High-strength bolting and RCSC inspection for structural steel

High-strength bolting, and why the method makes the joint

High-strength bolting is the practice of installing heat-treated structural bolt assemblies to a known clamping force so a steel connection behaves the way the engineer designed it. The bolt is only half the story. A bolt lying in a keg has a tensile strength stamped on the head and means nothing structurally until it is installed, and how it is installed is what turns a hole and a fastener into a joint that carries load.

That is the part people new to steel miss. The grade of the bolt tells you what it can do. The installation tells you what it is doing. The same A325 bolt can be run snug for a bearing connection, pretensioned for a joint that cannot be allowed to loosen, or pretensioned with a prepared faying surface for a slip-critical connection that carries load through friction. Three different jobs, one bolt, and the difference lives entirely in the installation and the surface, not in the part number.

On a datacenter or mission-critical job this matters because the steel holds equipment that has to stay anchored and operational through a seismic event, and the connections that resist that event are bolted as often as they are welded. The Research Council on Structural Connections, the RCSC, publishes the Specification for Structural Joints Using High-Strength Bolts, and that document is the rulebook for how these joints get installed and inspected. The bolting inspector's job is to confirm the right method was used on the right joint and to have the record that proves it.

The bolt assembly: ASTM F3125

ASTM F3125 is the consolidated specification that pulled the old separate standards into one document. It absorbed A325, A325M, A490, A490M, F1852, and F2280, and it carries them forward as grades inside a single specification. So the bolt the trade still calls an A325 is now Grade A325 under F3125, and the same goes for A490. The legacy names did not vanish from the field, but the governing document changed, and a submittal that cites F3125 by grade is citing it correctly.

The two strength levels are what you actually choose between. Grade A325 is the 120 ksi minimum tensile bolt, and Grade A490 is the 150 ksi minimum tensile bolt, the higher-strength alloy. F1852 and F2280 are the twist-off tension-control versions of those same two strength levels, A325 and A490 respectively, with a splined end built for installation. Sizes run from 1/2 in to 1-1/2 in in the inch series. Verify the exact grade, type, and finish against the adopted F3125 edition and the project, because chemistry and finish options vary by grade.

A bolt is never just a bolt on this work. It is an assembly: the bolt, a heavy hex nut to a matching ASTM specification, and washers where the method or the geometry requires them. The assembly is also lubricated or coated as supplied, and that lubrication is part of how the torque turns into tension. Mix a nut from one lot onto a bolt from another, strip the lube, or substitute a washer the procedure did not call for, and you have changed the assembly the qualification was based on.

The lot and the rotational-capacity test

High-strength bolts are bought, tested, and traced by lot, and the lot is the unit that matters for everything downstream. A lot is a production run of bolts, nuts, or washers that share a heat and a process, and the assembly as installed is only as good as the matched components in that lot. When a problem turns up, the question is always which lot, because the fix and the quarantine follow the lot, not the individual bolt.

The manufacturer's rotational-capacity test is the assembled-lot proof. It runs the bolt, nut, and washer together as a unit and confirms the assembly can be tensioned through the required rotation without stripping or twisting off prematurely, and that the lubrication is doing its job. The mill test report and the rotational-capacity results travel with the lot, and the inspector checks that the certs match the bolts that actually showed up. A keg with no traceable lot and no certs is a keg you cannot use on inspected work.

Keep the components matched as supplied. The bolt, nut, and washer that were tested together as a lot are the assembly that performs as tested. Pulling a nut from a different bin to finish a connection breaks that match, and on a job where the rotational-capacity test is the basis for acceptance, it breaks the acceptance with it.

What is the difference between snug-tight, pretensioned, and slip-critical joints?

The RCSC specification defines three joint types, and the joint type, not the bolt, sets how the bolt has to be installed. A snug-tight joint needs only the plies brought into firm contact. A pretensioned joint needs the bolt installed to a specified minimum tension. A slip-critical joint needs that same pretension plus a prepared faying surface, so the connection carries load through friction between the plies rather than by the bolts bearing on the holes.

Which type a connection is is a design decision the engineer of record makes and shows on the drawings, and it is not the field's call to change. Snug-tight is allowed wherever pretensioned and slip-critical are not required, and it covers a large share of ordinary bearing connections in statically loaded structures. Pretensioning is required where the joint cannot be allowed to loosen or where the code demands it, including many connections in the seismic and wind force-resisting systems and joints subject to fatigue or load reversal. Slip-critical is specified where any slip into bearing is unacceptable, such as joints with oversized or slotted holes, fatigue-loaded joints where slip would matter, and connections the EOR designates.

The trap is treating all three as the same work because they use the same bolt. They do not. A slip-critical joint that was run snug is a defective joint that looks finished, and a pretensioned connection installed without a verified method is a connection nobody can prove. The bolting requirements on a mission-critical job run inside the broader special-inspection program for the structure, where high-strength bolting is one of the listed items the inspector witnesses and reports.

What is a snug-tight joint?

A snug-tight joint is one where the connected plies have been pulled into firm contact, with no specified minimum bolt tension beyond that. The RCSC describes snug as the condition reached by the full effort of an ironworker on an ordinary spud wrench, or a few impacts of an impact wrench, enough to bring the steel together and seat the connection. It is a defined condition, not a loose guess, and it is also the starting point for every pretensioning method.

Snug-tight is for bearing connections that are not subject to fatigue, load reversal, or the kind of vibration that backs a nut off. In those joints the bolt carries shear by bearing against the side of the hole, and a controlled high pretension buys nothing the design needs. Specifying pretension where snug would do just adds cost and inspection for no structural gain, which is why the RCSC tells you to use snug-tight wherever the higher types are not required.

The honest failure here is the gap. Snug means firm contact, and on a multi-bolt connection the bolts farther from the stiff part of the joint can leave the plies apart while the first ones look tight. Bring the whole joint into contact, working the bolts systematically, or you have a snug-tight joint on paper with daylight in the steel.

What are the RCSC pretensioning methods?

The RCSC specification provides four pretensioning methods, and all of them start the same way, by bringing the joint to snug-tight first. From snug, the methods are turn-of-nut, calibrated wrench, twist-off-type tension-control (TC) bolts, and direct-tension-indicator (DTI) washers. Each one is a different way to put a known minimum tension in the bolt, and each one is verified differently.

The methods split into two families by how they prove tension. Turn-of-nut and calibrated wrench control the installation: you do something to the nut, a measured rotation or a calibrated torque, and the tension follows from it. TC bolts and DTI washers are indicating methods: a physical feature on the assembly changes when the bolt reaches tension, the spline shears or the bumps crush, and you read the indicator. Both families are accepted by the RCSC. The choice is usually the fabricator's and erector's, balanced against what the inspector can verify on that joint.

Whatever method is chosen, the rule underneath all four is the same. The job is the minimum specified pretension in the bolt, and the method is just the means. The inspector accepts the method as the RCSC defines it for that method, not by re-torquing every bolt to a number, which is the most common misunderstanding on the deck.

Turn-of-nut: the rotation past snug

Turn-of-nut is the oldest and, on a lot of work, the most reliable method, because it relies on the geometry of the bolt stretching rather than on friction. You bring the joint to snug, mark the nut and the protruding bolt end with a match-mark, then rotate the nut a specified additional amount while holding the bolt from turning. The bolt stretches into its elastic range as the nut advances, and that stretch is the pretension.

The required rotation depends on the length of the bolt relative to its diameter, because a longer bolt stretches more for the same turn. For bolts up to about four diameters long with faces normal to the axis, the RCSC calls for 1/3 turn, 120 degrees. From about four to eight diameters it is 1/2 turn, 180 degrees. From about eight to twelve diameters it is 2/3 turn, 240 degrees. Sloped outer faces add to the required rotation. These are the prequalified values, so confirm the exact rotation against the adopted RCSC table for the bolt length, geometry, and any beveled surfaces on your joint.

The match-mark is the genius of the method and the thing crews skip. A paint or marker line across the nut, the washer, and the steel before the final rotation lets anyone, the next shift or the inspector days later, see exactly how far the nut turned. No mark, no proof. A turn-of-nut joint without match-marks is an assertion, not an installation, and the inspection leans on those marks more than anything else.

Is torque enough to accept a high-strength bolt?

No. Torque alone is not acceptance, and that is the single most important thing to understand about the calibrated-wrench method. The spec is tension in the bolt. Torque is only a proxy, and a sloppy one, because the relationship between torque and tension swings with the lubrication, the thread condition, rust, and dirt. The same torque on a clean lubricated assembly and a weathered dry one puts wildly different tension in the two bolts.

That is why the calibrated-wrench method is built on the Skidmore-Wilhelm tension calibrator, not on a torque chart. The Skidmore is a hydraulic device that reads the actual tension a bolt develops as you tighten it, so you can find the torque that produces the required tension for that specific lot of bolts, with their lubrication, on that day. The RCSC requires this calibration to be done daily, and again whenever the lot, the assembly, or the conditions change, on at least three bolts of each diameter from the lots in use.

Once the wrench is calibrated to the tension on the Skidmore, you install to that torque. The acceptance still traces back to tension, through the daily calibration record, not to a number off a generic chart. A crew running a published torque value with no Skidmore calibration is not running the calibrated-wrench method. They are guessing, and the guess fails the moment a lot of dry, rusty bolts comes onto the deck.

Twist-off tension-control (TC) bolts

A twist-off TC bolt, F1852 at the A325 strength level or F2280 at the A490 level, builds the indicator into the bolt itself. The end of the bolt has a splined tip beyond the threads, separated by an annular groove. A special wrench grips the splined tip and the nut and turns them against each other, and when the bolt reaches the design tension the torque in the groove exceeds its strength and the spline shears off. The sheared spline is the signal that tension was developed.

Inspection is fast when the method is run right. The accepted bolt has its spline gone, and the nut is not spinning free, which would mean the bolt failed to develop tension or stripped. The inspector walks the joint looking for snapped splines and confirms the snug-then-tension sequence was followed for the whole connection, not just the last bolt. The collected splines in a coffee can at the base of the column are a rough tally that the bolts were run.

The weakness is weathering and the gap. TC bolts depend on the as-supplied lubrication to shear at the right tension, and a lot that has sat outside, rusted, or lost its coating can shear at the wrong tension, low or high, so the snapped spline lies. The RCSC handles this by requiring the same pre-installation verification on a Skidmore for TC bolts as for any method, and weathered lots get re-verified or rejected. And like every method, TC bolts only develop full tension if the plies were brought into firm contact first. Shear the spline on a joint with a gap and you have a snapped spline and a loose connection.

Direct-tension-indicator (DTI) washers

A direct-tension-indicator washer, ASTM F959, is a hardened washer with raised bumps, or arches, pressed into one face. As the bolt is tensioned, the bumps flatten, and the gap under them closes by a controlled amount. The DTI is effectively a little load cell under the head: when the gap has closed enough, the bolt has reached the required tension. You read the indicator, you do not read a torque.

Verification is the feeler gauge. You try to slide a feeler of the specified thickness, commonly 0.005 in for a coated or project-specific condition, into the remaining gaps between the bumps. The bolt is acceptable when the gauge is refused in the required number of gaps, generally more than half of them, which means the bumps compressed enough to develop tension. Refusal in too few gaps means the bolt is under-tensioned. The exact gauge thickness and the number of required refusals depend on the DTI type and finish, so confirm them against the manufacturer's instructions and the adopted RCSC and F959 requirements.

Orientation is the field error that quietly ruins DTIs. The washer goes in with the bumps facing the right way and under the correct element, and a hardened flat washer is used under the turned element so the bumps are not chewed up by a spinning nut. Put the DTI in backward, or let the turned nut bear directly on the bumps, and the gaps read wrong. Check the orientation as the joint goes together, because once it is tensioned the only fix is to take it apart.

The faying surface and why slip-critical cares about it

The faying surface is the contact area between the plies, the surfaces that press together when the bolt is pretensioned. In a slip-critical joint that surface is the load path. The pretension clamps the plies, friction between the faying surfaces resists slip, and the load transfers across the joint without the bolts ever bearing on the holes. That is why slip-critical cares about the surface and snug-tight bearing joints do not: in a bearing joint the steel takes load by contact at the hole, and the surface finish between the plies is irrelevant.

Slip resistance is classified, and the class drives how the surface has to be prepared. Clean mill scale or surfaces blast-cleaned to a lower standard give a Class A surface, a slip coefficient commonly taken as 0.30. Surfaces blast-cleaned to bare metal give a Class B surface, with a higher coefficient commonly taken as 0.50. Hot-dip galvanized surfaces are usually treated as Class A. Specified coatings on the faying surface have to be qualified for a slip class by test, because an unqualified paint can drop the coefficient to near nothing and the joint slips under load.

The field consequences are blunt. You cannot paint, oil, or let overspray land on a slip-critical faying surface unless that coating is qualified for the required class. Mill scale that the spec calls clean has to actually be clean. The drawing assumes a prepared surface and the jobsite delivers whatever the last trade left on the steel, so the slip-critical faying surface is a thing the inspector checks before the joint closes up, not after. Once the plies are together the surface is gone from view and gone from inspection.

Pre-installation verification and storing the bolts

Before any pretensioning method is used in the work, the RCSC requires pre-installation verification of the actual fastener assemblies as delivered, on a tension calibrator. You put representative bolts from the lots you are about to install into a Skidmore, run them by the method you will use, and confirm the assembly develops at least the required tension. For TC bolts you confirm the spline shears at the right tension. For DTIs you confirm the gaps close at the right tension. This is done for each combination of diameter, grade, length, and lot, and repeated daily.

The reason the lot matters this much is that the assembly performs as a system. The bolt, the nut, and the washer were tested together, lubricated together, and certified together as a lot. Verification proves that the specific lots on your deck, in the condition they are actually in, will develop tension by your method. A lot that passed its rotational-capacity test at the mill can still fail pre-installation verification on site if it sat in the rain and the lube washed off.

Storage is the cheap insurance that keeps the verification valid. Keep the assemblies in their sealed containers until use, keep them dry and out of the dirt, and keep the lubrication intact, because the lube is part of the engineered assembly and not a contaminant to wipe off. Take out only what gets installed in a shift, and protect the rest. Bolts left open to the weather are the most common reason a lot that should pass starts failing verification.

Snugging and the tightening sequence

Every method starts at snug, and snug has to be reached across the whole joint before any bolt is pretensioned, or the pretension you put in the first bolts gets relieved as later bolts pull the plies together. The RCSC requires the joint to be brought to snug systematically, working from the most rigid part of the connection toward the free edges, so the plies draw into firm contact everywhere before final tensioning begins.

The logic is the same one that bites people who skip it. Tighten an outer bolt to full tension while the middle of the joint still has a gap, then close that gap by tightening the inner bolts, and the outer bolt you already finished has lost tension because the steel moved under it. Snug the whole pattern first, from stiff to free, then make a second systematic pass for the pretensioning. On a big connection you may snug, then re-snug, before you ever start the final method.

This is why a one-pass blitz with an impact wrench, hitting bolts in whatever order is reachable, produces a connection that looks done and is not. The sequence is part of the procedure, and on inspected work it is part of what gets watched.

How is each pretensioning method inspected?

The inspector's role on high-strength bolting is mostly to observe and verify, not to re-install the joint. The RCSC frames inspection around confirming that the right method was used correctly, and the heart of it is that the inspector observes pre-installation verification and observes the work in progress enough to know the method is being run as specified, then checks the finished joints. The inspector is not required to re-tension bolts that were properly installed by an accepted method.

What gets checked is method-specific. For turn-of-nut, the inspector confirms snug was reached and reads the match-marks to see the required rotation was made. For calibrated wrench, the inspector confirms the daily Skidmore calibration was done and the wrench is set to it. For TC bolts, the inspector confirms the splines are sheared and the nuts are not spinning. For DTIs, the inspector checks the feeler-gauge refusals in the gaps. Across all of them the inspector confirms the joint was brought to firm contact and the faying surface, on slip-critical work, was prepared before closing.

When the installation and the inspection disagree, the RCSC provides an arbitration procedure. A representative sample of bolts is tested, and the result of that testing settles whether the joints are accepted or have to be reworked. The arbitration uses an inspection torque applied to the already-tensioned bolt and looks for movement, which is a check of the as-installed condition, not a re-installation. Run the arbitration by the procedure and document it, because it is the path that resolves the dispute without re-tensioning a whole connection on argument.

Can you reuse high-strength bolts?

Mostly no, and the rule splits by grade and finish. A490 bolts and galvanized A325 bolts are not reused, period. They have enough ductility to be pretensioned once, but not consistently enough to survive a second pretensioning without risk of failing, so once they have been tensioned and backed off they come out of service. This is a hard rule on inspected work, not a preference.

Plain-finish A325 bolts are the limited exception. The RCSC permits reuse of plain A325 bolts when the engineer of record approves it, because they keep enough ductility for more than one pretensioned installation. Even then it is bounded, touching up a nut that was run down and backed off during fit-up is not the same as a bolt that was fully pretensioned, cut loose, and reinstalled in a new joint. When in doubt, treat a bolt that has been pretensioned as spent.

The blunt field version: if you tensioned an A490 or any galvanized bolt and took it back off, it goes in the scrap bucket, not back in a hole. The cost of a new bolt is nothing against the cost of a connection that fails because somebody reused a fastener that had already given up its one good stretch.

Common field problems on the deck

The problems that show up on bolted steel are predictable, and most of them are geometry and discipline rather than the bolt itself. Bolt length is the first one. A bolt too short does not get enough thread engagement past the nut, and a bolt too long bottoms the nut on the thread runout before it clamps, so the connection never reaches tension even though the wrench stalled. The fix is the right grip length for the actual ply thickness, checked before the connection is buttoned up.

Missing washers are next, and they are method-dependent. Some methods and geometries require a hardened washer under the turned element, under a DTI, or over a slotted or oversized hole, and leaving it out changes how the load bears and how the indicator reads. The third recurring problem is the gap: the plies were never brought into firm contact, so the bolt that metered or sheared or turned correctly is still in a joint that is not closed. The pretension is real and the joint is still wrong.

Then there are the weathered TC bolts that will not shear, or shear at the wrong tension, because the lot sat outside and lost its lubrication. A spline that snaps at a tension the rusty assembly happened to reach is a false pass. Catch these at pre-installation verification, not after the splines are in the can. The pattern across all of them is the same: the indicator can look right while the joint is wrong, which is exactly why the method and the sequence get inspected, not just the finished bolt.

The bolting record

The bolting record is what proves, after the steel is clad and the connection is buried, that the joints were installed and inspected to the specification. For each connection or bolt group it ties together the joint type, the bolt grade and diameter, the pretensioning method, the pre-installation verification for that lot, and the inspection result. Without it, a finished joint is just a statement that it was done right, with nothing behind it.

The record has to connect to the lot and to the location. The mill test reports and rotational-capacity results identify the lot, the daily Skidmore calibration identifies that the method was set up that day, and the connection mark on the erection drawings identifies which joint the bolts went into. That chain is what the special inspector, the engineer of record, and the building official as the authority having jurisdiction read at turnover.

This sits inside the larger datacenter structural QA program, alongside the welding and anchorage records, as one of the listed special-inspection items for the building. Keep the bolting documentation aligned with that overall QA package so the steel closes out clean with the rest of the structure, and keep it tied to the same connection marks the rest of the steel package uses.

What to document

Document each connection so a stranger could reconstruct which bolt, in which joint type, was installed by which method and how it was verified. The joint type and the method drive everything else, because they determine what acceptance even looks like, and the lot ties the bolt back to its certs and its verification.

If a joint was reworked, a stripped bolt replaced or a faying surface re-prepared, that belongs in the record with the new lot and the re-inspection. An honest record that shows a flagged joint and its repair is worth more than a too-clean record across a building full of connections.

Common mistakes

• Accepting a bolt on torque alone with no Skidmore calibration, when the spec is tension, not torque.
• Skipping pre-installation verification on the actual lot, then trusting an indicator that may be lying.
• Reusing an A490 or galvanized bolt that was already pretensioned, instead of scrapping it.
• Installing DTIs backward or letting the turned nut crush the bumps, so the feeler-gauge gap reads wrong.
• Treating a snug-tight installation as if it were a pretensioned or slip-critical joint, or the reverse.
• Shearing a TC spline or making the turn-of-nut rotation on a joint whose plies are not in firm contact.
• Running weathered TC bolts that lost their lubrication, so the spline shears at the wrong tension.
• Painting, oiling, or allowing overspray onto a slip-critical faying surface with an unqualified coating.

Field checklist

Standards and references

The RCSC Specification for Structural Joints Using High-Strength Bolts, published by the Research Council on Structural Connections, is the central document. It defines the three joint types, the four pretensioning methods, the snug-tight condition, pre-installation verification, the faying-surface classes, and the inspection and arbitration procedures. AISC adopts it by reference, so the RCSC provisions reach the job through the AISC 360 specification for structural steel buildings and the AISC 303 code of standard practice. Confirm requirements against the adopted edition rather than carrying an old clause number onto a new job.

The bolt assemblies are specified by ASTM F3125, the consolidated standard that absorbed A325, A325M, A490, A490M, F1852, and F2280 and carries them as grades. DTI washers are ASTM F959. The nuts and washers are specified by their own matching ASTM standards, and the steel and base metals are specified separately on the structural drawings. Seismic force-resisting connections carry the additional requirements of AISC 341, the seismic provisions, where the design category triggers them.

The engineer of record sets the joint types, the slip-critical designations, and the inspection extent on the contract documents, and the authority having jurisdiction enforces the adopted building code and its special-inspection requirements under IBC Chapter 17. Where the contract is stricter than the code, the contract controls. Verify the adopted editions and any local amendments before citing a specific clause on a submittal.

Units, terms, and conversions

High-strength bolting carries a vocabulary that means specific things, and the same connection can be described differently across a drawing set, a fabricator's procedure, and an inspection report. Bolt tension and minimum pretension are stated in kips or pounds, bolt strength in ksi, and torque in lbf-ft, but torque is never the acceptance criterion on its own. Rotation is given in fractions of a turn and in degrees, where 1/3 turn is 120 degrees, 1/2 turn is 180 degrees, and 2/3 turn is 240 degrees.

Grades carry their legacy names in the field even though F3125 is the governing specification, so A325 and Grade A325 mean the same bolt, as do A490 and Grade A490. Slip coefficients are dimensionless, commonly 0.30 for Class A and 0.50 for Class B faying surfaces. Confirm any value against the adopted RCSC and ASTM editions and the project documents before you rely on it for acceptance.

ASTM F3125

The consolidated bolt specification carrying A325, A490, F1852, and F2280 as grades

A325 / A490

The 120 ksi and 150 ksi structural bolt grades, now grades within F3125

Snug-tight

Plies brought into firm contact by full effort on a spud wrench or a few impacts

Pretensioned joint

A joint with the bolt installed to a specified minimum tension by an RCSC method

Slip-critical

Pretensioned joint that carries load by friction across a prepared faying surface

Faying surface

The contact area between plies; Class A is 0.30, Class B is 0.50 slip coefficient

Turn-of-nut

Pretensioning by a specified nut rotation past snug, verified by match-marks

DTI

Direct-tension-indicator washer, ASTM F959, whose bumps crush as the bolt tensions

TC bolt

Twist-off tension-control bolt, F1852 or F2280, whose spline shears at tension

Skidmore

Skidmore-Wilhelm tension calibrator that reads actual bolt tension for verification

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What is a special inspection on a construction project?
• How do you calculate weld heat input?
• How much does a high-density AI rack weigh?
• What is structural steel erection?
• What is a rear-door heat exchanger?