Measuring the two numbers that matter
The major diameter is measured across the crests of the thread with calipers, on a length of thread that has not been chewed by a wrench or a stripped nut. Expect the reading to come in slightly under the nominal size: an M8 bolt is nominally 8.00 mm, and a real one usually measures somewhere around 7.8 to 7.9 mm because the specification puts the manufacturing tolerance entirely below nominal, never above. A nominal-size fastener that measures over its nominal diameter has something on it — paint, plating applied after threading, or corrosion product.
The pitch is the harder measurement and the one people skip, which is why so many fasteners get misidentified. A pitch gauge is the direct way: a set of leaves, each cut with a known thread form, that you offer up until one seats with no light showing. Without one, lay the fastener against a rule, line the rule up with a crest, count the gaps between crests over as long a run as you have, and divide the measured distance by the count. Counting gaps rather than crests is the part that goes wrong — ten crests span nine gaps. This tool takes the count in that form and converts it both ways.
Diameter alone identifies nothing
This is the reason for the whole page. Coarse and fine threads of the same nominal size are identical across the crests and completely incompatible in a hole. M8 × 1.25 and M8 × 1 both measure 8 mm. So do 1/4-20 and 1/4-28, and 1/2-13 and 1/2-20. Threading a fine bolt into a coarse hole binds after a turn or two, and forcing it destroys the tapped hole, which is usually the more expensive half of the assembly.
Worse, some imperial and metric sizes are close enough to start. M6 is 5.94 mm at nominal against 1/4 inch at 6.35 mm, which will not start; but 3/8-16 at 9.525 mm and M10 at 10 mm are close enough that a worn 3/8 bolt sometimes enters an M10 hole for a thread or two before it locks. The pitch numbers are what separate them — 1.588 mm against 1.5 mm — and no amount of feel at the wrench will.
Reading the head, which is description not identification
| Feature | What it tells you |
|---|---|
| Hex head, six flats | Wrench size across flats. Related to the thread by convention only, and the convention differs between standards, so measure it rather than deducing it |
| Socket cap, cylindrical head | Hex key drive, sized in the fastener system: metric keys in metric caps, inch keys in inch caps |
| Countersunk / flat head | Sits flush. The included angle differs between systems — 82 degrees is the common imperial figure and 90 degrees the metric one, which is why a mixed pair sits proud or bottoms early |
| Radial lines on a hex head | An SAE grade marking. Three lines and six lines are the two commonly met. It identifies the grade; the strength figures behind that grade are outside what this page will state |
| Stamped number such as 8.8 or 10.9 | A metric property class. Again identification only |
| No marking at all | Commodity fastener with no declared class, or a head too small to mark. Treat an unmarked fastener as unknown rather than as low grade |
Head type, drive type and length are all independent of the thread designation. A 1/4-20 fastener can be a hex head, a socket cap, a button head, a flat head or a carriage bolt, and none of that changes what hole it fits.
What this table deliberately leaves out
Pipe threads are the big omission and the most common source of confusion. NPT is tapered — the diameter changes along the thread, sealing by interference — and it is designated by a nominal pipe size that has no relationship to any measurement you can take with calipers. A 1/2 inch NPT thread measures about 0.84 inches across. BSPP and BSPT are a separate family again with a different thread form angle. None of them belong in a machine screw table and none of them are here.
Self-tapping, sheet metal, wood and drywall screws are also absent. They are specified by a gauge number and a threads-per-inch figure, they have thread forms designed to cut or form their own thread rather than match one, and matching them against a machine screw table produces a plausible answer that is wrong. Acme, trapezoidal and buttress threads for lead screws are out too, as are British Association, Whitworth and BSF, which turn up on older equipment and use a different flank angle from Unified threads even where the diameters coincide.
Once you know what it is
If the next step is drilling and tapping a matching hole, the tap drill size calculator takes the designation and gives the tap drill for the engagement you want, plus clearance drills for the through hole. If the next step is turning it, the wrench and hex key size chart covers what fits the head and where the deceptively close metric and inch pairs are. And if it is going into a joint that has to hold a specified load, that is a question for the drawing or the manufacturer instruction, not for a chart.
Questions people ask
My bolt measures 7.85 mm. Is it M8 or damaged?
Almost certainly a perfectly normal M8. Thread tolerance is one-sided: nominal is the maximum, and production sizes sit below it, so a healthy M8 typically measures around 7.8 to 7.9 mm across the crests. Rolled threads in particular come out a little under. What would be suspicious is a reading above 8 mm, which means something is on the surface rather than that the thread is oversize.
How do I tell coarse from fine without a pitch gauge?
Count over a measured run. Lay a steel rule along the thread, align a graduation with one crest, and count the gaps between crests to another graduation as far along as the thread allows. Divide the distance by the number of gaps. Over ten gaps the arithmetic is easy and the error from a slight misalignment is divided by ten. A phone photo zoomed in helps a lot on small fasteners, since the counting is the error-prone half.
Two entries came back within a percent of each other. Which is it?
When an imperial and a metric candidate both land close, the pitch is the discriminator, not the diameter, because the diameters were always going to be near. Measure the pitch over the longest run you can and compare against both figures shown. If they are still too close to call, thread it into a known nut of one system by hand — a wrong-system thread will feel gritty and stiff within two turns, and you should stop there rather than forcing it.
Does it cover left hand threads?
The designations here are all right hand, which is the overwhelming default. A left hand thread has the same diameter and pitch and is identified by the direction the helix runs, which a caliper cannot see — look at the thread with the fastener pointing away from you and see which way the crests climb. Left hand threads carry an LH suffix in a designation and turn up in specific places, such as one side of a turnbuckle or a spinning assembly where normal rotation would loosen a right hand thread.
Can I get a torque figure from the designation?
Not from this page, and not from a designation alone anywhere. Tightening torque depends on the grade or property class of the fastener, the clamped material, whether the threads and the bearing face are dry, oiled or coated, whether a washer is present, and what the joint is supposed to achieve. The same M8 bolt has very different correct torques in two different joints. Where a value matters, it comes from the assembly instruction or the engineer responsible for the joint.