UT-Trainer — Ultrasonic Testing Simulator

Full version & prices →

Real device photo with interactive controls — the same UFD 900 used in my Udemy courses UT Level 1 and UT Level 2.
In this trial version you calibrate (adjust) the probe on a step block, then find flaws in a steel plate and identify an unknown material by ultrasound. Use the Tour function to be guided step by step.

◍ POWER OFF

COUPLE

DAC

ALARM

Reference

Scan conditions and display settings

Sound field

Position: 150 mm

Single-element straight-beam probe.

Drag the probe along the plate's surface. Click it to lift it off (line goes flat) or place it back down (coupled).

Live readout

SelectedRANGE

Gain12.0 dB

Range100 mm

Velocity (MATVEL)5920 m/s

Gate30–50 mm @40%

Gate reading– –

DAC/AVG points0

How to use

Press Power first. The 5 tiles on the right of the display are real menus, exactly like the original device: Range, Probe, Gate, Set and Mem jump straight into their own menu (BASE/PROBE/GATE/SET/MEM), DAC/AVG cycles DAC → AVG → off. The Menu key (between dB+ and ↑) opens the grid of function groups — move the cursor with the D-pad, Enter loads a group, Cancel closes it without changing anything. Inside a menu, tapping a tile selects it (pink outline) or fires the action, and ↑/↓ step through a menu's own functions, scrolling if it has more than 5. Change a value with the rotary encoder (click left/right or scroll wheel) or the ◀/▶ arrows. dB+/dB− change gain directly, even outside the GAIN menu. The round key with the zigzag icon is AutoGain. The second square icon is Zoom. Freeze shows "Freeze" on screen. On the test specimen to the right, drag the probe along the plate's surface or click it to lift it off / place it back down — without coupling the trace stays flat. A hidden flaw sits somewhere along the plate: scan slowly and watch for an extra echo ahead of the first backwall signal — like a real near-surface reflector, it rings its own repeat-echo train too (fading faster than the backwall's), so don't mistake those for deeper flaws.

Menus (from the original manual)

BASE (Range): RANGE/VELOCITY/D-DELAY/SCALE/THICKNESS — D-DELAY is a scan-level timebase offset the operator dials in (e.g. a known immersion water path), separate from the probe itself. Double-click VELOCITY to flip straight between steel's longitudinal and transverse velocity (5920 ↔ 3250 m/s) instead of dialling the whole way. PROBE (Probe, scrolls): TYPE/ANGLE/P-DELAY/X_VALUE/X_FREQ/ X_SIZE/DAMPING/CAL PROBE/Name — P-DELAY and X_FREQ are that probe's own fixed, factory specs (transit-time delay through its wear plate/wedge, and its build frequency) — both jump to a new default whenever TYPE changes (default P-DELAY 2.00/11.50/5.00 µs, X_FREQ 5.00/5.00/4.00 MHz for straight-beam / T/R / angle), exactly like swapping to a physically different probe. A P-DELAY that doesn't match the probe's real delay shifts every reading — CAL PROBE (or adjusting it by hand) finds the probe's real value. DAMPING switches the damping resistor between the manual's four values — the echo is tallest at 400 Ω, loses 5 % at 150 Ω, drops to 75 % at 100 Ω and to just 26 % at 50 Ω, so stepping through them and keeping the highest amplitude is a real search. CAL PROBE is the manual's semi-automatic calibration: key in the two thicknesses T1/T2, peak each backwall echo in Gate A, and the device calculates VELOCITY and P-DELAY. P/R (Menu grid only): ENERGY/P WIDTH/PRF/ RECTIFY/REJECT — RECTIFY switches full-wave, half-wave and true RF (RF is skipped while a DAC/AVG curve is shown, per the manual). GAIN (Menu grid only): GAIN/STEP/SCAN/dB CORR/AUTO GAIN. GATE: GATE SEL only picks which gate (A or an independent Gate B, cyan) the rest of the menu edits — it does not switch it on. LOGIC sets the selected gate to POS (alarm while an echo exceeds the gate height), NEG (alarm while no echo reaches it — e.g. to monitor the loss of the backwall echo) or OFF (invisible, no reading, no alarm — same as the real UFD900's GATE/LOGIK function) — then START/WIDTH/HEIGHT, DETECTION (PEAK = always the highest echo in the gate, distance and height, even below the threshold — which then only sets the alarm; shown as a large triangle overlapping a smaller one; FLANK = gate-crossing point, one triangle) and INDICATION (SINGLE, or DUAL = distance Gate B − Gate A). DAC (DAC/AVG key, then BUILD DAC): gate a reference echo (e.g. the side-drilled holes of the DAC block), save with ↑ (not Enter!) — the first point is brought to 80 % automatically — repeat (up to 16 points), Enter finishes. Besides the recorded curve (GL) the RL/SL/EL curves are shown, offset by DAC RL/SL/EL (STANDARD Custom, −50…+50 dB); SIZE REF picks the reference curve the gated echo is rated against in dB. During the recording BUILD DAC saves like ↑. EDIT DAC selects a point (◀/▶): ↑ records only that point again, the knob moves it, Delete removes it. TRANSFER sets the transfer correction (dB, on the real instrument a box after Enter); ATTENUATION (trainer extension, dB/m) corrects the curves for a test piece that attenuates more than the block. AVG (DAC/AVG key twice): DGS evaluation as in the manual, 5.4 — BUILD AVG shows the probe data and the reference reflector (BW/FBH/SDH); gate the reference echo, ↑ stores it, Enter calculates the curves AVG-UL/-ML/-DL (equivalent reflector sizes in mm, drawn in teal). The measuring line then shows the ERS of the gated echo. SET (Set key): SET NO. dials in a channel number 1–50 with ◀/▶ or the knob (like any numeric field) — Enter opens the FILE LIST, a table of all 50 channels; picking one applies it straight to the live instrument, so the A-scan updates immediately (matches the manual's own SET NO. + Enter behaviour, 4.32 — there is no separate RECALL function). SAVE first asks what inspection the setup is for and files it under that description plus its own number — type "V2 Test" on channel 2 and the file is called V2 Test - Set No.2; leave the box empty and it is just "Set No.2". RENAME changes that description later. The save itself is immediate (toast: "Saved to #N") and separately offers the file as a downloadable .json: watch for a save prompt from the app around the UT-Trainer and confirm it — until you do, the file only lives in the UT-Trainer, not on your computer (browsers can't write straight to a folder on the Desktop on their own, so that prompt is the only way a copy reaches your PC; pick "UT-Trainer" on the Desktop once and most browsers remember that folder after). Saving over an already-occupied channel asks for confirmation first, same as DELETE/DEL ALL. MEM (Mem key, scrolls): REPORT dials in a report number 1–100 the same way — Enter opens its own FILE LIST, sorted by name or date (toggle with ORDER). A report is proof a specific inspection was carried out — SAVE logs the whole setup and result there (not just the reading), asks for its description the same way (V2 Test - Report No.8) and downloads it like a SET, also confirming before it overwrites an existing report. Opening one from the list also applies its exact settings to the live instrument and shows every setting, the result and the A-scan as it was when saved, just like the manual's BERICHT function — Cancel leaves that picture on screen rather than reverting it. A report itself is never overwritten by this: it stays in the log exactly as saved. To actually repeat that inspection, save it as a SET afterward — that's what gives you a reusable, nameable channel for next time; the next SET SAVE (from there or anywhere else) is what replaces whatever's on screen. RENAME/DELETE/DEL ALL manage the log. In the report list (MEM › REPORT › Enter) the PDF button opens a large, readable report that can be saved as a PDF — this UT-Trainer's stand-in for the manual's report EXPORT. SET/REPORT entries also save automatically to this artifact's own storage (when that's available), so they survive a reload without any manual action — the .json download on SAVE is still there too, for a backup copy on your own machine (there's no way to load that file back in — the manual's IMPORT isn't modelled here).

Material calibration (Probe)

Picking a material here dials MATVEL to its reference velocity, same as typing it in on the real device. Every test specimen is steel (5920 m/s) except the DAC Block (aluminium, 6360 m/s — no mystery, it’s labeled on the block) and the Unknown Material block, whose real velocity is deliberately hidden — its known 30 mm thickness only reads correctly once MATVEL is right. Wrong MATVEL = wrong depth reading everywhere, just like on the real device.