 8	5.0	3.3	33.0	41.3	0.3	36	0.263
over 38 to 44	12 × 8	5.0	3.3	39.0	47.3	0.3	43	0.273
over 44 to 50	14 × 9	5.5	3.8	44.5	53.8	0.3	43	0.280
over 50 to 58	16 × 10	6.0	4.3	52.0	62.3	0.3	43	0.276
over 58 to 65	18 × 11	7.0	4.4	58.0	69.4	0.4	43	0.277
over 65 to 75	20 × 12	7.5	4.9	67.5	79.9	0.4	52	0.267
over 75 to 85	22 × 14	9.0	5.4	76.0	90.4	0.4	52	0.259
over 85 to 95	25 × 14	9.0	5.4	86.0	100.4	0.4	52	0.263
over 95 to 110	28 × 16	10.0	6.4	100.0	116.4	0.4	52	0.255
over 110 to 130	32 × 18	11.0	7.4	119.0	137.4	0.4	62	0.246
over 130 to 150	36 × 20	12.0	8.4	138.0	158.4	0.4	62	0.240
over 150 to 170	40 × 22	13.0	9.4	157.0	179.4	0.4	62	0.235
over 170 to 200	45 × 25	15.0	10.4	185.0	210.4	0.4	62	0.225

Which dimension is on your drawing?

This is the trap the page exists for. The gap between the first two readings is 39.0 mm on a 50 mm shaft — roughly eight times the shaft diameter’s own tolerance and more than a hundred times the ±0.1 mm the depth itself is usually held to — so there is no chance of confusing them by accident on a good drawing and every chance of confusing them on a bad one. A depth figure of the same order as the key height is t₁ or t₂; a depth figure of the same order as the shaft diameter is d − t₁ or d + t₂. The third row is the one that catches inspectors, and it is a measurement problem rather than a drawing one: a micrometre whose anvil is wider than the keyway — which is almost any micrometre on almost any keyway — bridges the slot and rests on the shoulders either side, so it reads the full shaft diameter and not d − t₁ at all. There is no partial reading to interpret. Use a depth micrometre with a narrow foot, a depth gauge, or a pin of known diameter laid in the keyway and measured over.
What the drawing says	What it means	For a 14 × 9 key on a 50 mm shaft
t₁ = 5.5, or “keyway depth 5.5”	Measured from the shaft’s outside surface down to the keyway floor, on the keyway centreline. The metric convention, and what DIN 6885-1 and ISO 773 tabulate	5.5 mm
44.5, or “S = 44.5”, or a dimension arrowed right across the shaft	Measured from the keyway floor across to the diametrically opposite surface of the shaft. d − t₁. What ANSI B17.1 tabulates as its S dimension, and the one you can inspect without knowing the shaft’s actual diameter	44.5 mm
50.0 — that is, the plain shaft diameter	What a micrometre reads if its anvil is wider than the keyway. The anvil bridges the slot and rests on the two shoulders either side of it, so it never reaches the floor and measures the shaft, not the keyway. On a 50 mm shaft those shoulders stand 1.00 mm proud of a chord across the keyway, so there is no halfway reading — it is the full diameter or nothing	50.00 mm
t₂ = 3.8	Hub keyway, measured from the bore surface outwards to the keyway floor	3.8 mm
53.8, or “T = 53.8”	Hub keyway, measured from the keyway floor across to the opposite side of the BORE. d + t₂, ANSI’s T dimension, and for a taper keyseat it is measured at the deeper end	53.8 mm

The three fit classes worked out for a 14 mm key, in micrometres

Read the last two columns together, because they are the point. Not one of these three classes is an interference fit in the sense that a press fit is. The class the standard calls close or interference pairs a P9 keyway with an h9 key, and at the tight end of both bands that is 61 µm of interference on a 14 mm key — but at the loose end of both bands it is 25 µm of CLEARANCE. All three keyway fits are transition fits; they differ in where the transition sits, not in whether there is one. That is why a “tight” key still rattles occasionally and why a reversing drive needs an interference fit on the HUB BORE rather than a tighter keyway. The bands here are computed from ISO 286 rather than copied: IT9 is 43 µm at 14 mm, N9’s upper deviation is −12 µm because n’s fundamental deviation is +12, and P9’s is −18 because p’s is +18. 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.
Fit class	Key width	Shaft keyway	Hub keyway	Shaft: tightest	Shaft: loosest	Hub: tightest	Hub: loosest	CAN be an interference?	MUST be an interference?
Free (sliding)	h9	H9	D10	0	86	50	163	no	no
Normal	h9	N9	JS9	-55	31	-22	65	yes	no
Close (interference)	h9	P9	P9	-61	25	-61	25	yes	no

Woodruff keys — a different family, for a tapered or stepped shaft end

A Woodruff key is a segment of a disc sitting in a semicircular pocket milled with a Woodruff cutter, and it exists for one reason: it self-aligns in a tapered bore and it cannot fall out while you assemble the hub, which a parallel key in a vertical shaft end certainly can. The price is depth. Compare a 6 × 11 × 28 Woodruff against the 6 × 6 parallel key for the same 17–22 mm shaft: the Woodruff pocket is about 8.5 mm deep against 3.5 mm, so it removes more than twice as much of the shaft section and the stress concentration at the pocket end is worse. The last two columns carry the same clearance identity as the parallel keys, and it holds at fifteen of these sixteen rows — the 6.0 × 11.0 × 28 row prints t₁ = 7.5 mm where the identity wants 8.5. That row is flagged, not quietly corrected: check it against your own supplier’s drawing before cutting. 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.
Designation b × h₁ × D	b (mm)	h₁ (mm)	D (mm)	t₁ shaft (mm)	t₂ hub (mm)	t₁ + t₂ − h₁
1 × 1.4 × 4	1.0	1.4	4	1.0	0.6	0.2
1.5 × 2.6 × 7	1.5	2.6	7	2.0	0.8	0.2
2 × 2.6 × 7	2.0	2.6	7	1.8	1.0	0.2
2 × 3.7 × 10	2.0	3.7	10	2.9	1.0	0.2
2.5 × 3.7 × 10	2.5	3.7	10	2.7	1.2	0.2
3 × 5 × 13	3.0	5.0	13	3.8	1.4	0.2
3 × 6.5 × 16	3.0	6.5	16	5.3	1.4	0.2
4 × 6.5 × 16	4.0	6.5	16	5.0	1.8	0.3
4 × 7.5 × 19	4.0	7.5	19	6.0	1.8	0.3
5 × 6.5 × 16	5.0	6.5	16	4.5	2.3	0.3
5 × 7.5 × 19	5.0	7.5	19	5.5	2.3	0.3
5 × 9 × 22	5.0	9.0	22	7.0	2.3	0.3
6 × 9 × 22	6.0	9.0	22	6.5	2.8	0.3
6 × 11 × 28	6.0	11.0	28	7.5	2.8	-0.7	← fails the identity
8 × 11 × 28	8.0	11.0	28	8.0	3.3	0.3
10 × 13 × 32	10.0	13.0	32	10.0	3.3	0.3

The staircase, the two depth conventions, and why no key fit is really an interference fit

This page gives the dimensions. It does not give the capacity. How much torque a key can carry is a shear and bearing calculation on the key’s flanks, and it has its own page on this site; what this one answers is the question that comes first and gets answered wrong more often — what size key does a shaft of this diameter take, how deep are the two keyways, and which of the three numbers on the drawing is the depth.

The size is a staircase, and the staircase is the definition. DIN 6885-1 and ISO 773 do not give a formula for the key width; they give a table of shaft diameter bands, and every shaft in a band gets the same key. A 44.5 mm shaft and a 50 mm shaft both take a 14 × 9. That matters because the key width is between a fifth and a quarter of the shaft diameter depending on where in the band you sit, so a shaft at the bottom of a band loses more of its section than one at the top. It also means the familiar “the key is about a quarter of the shaft” is an average of a staircase rather than a rule, and the series above shows exactly where the steps fall. Two of them are worth knowing: the 12 × 8 and the 25 × 14 add width without adding height, so those two rows buy bearing area and not depth.

The depth trap is real and it is dimensional, not conceptual. A shaft keyway can be dimensioned two ways. The metric standards tabulate t₁, measured from the shaft’s outside surface down to the keyway floor: 5.5 mm on a 50 mm shaft. American practice dimensions the same keyway as the distance from the floor across to the opposite side of the shaft — ANSI B17.1 calls it S, and Amesweb quote the definition as “the distance from the bottom of the shaft keyseat to the opposite side of the shaft” — which is 44.5 mm on the same shaft. The hub is the same story: t₂ = 3.8 mm from the bore, or T = d + t₂ = 53.8 mm across it. The two readings differ by 39 mm on a 50 mm shaft, so no competent drawing confuses them; an incompetent one does, and a keyway cut 44.5 mm deep into a 50 mm shaft is a shaft in two pieces. The calculator prints both for whatever you enter, so you can match the number on the paper in front of you.

And there is a third reading, which is the measurement trap rather than the drawing one. Put an ordinary micrometre across a keyed shaft and it will not give you d − t₁. Its anvil is wider than the keyway, so it bridges the slot and rests on the two shoulders either side of it — which on a 50 mm shaft stand 1.00 mm proud of a chord across a 14 mm keyway — and what it measures is the shaft, not the keyway. The reading is the full diameter. There is no partial figure to interpret and no correction to apply: it is d − t₁ if the contact genuinely reaches the floor and d if it does not. Use a depth micrometre with a narrow foot, a depth gauge, or a pin of known diameter laid in the keyway and measured over. (An earlier draft of this page claimed a chordal correction to d − t₁ for a bridging anvil. That was wrong and it is withdrawn: the keyway floor is a plane and the opposite contact is a tangent plane, so an anvil that does reach the floor reads d − t₁ exactly.)

The fit classes are all transition fits, which surprises people. The key is h9 in every class; the keyways change. The sliding fit puts H9 on the shaft and D10 in the hub and is always a clearance. The normal fit puts N9 on the shaft and Js9 in the hub. The close or interference fit puts P9 in both. Work the bands out from ISO 286 — which this page does rather than copying a printed table — and the close fit on a 14 mm key runs from 61 µm of interference at the tight end of both tolerance zones to 25 µm of CLEARANCE at the loose end. It is not a press fit. If you need a connection that genuinely cannot move, the answer is an interference on the hub bore, a taper bush, a shrink disc, or a spline — not a tighter keyway. The interference fit and shrink fit calculator does the hub-bore version, and the spline torque capacity calculator does the alternative. What a keyway does to the shaft’s fatigue strength is on the shaft fillet and stress concentration calculator.

Length, and the number that is quoted in both directions. Calcimator publish “the engaged portion should be at least 1 to 1.5 times the shaft diameter”, a floor. EICAC publish “the key length should be less than about 1.5 times the shaft diameter to ensure a good load distribution over the entire key length when the shaft becomes twisted when loaded in torsion”, a ceiling — and the ceiling comes with the physics, which is the more useful half. The shaft twists under torque, so the far end of a long key sees less relative rotation and carries less load; past about 1.5 diameters you are removing shaft section without buying capacity. This batch found no source for the 1.25 × figure that circulates. Surface finish class and roughness grades belong to the converters plugin, not here.

Frequently asked questions

What size key does a 50 mm shaft take?

A 14 × 9 mm parallel key to DIN 6885-1 / ISO 773, with a 5.5 mm deep keyway in the shaft and a 3.8 mm deep keyway in the hub. Fifty millimetres sits at the top of the 44 to 50 mm band, so the same key serves a 44.5 mm shaft — and on that shaft it is a larger fraction of the section. The standard steps rather than interpolating, and the calculator follows the steps.

Is keyway depth measured from the shaft surface or across the shaft?

Both conventions are in use and that is the trap. The metric standards, DIN 6885-1 and ISO 773, tabulate t₁ from the shaft’s outside surface down to the keyway floor — 5.5 mm on a 50 mm shaft. ANSI B17.1 tabulates S, the distance from the keyway floor across to the opposite side of the shaft — 44.5 mm on the same shaft. The hub is t₂ from the bore surface, or T = d + t₂ across the bore. A depth figure of the same order as the key height is t₁ or t₂; one of the same order as the shaft diameter is d − t₁ or d + t₂. This page prints all four.

How do I actually measure keyway depth?

Not with an ordinary micrometre across the shaft. Its anvil is wider than the keyway, so it bridges the slot and rests on the two shoulders either side — on a 50 mm shaft those shoulders stand 1.00 mm proud of a chord across a 14 mm keyway — and the reading is the full shaft diameter, not d − t₁. Use a depth micrometre with a narrow foot, a depth gauge, or a pin of known diameter laid in the keyway and measured over, subtracting the pin. Note that a contact which genuinely reaches the flat floor reads d − t₁ exactly — the floor is a plane and the opposite contact is a tangent plane, so there is no chordal correction to make.

Which key fit should I use — free, normal or close?

Normal for almost everything: N9 keyway in the shaft, Js9 in the hub, key h9, and the key goes in with light tapping. Free (H9/D10) only when the hub has to slide along the shaft under load, such as a shifting gear. Close (P9/P9) for reversing drives and high shock, but understand what you are buying: the calculator shows that even the close fit can come out with clearance at the loose end of both tolerance zones. None of the three is a press fit.

How long should a key be?

Between about one and one and a half shaft diameters of engaged length, and the upper figure matters more than the lower. The shaft twists under torque, so a key much longer than 1.5 diameters has a far end that sees little relative rotation and carries little load, while the keyway it sits in has still removed shaft section all the way along. If the capacity is not there at 1.5 diameters the published answers are two keys at 180°, a bigger shaft, or a keyless connection — not a longer key.

What about Woodruff keys?

A different family, DIN 6888 / ISO 3912 / BS 4235-2, and a different job: a disc segment in a semicircular pocket, which self-aligns in a tapered bore and cannot fall out during assembly. The price is depth — a 6 × 11 × 28 Woodruff pocket is about 8.5 mm deep where the parallel key for the same shaft needs 3.5 — so it removes far more section and concentrates stress worse. The table on this page carries the sizes with both keyway depths, and flags the one published row whose depths do not satisfy the same internal check as all the others.

Does this page calculate how much torque the key can take?

No, deliberately. Key shear and bearing capacity is a separate calculation with its own page on this site, and duplicating it here would split the subject. This page is the dimensions: what size, how deep, which convention, what the fit actually delivers, and how much hub wall is left. Get those right first — a capacity calculation on the wrong key size is worse than no calculation.

Related calculators

References

DIN 6885-1:1968, Mitnehmerverbindungen ohne Anzug — Parallel keys, keyways, deep pattern, and the equivalent ISO/R 773. Cited by number; the standards are copyrighted and were not fetched. The dimensions printed here are attributed to the catalogues below, and the internal structure of the table — that t₁ + t₂ − h is a three-step ladder of 0.2, 0.3 and 0.4 mm keyed on the key width — is computed here, not quoted.
AIMS Industrial Solutions. Keyways & Keys: Parallel Keys, Woodruff Keys, Key Steel & Size Charts. The source for the nineteen ISO 773 / DIN 6885-1 rows on this page — shaft diameter band, key b × h, shaft keyway depth t₁ and hub keyway depth t₂ — and for the DIN 6888 / ISO 3912 Woodruff designations. Every row was checked against the top-clearance identity before use and all nineteen satisfy it exactly.
Aspen Fasteners. Metric DIN 6885 Parallel Keys High Type “A” specification sheet. The independent check on the b × h ladder: it carries the width and height columns only, agrees with AIMS at all nineteen shared rows, and continues the ladder to 100 × 50 for a 500 mm shaft. Used for nothing else, because it has no depth columns.
National Machine Tool (keyseaters.com). Keyway Tolerance Chart: ANSI B17.1 & DIN 6885 Fit Classes. The source for the three fit classes and the tolerance class on each member: h9 on the key in every class; sliding H9 shaft with D10 hub; normal N9 shaft with Js9 hub; tight P9 shaft with P9 hub. It also carries the ANSI B17.1 class 1 / 2 / 3 scheme, which is a different construction entirely — it works from bar-stock against key-stock tolerances rather than from ISO 286 letters.
Miki Pulley. Parallel Key and Keyway Sizes and Tolerances and Dimensions and Tolerances for Parallel Keys and Keyways (JIS B 1301). The second, independent source for the keyway width classes: it prints N9 on the shaft for the normal fit and P9 on the shaft for the tight fit, agreeing with National Machine Tool. Its hub column came through the fetch as Js9 for BOTH fits, which cannot be right — a tight fit with a symmetric hub tolerance is not a tight fit — so the hub class is taken from National Machine Tool and this reference is used only to corroborate the shaft classes.
Amesweb. Depth Values for Shaft Keyseats — Hub Keyways (ANSI B17.1-1967). The source for what the American convention actually dimensions: “the distance from the bottom of the shaft keyseat to the opposite side of the shaft is specified by dimension S”, and “the distance from the bottom of the hub keyway to the opposite side of the hub bore is specified by dimension T. For taper keyseats, T is measured at the deeper end.” That is the d − t₁ and d + t₂ convention, and it is why this page prints both.
RoyMech. Woodruff Key Dimensions, Sizes and Keyway Data, to BS 4235-2:1977. The source for the sixteen Woodruff rows, with the shaft depth t₁ and hub depth t₂ this page needs. Fifteen of the sixteen satisfy the same top-clearance identity as the parallel keys; the 6.0 × 11.0 × 28 row prints t₁ = 7.5 mm where the identity wants 8.5, and the page flags that row rather than using it.
EICAC. Metric Keyway Calculator. The source for the key length CEILING and, more importantly, for the reason: “the key length should be less than about 1.5 times the shaft diameter to ensure a good load distribution over the entire key length when the shaft becomes twisted when loaded in torsion”. Note that this is the same number that other publishers give as a minimum.
Calcimator. Keyway Design Calculator — Shear, Bearing & Key Length. The source for the key length FLOOR: “the engaged portion should be at least 1 to 1.5 times the shaft diameter”. Printed beside EICAC’s ceiling on this page because the two publishers use the same figure in opposite directions and the reader deserves to know that.
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 clause; not fetched. Its defining relations are what this page computes from: the standard tolerance factor i = 0.45 D^(1/3) + 0.001 D at the geometric mean of each size band, the grade multipliers IT6…IT11 = 10i, 16i, 25i, 40i, 64i, 100i, and the fundamental-deviation rules for the interference letters. The table values were checked against those relations rather than trusted, and the two relations disagree below 6 mm because the standard rounds small sizes to convenient numbers.
Dalloway Precision. ISO 286 Tolerance Grade Chart — IT5 to IT11 in Microns. The source for the IT table. Its column HEADINGS come through in the order IT5 IT6 IT8 IT7 while its VALUES are in the order IT5 IT6 IT7 IT8; the R5 ratio settled which reading is right, because each grade step must be about 10^(1/5) = 1.585 and 10 → 1.67 satisfies that where 14 → 2.33 does not. Every one of the 78 values was then checked against the tolerance-factor formula.
Impex Industrial Corporation. Parallel Key Size Chart by Shaft Diameter — DIN 6885 / IS 2048. Fetched and NOT used: the extract returned the hub keyway depth t₂ in a column headed “h” and the shaft depth t₁ in a column headed “t4”, so on the face of it the key height for a 22–30 mm shaft was 3.3 mm rather than 7 mm. The top-clearance identity identified the shift immediately, because the “h” column satisfied h = t₁ + t₂ − 0.3 instead of being h. Recorded here because it is the single most common failure mode of a standards table taken off the web.