14.1 Calibration Blocks, DAC, TCG, and Sensitivity Setting

The ASNT NDT Level II ultrasonic general exam still lists Evaluation of Base Material Product Forms and Evaluation of Weldments as official UT topics 2 and 3 on the outline administered prior to 5 February 2027. Those headings do not begin with a picture of a slag line. They begin with a legal instrument: range that means metal path, zero that means the test surface, velocity that matches the alloy, and sensitivity that is traceable to a hole or a notch. A Level II who can name every product-form discontinuity but cannot say what the IIW 100 mm radius, a DSC 1 in radius, or a 1.5 mm side-drilled hole is doing on the screen will miss the calibration items that sit in front of every evaluation item.

Calibration is not a warm-up. It is the measurement system. Every later call — lamination versus lost backwall, root geometry versus lack of fusion, an indication 6 dB over DAC — inherits the last calibration. This section teaches the blocks the general exam expects you to recognize, the four instrument settings that make a screen reading true, how a distance-amplitude correction (DAC) and time-corrected gain (TCG) are built, how decibel arithmetic works, what transfer correction is conceptually, and which field changes make the calibration invalid.

ASME Boiler and Pressure Vessel Code, Section V, Article 4 (welds) and Article 5 (materials), AWS D1.1, ISO 2400, and ISO 7963 are industry technique references for blocks and sensitivity. They are not ASNT exam publications. Attribute the number to the standard. The employer's written procedure states which block, which reflector, which reference amplitude, and which verification interval apply on that job.

Why the screen is a clock until you calibrate it

An ultrasonic instrument plots time on the horizontal axis and amplitude on the vertical axis. Time becomes metal path only after you assign a velocity and a zero. Amplitude becomes a rejectable fraction of a reference only after you assign a gain to a stated reflector. Until those assignments exist, a peak at division 4 is just a peak at division 4.

Four settings turn the clock into a measuring tool:

1. Screen range (time base) — how much metal path occupies the full horizontal display.
2. Delay / zero — how much of that time is wedge, delay line, or cable, so that screen zero is the entry surface of the test material.
3. Velocity — the longitudinal or shear speed you claim for this alloy and this mode.
4. Sensitivity — the gain that puts a named reference hole or notch at a named screen height (often 80% full-screen height, FSH).

Angle-beam work adds two mechanical facts you measure on the same block: the probe index (exit point) and the actual refracted angle. Those are not gain knobs, but without them skip distance and depth plots are fiction.

The blocks the general exam expects you to tell apart

Several steel blocks appear in Level II teaching. They overlap in purpose and differ in which feature you use. Do not treat them as one generic "cal block."

IIW Type 1 (V1)

The classic International Institute of Welding (IIW) block — often called Type 1 or V1 — is the large shop block. A common size is 300 mm long × 100 mm high × 25 mm thick. The features that matter on the exam are the 100 mm radius quadrant, the 50 mm radius, a large 50 mm diameter hole, a small 1.5 mm diameter hole, and resolution notches or slots depending on the edition.

What you actually do with it:

• Peak the 100 mm or 50 mm radius to mark the probe index on the wedge. The radius center is a known point; the mark on the wedge is where the beam leaves the plastic.
• Use the engraved degree scale and the 1.5 mm hole (or the radius-to-hole geometry) to measure the true refracted angle. Wedges are stamped 45°, 60°, or 70°; wear and temperature move the real angle.
• Set screen range from a known radius or from multiple back reflections on the 25 mm thickness.
• Set delay/zero so that known radius or hole metal path lands on the correct screen division.
• Use the 1.5 mm hole or a stated notch as a first sensitivity reference when the procedure names that reflector.

The IIW Type 1 is bulky. It is the right block when you have bench space and you need index, angle, range, and a resolution check on one piece of steel.

IIW Type 2

IIW Type 2 is the smaller, often 12.5 mm (0.5 in) thick United States field cousin. It keeps a 50 mm (2 in) radius and a working set of holes or notches, but it is light enough to carry up a column. Functions are the same family — range, delay, index, angle, a sensitivity hole — on a shorter sound path. A Type 2 does not automatically replace a Type 1 if the procedure names the Type 1 1.5 mm hole or the 100 mm radius. Read the procedure's block clause.

DSC (distance/sensitivity calibration) block

The AWS DSC block is the compact distance and sensitivity block used heavily in AWS D1.1 structural work. The features the exam cares about are the 1 in radius and the 3 in radius. You set range and delay from those known curved paths, then set a reference level on the radius or on the hole the procedure names. The DSC exists because a full IIW Type 1 is a poor companion on a girder flange. It is not a substitute for an ASME basic calibration block with a ladder of side-drilled holes when the procedure demands a DAC from 1/4 T, 1/2 T, and 3/4 T holes.

IIW-ISO blocks (ISO 2400 and ISO 7963)

ISO 2400 specifies Calibration Block No. 1, the ISO counterpart of the large IIW/V1 block. ISO 7963 specifies Calibration Block No. 2, the miniature V2 block with short radii (commonly 25 mm and 50 mm) for field range and sensitivity. When a procedure or a multinational traveler says "IIW-ISO," it is pointing at these dimensional standards, not at a different physics. Hole diameters, radii, and thickness are as specified in the ISO edition the procedure cites. Do not invent a hole size. Do not assume an ISO Block No. 2 hole is the same reflector as an ASME 3/64 in side-drilled hole.

ASME basic calibration block (why it still belongs in this section)

Welds examined to ASME Section V, Article 4 are commonly calibrated on a basic calibration block of similar material (P-number grouping) that contains side-drilled holes (SDHs) at 1/4 T, 1/2 T, and 3/4 T, plus any notches the procedure adds. That block is how you build a DAC. The IIW/DSC family sets the instrument clock and a first reference; the ASME SDH ladder sets the distance-amplitude relationship for the examination volume. A Level II who confuses "I peaked the IIW radius" with "I have a DAC" will write an illegal weld report.

Screen range (time base)

Screen range is how many inches or millimeters of metal path you have assigned to the full width of the display. It is a time-base setting, not a gain setting. If you set 10 in of shear-wave metal path full screen, a reflector at 5 in metal path must sit at mid-screen after delay and velocity are right.

How you set it, in practice:

• Straight beam: multiple back reflections from a known thickness (the IIW 25 mm path, a step wedge, or the part itself if the procedure allows a known sound wall).
• Angle beam: a known radius (IIW 100 mm or 50 mm, DSC 1 in or 3 in, V2 25 mm or 50 mm) or a known SDH metal path.

Choose a range that puts the examination volume — first-leg root, second-leg cap, or the far-side skip the procedure requires — on scale with room to see the next node. A 5 in range that clips a 70° second-leg cap on 1 in plate is an illegal display, not a tidy screen. A 20 in range that crushes the root into the first two divisions makes plotting sloppy. The procedure's scan plan states the legs you must see; the range must include them.

Range errors hide as wrong location. The indication is real. It is just not where the screen says it is. That is how a root reflector gets plotted in the cap, and how a mid-wall lamination gets reported as a backwall.

Delay and zero

Delay (also called zero offset or probe zero) subtracts the time the pulse spends in the wedge, delay line, or wear face so that screen zero is the test-surface entry point. Until delay is set, every metal path is long by the plastic path.

Angle-beam delay is not optional. A 70° polystyrene or acrylic wedge can hold a microsecond-scale path. At steel shear velocity that is a large fraction of an inch. If you skip delay and then plot depth from metal path and angle, every reflector is systematically deep.

Straight-beam delay is the same idea for a delay-line or dual-element probe: the delay shoe or the roof-angle wear face is not steel. Thickness gages hide this in a zero or probe-zero routine on a reference step. Contact straight-beam work on a thin plate still needs a zero if the instrument includes cable and wear-face time in the clock.

Set delay from a known metal path, not from hope:

1. Peak a known radius or a known-thickness backwall.
2. Adjust delay until that echo sits on the screen division that matches its true metal path at the velocity you have set.
3. Confirm with a second known path (the other radius, a second back reflection, or another SDH). One echo can be forced onto a line with a wrong velocity plus a compensating delay. Two echoes expose the lie.

If delay is right and velocity is wrong, paths scale. If velocity is right and delay is wrong, every path is offset by a constant. The exam likes that distinction.

Velocity

Velocity converts time to distance. You must use the mode you are running:

• Carbon-steel longitudinal (straight-beam, 0°) is on the order of 0.233 in/µs (about 5.9 km/s).
• Carbon-steel shear (angle-beam after refraction) is on the order of 0.128 in/µs (about 3.2 km/s).

Those numbers are starting values, not laws of nature. Stainless, duplex, nickel alloys, castings, and cladding run differently. Grain-flow direction in heavy plate and forgings can shift velocity a few percent. A few percent on a long 70° path is a large location error.

Set or verify velocity on a known thickness of the same product form and alloy family, or use the value the procedure locks after a documented check. Do not leave a shear-wave examination on the instrument's longitudinal default. That single mistake doubles metal path and makes every skip plot impossible.

Wrong velocity is one of the cleanest ways a calibration becomes invalid even though the holes still light up. The screen looks lively. The ruler is in the wrong units.

Sensitivity: reference hole or notch

Sensitivity is the gain that makes a named reflector reach a named screen height. Common teaching references:

• The 1.5 mm IIW hole or an ISO Block hole the procedure names.
• An ASME side-drilled hole at a stated depth (often the hole that produces the highest response among the DAC holes, then the others are plotted relative to it).
• A surface notch of stated depth and length (root or ID notch for piping, OD notch when the procedure says so).
• An AWS reference level on a DSC radius or on the hole D1.1 names for that thickness.

A typical shop phrase is "set the 1.2 mm (3/64 in) SDH to 80% FSH and build DAC." The number that matters is whatever the procedure wrote. Sensitivity is not "turn it up until the grass looks healthy." Grass is noise. Noise is not a reference.

Two sensitivity ideas the exam separates:

• Reference level — the gain at which the calibration reflector is at the stated height. All later dB statements ("6 dB over reference," "14 dB over DAC") are relative to this level, after any scanning gain the procedure allows.
• Scanning gain / evaluation gain — extra gain used so you do not walk past a weak reflector, then returned to reference (or to the DAC/TCG presentation) for measurement. Adding scanning gain without recording it, then evaluating at that hotter gain, invents rejectable indications.

Notches and holes are not interchangeable. A side-drilled hole is a cylindrical reflector used for volume sensitivity and DAC. A notch is a surface-breaking rectangular reflector used when the examination is aimed at ID/OD connected discontinuities (piping, shaft, corrosion). A flat-bottom hole (FBH) is a different amplitude standard (often immersion or forging work). If the procedure says SDH, do not "close enough" a notch.

DAC from side-drilled holes

A distance-amplitude correction (DAC) is a curve on the screen that shows how amplitude from equal-size reflectors falls with metal path because of beam spreading, attenuation, and the changing intersection of the beam with the hole.

How you build it:

1. Use the ASME basic calibration block (or the procedure's equivalent) with same-diameter SDHs at increasing metal paths — classically 1/4 T, 1/2 T, 3/4 T, and additional holes when thickness or the scan plan needs them.
2. Peak each hole from the same surface and the same angle you will scan with.
3. Set the highest of those peaked signals (or the hole the procedure names first) to the reference height, often 80% FSH.
4. Mark each hole's peaked amplitude on the screen. Connect the marks. That broken line is the DAC.
5. Record reference gain. From this point, an indication's amplitude is spoken as dB above or below DAC, not as a raw percent of screen that ignores distance.

Why SDHs, not the IIW 1.5 mm hole alone? One hole at one path gives one sensitivity point. A weld examination volume spans a range of metal paths (first-leg root is a short path; second-leg cap is a long path). Without a DAC, a deep slag line looks "small" only because it is far, and a shallow geometric nick looks "large" only because it is near.

DAC rules of honesty:

• Holes must be the same diameter. Mixing a 1.5 mm hole with a 3/64 in hole is not a DAC.
• The block material must be acoustically comparable to the part (Article 4's P-number grouping idea). A carbon-steel DAC on a coarse stainless weld is a different examination.
• Rebuild the DAC when you change angle, frequency, probe size, wedge, or instrument settings that affect the beam.
• A DAC is not a transfer correction. It equalizes distance on the block. The part may still be rougher, more attenuative, or curved.

TCG

Time-corrected gain (TCG) — also called swept gain or distance-amplitude compensation in some menus — is the electronic version of a DAC. Instead of drawing a falling curve and comparing indications to it, the instrument raises gain as time increases so that the same-size SDHs all peak at the same screen height. A 1/4 T hole and a 3/4 T hole both sit at 80% FSH. Evaluation then uses a horizontal reference line.

TCG is not a different physics. It is a different presentation of the same SDH ladder:

• You still peak the same holes.
• You still need the same comparable block.
• You still record a reference.
• You still invalidate it when the beam or the material assumption changes.

What changes is the arithmetic on the screen. With DAC, a signal 6 dB over the curve at that metal path is over the reference. With TCG, a signal 6 dB over the flat reference line is over the reference. Do not mix the languages. A Level II who applies a DAC dB offset on a TCG screen, or who turns TCG on after setting a single-hole sensitivity and calls it "calibrated for distance," has not completed the calibration.

Some digital instruments can store both. The procedure says which presentation is legal for recording. If the traveler requires a DAC screenshot, a TCG-only file is incomplete.

Decibel arithmetic the exam will actually run

Amplitude ratios in ultrasonics are reported in decibels:

dB = 20 log₁₀ (A₁ / A₂)

where A is screen height (or a linear amplifier voltage), not energy and not "how bright it looked."

Memorize the handful the general exam uses as furniture:

Signs matter. +6 dB of gain doubles the height of a given echo (or brings a half-height echo back to the original height). −6 dB (or a drop from 80% to 40% at the same gain) is a halving. The 6 dB drop method of length sizing uses that half-amplitude point as the reflector's apparent end — a sizing method, not a calibration method. Do not confuse "I dropped 6 dB to size" with "my calibration is 6 dB off."

Two working examples:

• Reference hole is 80% FSH at 46 dB. The same hole later reads 40% FSH at 46 dB. Sensitivity has fallen 6 dB. That is a calibration failure, not a new DAC point.
• An indication peaks at 80% FSH when you have added 12 dB of scanning gain above the 80% DAC reference. At reference gain it would be about 20% FSH, which is 12 dB under a DAC that was drawn at 80%. Whether that is recordable depends on the procedure's recording level, not on the bright screen you used to find it.

If an item gives two percentages and asks for dB, use 20 log₁₀ of the ratio. 80% versus 40% is exactly 6 dB. 80% versus 20% is 12 dB. You do not need a calculator for those two.

Transfer correction, conceptually

The calibration block is machined, usually flat, and often a finer surface than the weld cap, the mill scale, or the vessel cladding you will scan. Transfer correction is the dB you add or subtract so that a given-size reflector in the part would produce the same evaluation amplitude it would have produced in the block.

Conceptually you are comparing like with like:

• A back-reflection or a side-drilled-hole-equivalent response on the block versus a comparable reflector or backwall on the part.
• Or a documented attenuation/surface-loss measurement the procedure describes.

You then raise gain if the part is rougher, more attenuative, curved, coated, or coarse-grained, and lower gain only if the procedure allows a credit (rare, and never informal).

Transfer correction is not:

• Scanning gain you added because you were tired of a quiet screen.
• A substitute for building the DAC.
• A number you invent because "stainless is usually +6 dB."

If the procedure is silent, you do not freelance a transfer number. You run the stated calibration and you record surface condition. If the procedure requires transfer, you measure it and write the dB. The general exam tests the idea: block and part are not acoustically identical, and ignoring that fact makes every amplitude call optimistic or pessimistic.

When the calibration is invalid

A calibration dies when any assumption that went into range, delay, velocity, angle, or reference amplitude is no longer true. The usual killers:

Temperature. Acrylic and polystyrene wedges change velocity and path length with temperature much faster than steel. A wedge that was zeroed in an air-conditioned trailer will have a new delay and a new refracted angle on a sun-hot flange. Steel velocity also moves, but the wedge is the first suspect when the exit-point mark no longer peaks the radius. Recalibrate on the block at the working temperature, not at the truck temperature.

Worn wedge. Grinding the wear face to restore coupling shortens the plastic path, moves the exit point toward the front, and can change the incident angle. Yesterday's index mark is now a lie. Delay is short. Skip plots land long or short. A shiny wedge is not a calibrated wedge.

Wrong material velocity. Instrument left on aluminum or on longitudinal while you shear-wave steel. Or a clad overlay velocity applied to the carbon-steel substrate. Locations and thickness readings are systematically wrong even if the DAC holes, on their own block, still look pretty.

Probe, cable, or instrument change. A different crystal diameter, frequency, damping, or pulser setting is a different beam. The stored DAC is for the old beam.

Couplant family change. A heavy paste versus a thin oil changes delay slightly and can change transfer. Follow the procedure; if it freezes a couplant type, that type is an essential variable.

Time. Procedures written against Article 4 or AWS practice commonly require a calibration check at the start of work, after any equipment change, after a pause, and at a stated interval (often on the order of every four hours). A check that shows the reference hole off the stated height or the radius off the stated division is a failed check. The response is recalibrate, then re-examine the work done since the last good check if the procedure says so. The response is not "add 6 dB and carry on" just because you know 6 dB is half amplitude.

A compact field test of honesty: put the probe back on the same radius or the same SDH. If index, range, and amplitude are not where you left them, every indication you plotted since the last check is a suspect measurement.

Realistic exam scenarios

The general exam likes short stories that are really one of the settings above.

• Radius peaks off the old index mark after lunch on a hot deck. Temperature or wear moved the exit point. Recalibrate delay, index, and angle. Do not "aim off the old crayon mark."
• Thickness on a known 1.000 in step reads 1.82 in. You are almost certainly on a longitudinal default while thinking in shear, or the reverse. Fix velocity, then reset zero.
• DAC holes at 1/4 T and 3/4 T will not both sit on a smooth curve. You mixed hole sizes, peaked one hole off the side of the beam, or the instrument filter/damping changed between holes. Rebuild; do not average the mess.
• Part backwall is 12 dB quieter than the block backwall on the same thickness. That is a transfer conversation, not proof the part is laminated. Check surface, couplant, curvature, and cladding before you add a number.
• Reference hole that was 80% FSH is now 40% FSH at the same gain. Sensitivity has dropped 6 dB. Calibration is invalid. Recalibrate. Then decide, per procedure, what must be rescanned.

What topics 2 and 3 assume you already did

Evaluation of plate, bar, forgings, cladding, and weldments is only as good as the clock and the ruler. Topic 2 will ask whether a lost backwall is a lamination or a taper; that question is unanswerable if zero and velocity are wrong. Topic 3 will ask whether a peak is root geometry or lack of fusion; that question is unanswerable if the skip is plotted from a worn-wedge index. Set range, delay, velocity, and sensitivity on the named block. Build DAC or TCG from equal side-drilled holes. Apply transfer only as the procedure measures it. Recalibrate when temperature, wear, or a wrong velocity break the assumptions. That is the whole of calibration as the general exam uses it.