Ultrasonic bolt testing is a contact pulse-echo examination that does two jobs. It finds cracks in bolts, studs and anchor rods with a 0° longitudinal beam sent from the bolt end, and it measures preload by timing the echo before and after tightening under ASTM E1685, which applies to bolts at least 6.3 mm in diameter.
The two jobs use the same transducer on the same end face, but they answer different questions. Flaw detection asks "is there a crack at the threads, at the concrete line or under the head?" Preload measurement asks "how much has this bolt stretched, and what load does that represent?" Mixing them up is the most common mistake in bolt UT specifications, so this guide treats them separately.
Contents: What it is · Finding cracks · Flaw example · Preload · Preload example · Procedure · Standards · Research methods · Limits · FAQ
What is ultrasonic bolt testing, and what can it tell you?
Ultrasonic bolt testing sends a sound pulse down the bolt axis and reads the echoes that come back. Where an echo arrives tells you where a reflector is; when the back-end echo arrives tells you the acoustic length of the bolt.
| Question | Technique | What is measured | Main reference |
|---|---|---|---|
| Is there a crack in the shank, threads or embedded length? | 0° straight-beam longitudinal wave from the end face; low-angle (up to about 10°) beams for off-axis areas | Echoes arriving before the back-end echo, and loss of back-end signal | ASTM E114 practice plus the owner's written procedure |
| Is the bar stock sound before it becomes a bolt? | Straight-beam and angle-beam UT of wrought bar | Internal discontinuities against an acceptance class | ASTM E2375 |
| How much has the bolt stretched? | Pulse-echo time of flight before and after loading | Change in transit time, converted to elongation | ASTM E1685 |
| What preload does that stretch represent? | Calibration of the same bolt type on a load frame | Load per unit change in transit time | E1685 plus a calibration under the project procedure |
Key takeaway: specify which question you need answered. A crack search and a preload measurement need different procedures, reference standards and acceptance criteria.
How does ultrasonic testing find cracks in bolts and anchor rods?
The transducer sits on the flat end of the bolt or rod and sends a longitudinal wave along its length. In a sound bolt the screen shows a back-end echo and, along the threaded length, a regular row of small thread-root echoes. A crack that runs across the bolt, which is the usual fatigue crack orientation, reflects the beam back toward the probe and shows as a larger echo ahead of the back-end echo, often with a drop in the back-end signal.
The Federal Highway Administration describes the method this way:
"Straight-beam testing launches a longitudinal wave normal or nearly normal to the surface of the material being tested."
FHWA's bridge ultrasonic testing page lists pins, trunnion shafts and anchor bolts among the steel bridge components tested this way. It adds that small angles of up to 10° may be used to align the beam with a defect or to improve coverage where access is limited. The physics of a straight beam is covered in our guide to pulse-echo ultrasonic inspection.
Where bolts crack, and how each location is reached
| Location | Why it cracks | How the beam reaches it | Notes |
|---|---|---|---|
| First engaged thread inside the nut | Highest stress concentration under tension and bending | 0° from the threaded end, or from the head end on long bolts | The thread echo pattern changes; compare against a sound bolt of the same size |
| Thread runout (thread-to-shank transition) | Change of section | 0° from either end | Easy to confuse with geometry echoes; a reference bolt helps |
| Under the head (head-to-shank fillet) | Fillet stress concentration | 0° or low angle from the opposite end | From the head end, the fillet is close to the probe and may fall in the near field or dead zone |
| Anchor rod at the top of concrete or grout | Corrosion and bending where moisture collects | 0° from the exposed top end | The concrete line is a known depth, so its location can be marked on the screen |
| Embedded length of an anchor rod | Corrosion, construction damage | 0° from the top end | Beam spread and sidewall echoes limit what can be read on long rods |
Low-angle beams, typically a few degrees off the axis, aim the sound toward a thread root or fillet on one side of the bolt. They also help where a nut, washer or bent rod end prevents a full seat on the end face. Shear-wave angle beams at 45° to 70° are used on welds and plate, not along a bolt axis; see our guide to shear wave ultrasonic testing for that technique.
End-face preparation and the reference standard
A reliable test needs a flat, clean end face square to the axis. Paint, galvanizing runs, a domed or stamped end, or a damaged thread end all scatter the beam. A reference standard, ideally a length of the same bolt or rod with saw cuts or notches at known positions, sets the screen range and the sensitivity. Owner procedures for anchor rods often cite ASTM E114-20, the practice for ultrasonic pulse-echo straight-beam contact testing, as the base method. E114 does not set pass/fail limits, so acceptance must come from the owner or the project specification.
How this differs from bolt hole cracks in rail
EMA's guide to bolt hole cracks in rail joints covers a different defect. A rail bolt hole crack runs through the rail web from a joint bolt hole, and it is found by angle-beam probes scanning the rail, not by testing the bolt. Its remedial action is set by crack length in the 49 CFR 213.113 remedial action table, starting with cracks over 1/2 in. Ultrasonic bolt testing examines the fastener itself: a structural bolt, a stud or an anchor rod.
Worked example: locating a crack in an anchor rod
Inputs: a 1 in. diameter ASTM F1554 anchor rod, 36.0 in. long, with 30.0 in. embedded and the top of concrete 6.0 in. below the top end. The probe is 5 MHz, 0.5 in. diameter, 0° contact. The longitudinal velocity is taken as 5,890 m/s (0.2319 in./µs), the Evident value for 1020 steel; the instrument is calibrated on the reference rod, so the exact velocity sets only the screen scale.
| Step | Calculation | Result |
|---|---|---|
| Back-end echo time | 2 × 36.0 in. ÷ 0.2319 in./µs | 310.5 µs round trip |
| Echo seen at | Sound path read from the calibrated screen | 30.5 in. from the top end (263.1 µs) |
| Depth below top of concrete | 30.5 in. − 6.0 in. | 24.5 in. into the embedded length |
| Near field of the probe | N = D²f / 4v = (12.7 mm)² × 5 MHz ÷ (4 × 5.89 mm/µs) | 34.2 mm (1.35 in.) |
| Beam half-angle at −6 dB | sin(α/2) = 0.514 × v / (f × D) = 0.514 × 5.89 ÷ (5 × 12.7) | 2.73° |
| Beam radius at 36 in. | 36.0 × tan 2.73° | 1.72 in., more than the 0.5 in. rod radius |
What it tells you. The echo sits 24.5 in. into the embedded length, so it is not the concrete-line corrosion site and it cannot be checked visually. Because the beam radius at the far end is larger than the rod, part of the sound strikes the rod wall and mode-converts. That produces trailing echoes after the back-end echo and broadens the signals near the far end. A reflector near the bottom of a long rod therefore needs confirmation by comparing it against a reference rod of the same length, and by testing neighboring rods for the same signal. The near-field and beam-spread formulas are explained in our guides to near field vs far field and ultrasonic testing formulas.
How does ultrasonic bolt testing measure preload?
A tightened bolt stretches, and the stress also slows the sound inside it, so the echo from the far end arrives later. The change in transit time, measured from the same end before and after tightening, is converted to elongation and then to load.
Two effects add together:
- Physical elongation. The stressed length of the bolt grows, so the sound has farther to travel.
- The acoustoelastic effect. Tensile stress along the axis lowers the longitudinal velocity, so the sound travels more slowly over the stressed length.
Because the second effect depends on the material, heat treatment and stressed length, the time change cannot be turned into load from a handbook velocity alone. Each bolt type is calibrated on a load frame. ASTM E1685-20(2026) is the practice for this measurement. Its scope is threaded metal bolts "6.3 mm or more in nominal diameter" with an effective length-to-diameter ratio of 2:1 or greater, and it measures from one end, so the far end doesn't need to be reached. ASTM lists structural steel connections, pressure vessels, turbines and laboratory testing among its uses, and states that the method:
"is used to monitor changes in length of fasteners and as a tool for industrial quality control."
E1685 is written in SI units only. US-unit examples, like the one below, must be converted before they are compared with the standard's text.
Worked example: from time of flight to elongation to load
Inputs: a 7/8-9 UNC structural bolt to ASTM F3125/F3125M Grade A325 (120 ksi minimum tensile strength), 5.0 in. long, in a joint whose effective stressed length is taken as 3.5 in. Longitudinal velocity 0.2319 in./µs; E = 29,000 ksi.
Step 1: the target load. The tensile stress area is As = 0.7854 × (d − 0.9743/n)² = 0.7854 × (0.875 − 0.1083)² = 0.462 in.² The RCSC Specification for Structural Joints sets minimum pretension at 70% of the bolt's specified minimum tensile strength (check the edition in your contract). For this bolt that is 0.70 × 120 ksi × 0.462 in.² = 38.8 kips. Use the value in the pretension table of your contract's RCSC edition.
Step 2: calibrate the bolt type. Bolts from the same lot, in a fixture that reproduces the joint's grip, are loaded on a calibrated load frame:
| Load (kips) | 0 | 10 | 20 | 30 | 40 |
|---|---|---|---|---|---|
| Measured change in transit time (ns) | 0 | 77 | 154 | 231 | 308 |
The fit is linear at 7.70 ns per kip (308 ns ÷ 40 kips). These calibration readings are illustrative inputs for the example, not published constants; your own calibration will differ.
Step 3: measure the installed bolt.
| Reading | Value |
|---|---|
| Before tightening (round trip, 5.0 in.): 2 × 5.0 ÷ 0.2319 | 43.122 µs |
| After tightening | 43.422 µs |
| Change in transit time, Δt | 0.300 µs = 300 ns |
| Load: 300 ns ÷ 7.70 ns per kip | 38.96 kips |
38.96 kips is above the 38.8 kip minimum, so the bolt passes.
Step 4: check the elongation. At 38.96 kips the stress is 38.96 ÷ 0.462 = 84.3 ksi (70.3% of 120 ksi), the strain is 84.3 ÷ 29,000 = 0.002908, and the elongation over 3.5 in. is 0.0102 in. That stretch alone would delay the echo by 2 × 0.0102 ÷ 0.2319 = 88 ns. The other 212 ns of the measured 300 ns comes from the acoustoelastic velocity drop.
Why the calibration matters. Converting Δt directly to length with the unloaded velocity (0.300 × 0.2319 ÷ 2 = 0.0348 in.) would overstate the real elongation by a factor of 3.4 in this example. Instruments built for E1685 work handle this with a material stress factor, which is why a bolt type needs its own calibration and why a calibration made for one grip length doesn't carry over to another.
What is the step-by-step procedure for ultrasonic preload measurement?
A preload procedure records each bolt's unloaded reading, applies a calibration made on the same bolt type, and remeasures at the same temperature and probe position. The sequence below completes the laboratory procedure into a field-ready one. Your written procedure, approved by the responsible Level III, governs.
- Define the bolts. Record the specification and grade, diameter and thread, length, lot or heat number, coating, and the joint grip length. Calibrations apply only to this combination.
- Prepare and identify the ends. Face or grind both ends flat and square where the procedure requires a reflecting back end. Mark each bolt with a unique number so its unloaded reading can be matched later.
- Stabilize the temperature. Let bolts reach a steady temperature and measure it on the bolt. Temperature changes both the bolt length and the sound velocity, so readings are taken at, or corrected to, the same temperature.
- Set up the instrument. Use the transducer, frequency and couplant named in the procedure. Verify the zero and range on a reference bar of known length before the first reading and at intervals the procedure sets.
- Take the unloaded reading. Seat the probe in the same marked position every time, use a consistent couplant layer, and record the transit time (or ultrasonic length) for each bolt.
- Calibrate on a load frame. Load sample bolts from the same lot in a fixture that reproduces the grip, at several load steps up to at least the target load, and record Δt at each step. Unload and repeat to check repeatability.
- Fit and check the calibration. Fit the load-versus-Δt line, check linearity across the working range, and record the residuals. Don't use the line above the highest calibrated load or in the plastic range.
- Install and tighten. Tighten by the project's installation method. Ultrasonic measurement verifies the result; it doesn't replace the specified pretensioning method unless the specification allows it.
- Take the loaded reading. Re-seat the probe at the marked position, measure the bolt temperature again, and record the transit time.
- Compute and compare. Apply the temperature correction, compute Δt, convert to load with the calibration line, and compare it with the target pretension.
- Verify a sample. Re-measure a set of bolts, ideally by a second operator, to confirm the repeatability the procedure claims.
- Report. List each bolt ID, both readings, temperatures, calibration reference, computed load, pass or fail, and any bolt that couldn't be read.
Where unloaded readings weren't taken before installation, the conventional method can't give an absolute load. That is the problem the reference-free methods described below try to solve.
Which standards apply to ultrasonic testing of bolts?
No single standard covers every ultrasonic bolt test; choose by what you need to know. The table lists the current designations as shown on each publisher's page in October 2026.
| Standard | Title (short) | What it does for bolt UT | What it doesn't do |
|---|---|---|---|
| ASTM E1685-20(2026) | Measuring the change in length of bolts using the ultrasonic pulse-echo technique | The bolt elongation (preload) practice; ≥6.3 mm diameter, L/D ≥ 2:1 | Doesn't find cracks or set preload targets |
| ASTM E114-20 | Ultrasonic pulse-echo straight-beam contact testing | Base practice for 0° end-face flaw detection | No acceptance criteria |
| ASTM E2375-26a | Ultrasonic testing of wrought products | Bar stock and forgings 0.250 in. (6.35 mm) and thicker, with acceptance classes the purchaser selects | Not for installed bolts |
| ASTM F606/F606M-26a | Mechanical properties of threaded fasteners, washers, DTIs and rivets | Proof load and tension tests used to validate a calibration | Not an ultrasonic standard |
| ASTM F3125/F3125M-26 | High-strength structural bolts and assemblies | Defines A325, A490 and related grades | Contains no NDT requirement |
| ASTM F1554-20 | Anchor bolts, steel, 36, 55 and 105 ksi yield | Defines anchor rod grades and forms | Contains no NDT requirement |
| ISO 16810:2024 | Ultrasonic testing, general principles | General UT framework | Excludes acceptance criteria |
The original version of this article cited "ASTM E2375-26". The active edition is E2375-26a. E1685-20(2026) is a 2026 reapproval of the 2020 edition.
Acceptance criteria come from the job, not from these practices. For a crack search, the owner's procedure or engineer of record sets what is reportable and what is rejectable. For preload, the target comes from the design or the RCSC Specification. For weld acceptance on the same structure, see our guide to ultrasonic testing acceptance criteria standards.
What advanced methods are in development?
Research methods aim at three weaknesses of the conventional method: the need for an unloaded reading, stress that isn't uniform along the bolt, and couplant variation. These are laboratory results reported by their authors under their own test conditions, not field-proven accuracy.
| Method | What it solves | Named research |
|---|---|---|
| Longitudinal plus transverse (shear) wave, "two-mode" | Estimates preload without an unloaded reference reading, because the two wave modes respond differently to stress | Niwinski, Kraemer, Klein, Oechsner, Herter and Becker, Materialwissenschaft und Werkstofftechnik 54(8):945–956, 2023 (doi:10.1002/mawe.202100355) |
| Shape factor | Accounts for non-uniform stress through head, shank, threads and nut | Pan, Pan, Chang and Xu, NDT & E International 111:102210, 2020 (doi:10.1016/j.ndteint.2019.102210); tested on M16 and M20 bolts, with relative error reported within 5% |
| Coupling-error compensation (solid coupling) | Reduces error from couplant thickness, pressure and creep | Yuan, Sun, Wang, Zhao, Mu and Sun, Measurement 221:113484, 2023 (doi:10.1016/j.measurement.2023.113484) |
| Thin-film piezoelectric "smart bolts" | A sensor built onto the bolt head for repeated or continuous monitoring | Liu et al., Sensors 23(1):220, 2023, on nickel-based superalloy bolts (doi:10.3390/s23010220) |
| Laser ultrasound | Non-contact measurement with no couplant | Kitazawa, Lee and Patel, NDT & E International, 2023 (doi:10.1016/j.ndteint.2023.102838) |
| EMAT (electromagnetic acoustic transducer) | Generates sound without liquid couplant and can excite shear modes | Active research area; performance depends on the bolt's conductivity and magnetic properties |
For a survey of the principles, effective stressed length, temperature compensation and calibration methods, see the review by Pan, Pan, Shao, Chang and Xu, Chinese Journal of Mechanical Engineering 33:11, 2020 (doi:10.1186/s10033-020-0431-x).
The earlier version of this article quoted a 3.35% detection error for the solid-coupling work. That figure couldn't be confirmed from the paper's public record and has been removed.
What ultrasonic bolt testing cannot do
Ultrasonic bolt testing is limited by access to a flat end face, by geometry echoes from threads, and by calibration. An engineer specifying it should state these limits in the scope.
| Limit | Why | What to do |
|---|---|---|
| No usable end face | Domed, stamped, painted or damaged ends scatter the beam | Specify end preparation, or test from the other end |
| Near-surface zone under the probe | The near field and the dead zone hide reflectors close to the probe | Test critical areas such as the head fillet from the opposite end; see dead zone in ultrasonic testing |
| Thread echoes | Every thread root reflects some sound, which can hide or mimic a small crack | Compare against a sound reference bolt of the same size and length |
| Long, slender rods | Beam spread hits the rod wall and produces mode-converted trailing echoes | Confirm far-end reflectors against a reference rod; check neighbors |
| Cracks parallel to the axis | A lengthwise crack reflects little sound back up a 0° beam | Use magnetic particle testing on exposed surfaces; see our MT guide |
| Absolute preload without an unloaded reading | Conventional time-of-flight measures change, not absolute load | Take unloaded readings before installation, or use a validated reference-free method |
| Calibration transfer | The stress factor depends on material, heat treatment and grip length | Calibrate each bolt lot and grip combination |
| Temperature | Length and velocity both change with temperature | Measure the bolt temperature at each reading and correct |
| Couplant | Couplant thickness and pressure change the reading | Use the same couplant and seating method each time; see ultrasonic couplant |
UT also doesn't size a crack's depth precisely from a single 0° echo. A rising echo amplitude over repeat inspections, or a loss of back-end signal, is a stronger indication than one reading.
How EMA Industries applies this
EMA Industries provides AWS D1.1 structural weld UT, MT and PT, and NDT Level III services, from Southern California, mobilizing nationwide. On structural and industrial projects, a Level III can write or review the written procedure for anchor rod and bolt examination, including the reference standard, end-face preparation and reporting format, before field work starts. To scope a bolt or anchor rod examination, see our ultrasonic examination services or contact EMA Industries.
Related reading: AWS D1.1 weld inspection criteria · SNT-TC-1A requirements for NDT certification · Ultrasonic testing equipment · Phased array ultrasonic testing
Frequently asked questions
What is the ASTM standard for ultrasonic testing of bolts?
It depends on the purpose. ASTM E1685-20(2026) covers measuring the change in length (elongation) of bolts by ultrasonic pulse-echo, which is how preload is measured. For crack detection from the bolt end, owner procedures usually build on ASTM E114-20, the straight-beam contact practice. Neither sets acceptance limits; those come from the project specification.
Can ultrasonic testing find cracks in anchor bolts that are cast into concrete?
Yes, provided the exposed top end can be prepared flat. A 0° beam from the top travels down the embedded length and reflects from cracks that run across the rod, such as corrosion-fatigue cracks near the top of concrete. On long rods, beam spread produces extra echoes near the far end, so far-end signals should be checked against a reference rod.
How accurate is ultrasonic preload measurement?
Accuracy depends on calibration more than on the instrument. A calibration on the same bolt lot and grip length, with temperature correction and a consistent couplant, gives a repeatable result. Research methods such as the shape factor approach report errors within 5% in the laboratory, but that figure applies only to the bolts and conditions tested.
Why does a bolt need to be measured before it is tightened?
Conventional time-of-flight measurement reads a change in transit time, not an absolute load. Without the unloaded reading there is no starting point to subtract. Two-mode methods that combine longitudinal and transverse waves are being developed to remove this need.
What NDT methods are used on bolts besides ultrasonic testing?
Visual testing comes first, for corrosion, thread damage and bent rods. Magnetic particle testing finds surface-breaking cracks on exposed ferromagnetic bolt surfaces, and penetrant testing does the same on non-magnetic materials. Ultrasonic testing is the only one of these that reaches the hidden threads inside a nut or the embedded length of an anchor rod.
Sources
- ASTM E1685-20(2026), Standard Practice for Measuring the Change in Length of Bolts Using the Ultrasonic Pulse-Echo Technique, ASTM store page (scope and significance)
- ASTM E114-20, Standard Practice for Ultrasonic Pulse-Echo Straight-Beam Contact Testing, ASTM
- ASTM E2375-26a, Standard Practice for Ultrasonic Testing of Wrought Products, ASTM
- ASTM F606/F606M-26a, Mechanical Properties of Threaded Fasteners, Washers, DTIs and Rivets, ASTM
- ASTM F3125/F3125M-26, High Strength Structural Bolts and Assemblies, ASTM
- ASTM F1554-20, Anchor Bolts, Steel, 36, 55, and 105-ksi Yield Strength, ASTM
- Bridge – Ultrasonic Testing (Flaw Detection), FHWA InfoTechnology
- Moore, Phares and Washer, Guidelines for Ultrasonic Inspection of Hanger Pins, FHWA-HRT-04-042 (2004), FHWA via ROSA P
- 49 CFR 213.113, Defective rails, eCFR (rail bolt hole crack remedial actions)
- Evident, Material sound velocities
- Pan Q., Pan R., Shao C., Chang M., Xu X., "Research Review of Principles and Methods for Ultrasonic Measurement of Axial Stress in Bolts," Chinese Journal of Mechanical Engineering 33:11 (2020), doi:10.1186/s10033-020-0431-x
- Pan Q., Pan R., Chang M., Xu X., "A shape factor based ultrasonic measurement method for determination of bolt preload," NDT & E International 111:102210 (2020), doi:10.1016/j.ndteint.2019.102210
- Niwinski T. J. et al., "Ultrasonic in situ evaluation of the bolt preload without referencing an unloaded state using a combination of longitudinal and transverse waves," Materialwissenschaft und Werkstofftechnik 54(8):945–956 (2023), doi:10.1002/mawe.202100355
- Yuan B. et al., "Study on bolt preload measurement: An error compensation model for ultrasonic detection based on solid coupling," Measurement 221:113484 (2023), doi:10.1016/j.measurement.2023.113484
- Liu et al., "Ultrasonic Measurement of Axial Preload in High-Frequency Nickel-Based Superalloy Smart Bolt," Sensors 23(1):220 (2023), doi:10.3390/s23010220
- Kitazawa, Lee and Patel, "Noncontact measurement of bolt axial force in tightening processes using scattered laser ultrasonic waves," NDT & E International (2023), doi:10.1016/j.ndteint.2023.102838
- RCSC, Specification for Structural Joints Using High-Strength Bolts (2020): minimum pretension at 70% of specified minimum tensile strength (secondary sources; check the edition in your contract)