In a typical flash thermography experiment, energy is delivered to the sample via a pulse of light due to a plasma discharge in a xenon filled quartz flash-tube. Energy is provided by the discharge of a capacitor bank, and the characteristic time for this discharge is typically quoted as having duration in the range of 1 millisecond to maximum rating. However, closer inspection of the flash pulse reveals that the duration that is usually referred to is actually the full width, half maximum value, and that the actual flash duration is significantly longer.
Moreover, the duration values provided by flashtube manufacturers refer only to the visible-light component of the flash, when in fact, there is a considerable flash component in the near-IR as well. The IR component is due in part to the plasma discharge, but also, to the radiation from the tube envelope and hardware immediately after the pulse. Although the visible component duration is identical to the lifetime of the plasma, the IR component lasts longer (on the order of 30-50 milliseconds), and in practice it may contaminate the first several frames of acquired data.
The effect of this flash “contamination” is that for the first few frames, the signal detected by the IR camera becomes a combination of emitted and reflected radiation, where the IR “tail” of the flash pulse dominates the reflected component. In the worst case, these early frames become saturated so that no useful data can be extracted. Although such contamination does not preclude the use of later frames, it weakens most signal processing schemes, since the early data is typically used as a mathematical baseline for analysis of later data.
Improved detection and measurement
Thermal Wave Imaging, Inc. has addressed this problem by developing hardware that allows precise control of flash duration and timing with respect to the integration time of the camera. The Precision Flash Controller (patents pending) is software controlled and it can be added to existing EchoTherm or Thermoscope systems. It truncates the actual plasma discharge and yields a near-rectangular pulse. As a result, saturation and early frame contamination can be eliminated in most cases.
This development has resulted in significant performance improvement of the Thermographic Signal Reconstruction® algorithm used in EchoTherm and Thermoscope, for both detection of deeper defects and quantitative measurement, and closer agreement of experimental data with model predictions.