The fastest way to buy the wrong DR detector is sorting datasheets by pixel pitch and picking the finest. On paper the criterion holds; on a real weld, much less so.
For weld radiography the balance of resolution, sensitivity, active area, dynamic range, ruggedness, energy compatibility and software integration shifts with every job: the “best detector” doesn't exist.
Why welds are a specific application
A weld seam changes along its length: non-uniform profile, varying thickness, flaws from porosity to fine cracks. The detector must show small indications and stay stable, exposure after exposure.
Pixel pitch: important, but not enough
Smaller pixels improve resolution but collect fewer photons each at the same dose: noise rises, and a fine crack reads worse on a noisy “higher-resolution” image than on a cleaner one.
Active area and productivity
If pixel pitch governs detail, active area governs time: on long seams, pipes, steel structures and tanks a larger panel cuts exposures and repositioning, the item no datasheet declares.
Dynamic range, sensitivity and SNR
Dynamic range matters when one image holds different thicknesses and absorptions: almost every joint. Signal-to-noise does the rest: the reading risk is mistaking noise for a real flaw.
Scintillator and energy range
The scintillator is where photons become signal: efficiency, sensitivity and resolution pass through it. Chosen for the wrong energy versus source, material and thickness, it returns less signal where the joint demands more.
Ruggedness, IP rating and field use
One example. A panel flawless in the X-ray room meets dust, humidity, shocks and uneven surfaces: without suitable IP rating, weight and cabling, datasheet image quality never survives the shift.
Frame rate, calibration and software
Frame rate matters for radioscopy, fast positioning and semi-automatic inspection. Stability comes from calibration: dark field, gain correction, offset and periodic checks keep the measurement repeatable. Integration with acquisition, image processing, measurement, annotation, archiving, reporting and traceability decides what lands on the operator.
Common mistakes
- selecting only by pixel pitch;
- ignoring dose, source and real thickness;
- underestimating active area and productivity;
- not considering environment and mechanical protection;
- neglecting software, calibration and archiving.
Technical request checklist
- material and thickness;
- joint type, diameter or geometry;
- available X-ray source;
- required image quality;
- operating environment;
- expected productivity.
The final criterion: the most suitable detector is the one your joint, source and site let work at full specification. Everything else is paper.