Bolted joints
Five Ways to Verify Bolt Tension
Four of the five can only answer at the moment of installation. Only ultrasonic measurement reads a fastener that is already in the structure.
Not an accuracy contest
The methods divide on when, not on how well.
Comparisons of bolt tension methods usually argue about accuracy, and that framing hides the distinction that decides most procurements. Four of the five methods below can only answer at the moment of installation. A direct tension indicator's gap closes once. A tension-control bolt's spline shears once. A hydraulic tensioner is removed. A bench calibrator never touches the bolt that ends up in the structure at all.
That is not a criticism of any of them — it is what they were designed to do, and for a steel frame being erected this week it is the correct and cheaper answer. It becomes a problem only when the question arrives later: what force is in this flange now, on an asset assembled some years ago by a contractor who has moved on. At that point four of the five are no longer available, and the fifth is the only one that reads the fastener itself.
Side by side
What each method evidences, and what it cannot be asked.
| Method | What it measures | When it can be asked | Re-readable later | Governing document | Stated limit |
|---|---|---|---|---|---|
| Ultrasonic measurement | Axial force in one identified fastener, from its own length change | Before, during or years after installation | Yes — the same bolt can be read again | ISO 16047:2005, ISO 898-1:2013 | Needs a reference for that bolt, or a qualified equivalent for the population |
| Bolt tension calibrator (Skidmore-Wilhelm type) | Clamping force developed by a test bolt on a bench | Pre-installation, on a sample from the lot | No — it never touches the installed fastener | RCSC Specification, pre-installation verification | Published bolt sizes run 1/2" to 1-1/4", metric M12 to M30 |
| Direct tension indicator washer | That a specified minimum was reached, by a gap closing | At installation only | No — the indication is consumed | ASTM F959 | A threshold pass, not a value; no reading above or below it |
| Tension-control bolt | That the spline sheared at its calibrated torque | At installation only | No — the spline is gone | ASTM F3125 | Shear point is a lot property; RCSC requires verifying each lot |
| Hydraulic tensioning | The force the tool applied, before transfer to the nut | At installation only | No — residual load after transfer is not the applied load | Joint design specification | Load loss on transfer is estimated, not measured |
Nothing in that table is a ranking. Each row is what its own governing document says the method establishes, and a specification that calls for a DTI is satisfied by a DTI regardless of what any other row can do.
The consumed indication
An indicator that has fired cannot be read again.
ASTM F959 defines a direct tension indicator as a washer with protrusions that compress as load is applied, with the gap measured by feeler gauge against a specified value. It is a threshold: the gap either closed enough or it did not. There is no reading above the threshold and none below it, and once the joint is assembled the indication is spent. The RCSC Specification treats that as sufficient evidence of installation, which for structural steel it is.
A tension-control bolt carries the same logic into the fastener itself: ASTM F3125 covers the assemblies, and the spline shears when the installation tool reaches the torque the assembly was manufactured to shear at. That removes operator judgement, which is the largest single source of variation in torque control. It does not remove the friction condition, which is why the RCSC Specification requires pre-installation verification of each lot rather than accepting the assembly's certificate alone.
Hydraulic tensioning is the one method that applies force directly rather than through friction — and then hands it over. The load that remains after the nut is run down and the tool released is lower than the load applied, by an amount that is estimated from the joint's stiffness rather than measured on the bolt.
A published boundary worth knowing
The bench calibrator's published range stops below a tower flange.
Skidmore-Wilhelm's MZ bolt tension calibrator is the reference instrument for pre-installation verification in structural bolting, and the company publishes the bolt sizes it accepts: 1/2", 5/8", 3/4", 7/8", 1", 1-1/8" and 1-1/4", with the metric equivalents M12, M14, M16, M18, M20, M22, M24, M27 and M30 (skidmore-wilhelm.com product page, retrieved 2026-09-17).
That range covers structural steelwork comprehensively. It stops one size below the fasteners this site is most often asked about: a wind-turbine tower flange commonly runs M36, and the torque chart tabulates to M36 for the same reason. Nothing about that is a fault in the instrument — it was built for a different population of joints. It does mean that for large-diameter flange bolting, the pre-installation verification route familiar from structural steel does not simply transfer, and the verification question has to be answered another way.
The decision
Three questions settle it faster than a specification sheet.
Is the joint already built? If yes, only ultrasonic measurement remains, and even then only if a reference exists for those bolts or a qualified equivalent can be established for the population.
Does the acceptance criterion ask for a value or for a process? A specification satisfied by observed installation is satisfied more cheaply by a DTI or a tension-control bolt than by anything that produces a number. Buying a measurement against a process criterion is buying information the acceptance basis will not use.
Does the record have to be re-examinable? A gap that closed and a spline that sheared leave an inspection record, not a dataset. Where a later engineering review has to reopen the question — a repeat campaign, a dispute, a life-extension assessment — a per-bolt force value is the thing that can be re-read, and the comparison of pretensioning methods sets out how torque control differs from measurement on exactly that point.
Where the answer points at measurement, it can be bought as an instrument — the WIT-BW10 — or as work, through ultrasonic bolt load verification as a service. The measurement guide covers what the method needs before either is worth arranging.
Provenance
Sources and limitations
- ASTM F959 — compressible-washer-type direct tension indicators for structural fasteners. The definition of what a DTI evidences is taken from the standard, not from any manufacturer.
- ASTM F3125 — high-strength structural bolts, including twist-off tension-control assemblies.
- RCSC Specification for Structural Joints (2020) — pre-installation verification and the inspection provisions for each installation method. Published in full and free to read.
- Skidmore-Wilhelm product page — the bolt sizes quoted for the MZ calibrator are the manufacturer's own published list, retrieved 2026-09-17. No performance figure of theirs is characterised here beyond that list.
- ISO 16047:2005 and ISO 898-1:2013 — torque/clamp-force testing and fastener mechanical properties.
- Limitation. This page compares what each method evidences, taken from each method's own governing document. It is not a test of competing products, no product was measured against another, and no accuracy figure is asserted for any method including our own. Comparative performance in a specific joint depends on the fastener, its condition, the calibration actually performed and the procedure the result is read against.
Decision support
Frequently asked questions
What is a bolt gauge, and is it the same as a tension gauge?
The phrase covers two different instruments. A bolt tension calibrator — the Skidmore-Wilhelm type — is a bench device that holds a test bolt and reads the clamping force it develops, used to establish the torque a wrench should be set to. An ultrasonic bolt gauge measures the length change of an individual fastener and converts it to axial force, and it works on a bolt in the structure. The first calibrates a procedure; the second measures a specific installed bolt.
Do DTI washers tell you the tension in a bolt?
They tell you that a specified minimum was reached when the gap closed, which is what ASTM F959 defines them to do. That is a pass at a threshold at a moment, not a value, and it is not re-readable later in any quantitative sense. For structural steel that is usually exactly what the specification asks for. For a joint whose current condition is in question years afterwards, it is not the same information.
Are tension-control bolts more accurate than torque?
They remove the wrench operator's judgement, which is a real improvement, but the spline still shears against a friction condition set when the assembly was manufactured and stored. ASTM F3125 governs the assemblies and the RCSC Specification requires pre-installation verification of each lot, which exists because the shear point is a property of the lot rather than a constant.
Why does ultrasonic measurement cost more per bolt?
Because it produces a different thing. The others are consumables or a one-off calibration that certify a tightening process; ultrasonic measurement produces a force value tied to an identified fastener, and it needs a reference for that bolt and a calibrated conversion. Where a process certificate satisfies the specification, buying a measurement is money spent on information nobody asked for.
Engineering review
If the joint is already built, four of these five are no longer available to you.
Tell us the fastener, the joint, the target preload and its range, whether a baseline exists and the procedure that governs acceptance. We will tell you what can be established before anyone quotes a day rate.