How to Calibrate a UT Flaw Detector (ASME V Article 4)
Your shift starts at 06:00 and the inspection foreman wants the first weld scanned by 07:00. Before you touch the part, the flaw detector has to be calibrated end-to-end against an IIW V1 block on the same probe, cable, and couplant you will use all day. Skip that and every indication you log is unverifiable. This walk-through is the exact sequence ASME V Article 4 §T-462 and ASTM E317 require, plus the gain and reference settings that survive a Level III audit.
For: Level II UT technicians performing manual contact pulse-echo weld inspection per ASME V Article 4, and Level III auditors verifying procedure compliance.
Equipment and reference standards required
Pull the flaw detector, probe, cable, and a calibration block that matches the material you will scan. For carbon-steel weld work, that is an IIW Type 1 (V1) block per ISO 2400. For curved or thin-wall pipe, swap to a V2 block (ISO 7963) or a procedure-specific basic calibration block (BCB) per ASME V Art. 4 §T-434.[1][2]
The probe must carry a current calibration certificate (typically 6 months for piezo, 12 months for the instrument body). The instrument itself needs a documented annual linearity check per ASTM E317 covering vertical, horizontal, and amplitude linearity within ±5%, ±2%, and ±20% respectively.[3]
Couplant must match the procedure — typically glycerin or a propylene-glycol gel for carbon steel between 5°C and 50°C. Avoid switching couplants mid-calibration; transfer correction values shift up to 6 dB between water and gel.
Step 1: Velocity and zero on a 0° straight beam
Couple a 2.25 or 5 MHz straight-beam probe to the 25 mm step of the V1 block. Set range to 100 mm, gain to 40 dB, and pulse-repetition frequency to auto. Adjust velocity until the first back-wall echo lands at 25 mm and the second multiple at 50 mm on the screen.
Now adjust the zero (delay) so the front-surface reflection sits at exactly 0 mm. Re-check both back-wall multiples — they should fall at 25, 50, 75, and 100 mm within ±0.5 mm. If they drift, repeat zero, then velocity. For carbon steel, velocity should land at 5920 ± 30 m/s; deviations beyond that point to a coupling problem or a worn wear-plate.[4]
Document the screen photo or A-scan capture. ASME V Art. 4 §T-462.2 requires the calibration record to show the actual A-scan, not just typed values.
Step 2: Angle beam refraction angle and exit point
Switch to your angle-beam probe — typically 45°, 60°, or 70° for weld work. Place the probe on the 100 mm radius of the V1 block and slide forward/back until the radius reflector peaks. The center of the probe at peak amplitude is the exit point; mark it on the wedge with a fine permanent marker. Acceptable drift from the engraved exit point is ±1 mm.
Next, measure refraction angle on the V1 5 mm side-drilled hole (SDH) or the 50 mm circle. Slide the probe until the SDH echo peaks, then read the angle off the engraved scale. Acceptable variance is ±2° from the nominal probe angle per ASME V Art. 4 §T-464.1.[1]
If angle drifts beyond ±2°, the wedge is worn — either replace it or recalculate skip distance and beam path using the measured angle, not the nominal. Log the measured value on the calibration sheet.
Step 3: DAC or TCG construction with reference reflectors
For weld inspection per ASME V Art. 4, build a DAC (Distance-Amplitude Correction) curve on a basic calibration block with side-drilled holes at 1/4T, 1/2T, and 3/4T. Set the gain so the closest reflector peaks at 80% full-screen height (FSH). Mark that point. Move to each subsequent SDH without changing gain, peak the signal, and mark amplitude vs sound-path.
Draw the DAC through the three points. The primary reference level is the gain at which the closest SDH reads 80% FSH — log this value. Typical primary reference for a 2.25 MHz 45° probe on 25 mm CS plate sits around 38-44 dB, depending on couplant and surface condition.[5]
Add scanning gain per the procedure — usually +6 dB for code work, +14 dB if the procedure follows AWS D1.1 Section 6.[6] Record the final scan-level gain. Re-check the DAC at end of shift and after every break longer than 30 minutes.
DAC reflector positions on a basic cal block (ASME V Art. 4 Fig. T-434.2.1)
- 1/4T side-drilled hole — closest reference, sets primary gain
- 1/2T side-drilled hole — mid-range reference
- 3/4T side-drilled hole — far reference, validates beam spread
- Notch on opposite face — optional, used for surface-breaking validation
Step 4: System linearity verification (daily and weekly)
Daily check: place the probe so the back-wall echo from a 25 mm step sits at 80% FSH. Drop gain by 6 dB — the echo must fall to 40% FSH ±2%. Add 6 dB back — return to 80% FSH ±2%. Drop by 12 dB — should sit at 20% FSH ±2%. Failure here means the receiver is non-linear and the unit goes out of service until repaired.[3]
Weekly: run the full ASTM E317 linearity protocol — vertical (amplitude) linearity at 20%, 40%, 60%, 80% FSH steps, horizontal (sweep) linearity using V1 multiples, and amplitude control linearity at -6, -12, -18, -24 dB steps. Document on the linearity log; retain for 5 years per most owner specs.
A common failure mode: the receiver attenuator clicks but the screen amplitude does not drop proportionally. This usually means a stuck attenuator relay — flag the unit, do not finish the shift on it.
Step 5: Recalibration triggers during the shift
Recalibrate any time the probe is changed, the cable is changed, the couplant is changed, the power is cycled, or 4 hours have elapsed — whichever comes first. ASME V Art. 4 §T-467 lists these triggers explicitly.[1]
Also recalibrate immediately if a reference reflector drifts more than 20% in amplitude or more than 10% in sound-path position. If a reflector drifts beyond those thresholds during the shift, every weld scanned since the last good calibration check must be re-scanned. There is no shortcut around this — record it on the calibration sheet and notify the Level III.
Common mistakes
Recurring errors that lead to failed inspections:
- Calibrating with one couplant and scanning with another — transfer correction can be 6 dB or more between water and glycerin gel; the DAC becomes meaningless. Always use the same couplant for cal and scan.
- Marking the exit point with a thick Sharpie and trusting it for the rest of the shift. The wedge wears under load; re-check exit point after every break and re-mark with a fine point.
- Skipping the linearity drop test "because the unit is new." A factory-fresh unit can fail vertical linearity if the receiver board took a shock in shipping. Always run the daily check before the first scan.
- Setting primary reference gain on a corroded or pitted SDH. If the reflector is degraded, the DAC sits too high and real indications get missed. Inspect the cal block face under raking light before every shift.
Frequently Asked Questions
How often must a UT flaw detector be linearity-checked under ASME Section V?
ASME V Article 4 §T-461 and §T-462 require system linearity verification at the start of each examination, when any probe or cable is changed, every 4 hours of continuous use, and at the end of each shift. The annual full-protocol ASTM E317 linearity check (vertical, horizontal, amplitude control) is separate — that one goes in the instrument calibration file and gets stickered onto the unit. Most procedures also require an extra mid-shift check after a power cycle or any drop, even if no damage is visible.
What is the difference between primary reference gain and scanning gain?
Primary reference gain is the dB value at which the closest reference reflector (typically the 1/4T side-drilled hole) reaches 80% full-screen height. It is the anchor for the DAC curve and the value reported on the calibration sheet. Scanning gain is the working gain you actually use during inspection — typically the primary reference plus 6 to 14 dB, depending on whether you are following ASME V (Art. 4 specifies +6 dB for transfer plus procedure-defined scanning sensitivity) or AWS D1.1 (often +14 dB). The scanning gain compensates for surface condition, beam spread, and attenuation between the cal block and the actual weld.
Why must I recalibrate after a couplant change?
Different couplants have different acoustic impedances and wet-out behavior. Switching from water to glycerin can change transfer signal amplitude by 4 to 8 dB on the same surface; switching from glycerin to a propylene-glycol gel can move it another 2 to 3 dB. Because the DAC curve is anchored to a specific signal amplitude on the cal block, any couplant change invalidates the reference. The procedure requires a fresh velocity, zero, and DAC build with the production couplant — short-cutting this is one of the top three audit findings from third-party Level IIIs on ASME pressure-vessel work.
Can I use the same calibration for a 25 mm and a 50 mm thick weld?
Only if the procedure-required calibration block covers both thicknesses with reference reflectors. ASME V Art. 4 §T-434.2.1 requires the basic calibration block to be within ±25% of the production part thickness, or to bracket the production thickness with reference reflectors at 1/4T, 1/2T, and 3/4T of the thinner part. For a 25 mm and a 50 mm scan, you either use a 38 mm cal block (within ±25% of both) or you build two separate calibrations. Mixing thicknesses on one cal sheet is an audit fail.
How do I document a failed linearity check?
Log the date, time, instrument serial number, probe ID, the failing parameter (e.g., "vertical linearity at 40% FSH read 36%, outside ±2% tolerance"), and the corrective action (typically: unit removed from service, sent for repair, replacement unit re-calibrated). Every weld scanned since the previous valid linearity check must be re-scanned with a verified instrument. Keep the failed-check record for 5 years minimum — many owner specs (Shell DEP, Saudi Aramco SAES-W-012) require it for the life of the asset.
References & Standards Cited
- ASME BPVC Section V, Article 4, 2023 edition — Ultrasonic Examination Methods for Welds, §T-461 through §T-467
- ISO 2400:2012 — Non-destructive testing — Ultrasonic testing — Specification for calibration block No. 1
- ASTM E317-21 — Standard Practice for Evaluating Performance Characteristics of Ultrasonic Pulse-Echo Testing Instruments and Systems
- ASTM E114-15 — Standard Practice for Ultrasonic Pulse-Echo Straight-Beam Contact Testing
- ASNT SNT-TC-1A:2020 — Recommended Practice for Personnel Qualification and Certification in Nondestructive Testing
- AWS D1.1/D1.1M:2020 — Structural Welding Code — Steel, Clause 8 (UT of Statically Loaded Connections)
Related on NDT Connect
Founder of NDT Connect and Atlantis NDT. 15+ years in industrial inspection across oil & gas, petrochemical, and offshore. ASNT Level III certified across five methods. Drives platform standards for the NDT Connect marketplace.