Linas Svilainis

Ultrasonic detection and characterization of delaminations in thin composite plates using signal processing techniques

Abstract For the detection of delaminations in thin multilayered composite plates, two different techniques based on the processing of band-limited ultrasonic pulses reflected from the sample under investigation are presented. The first technique exploits the differences between the signals reflected from the flawless and delaminated regions, which are revealed using the modified iterative L1 norm deconvolution. From filtered A-scans, two-dimensional C-type scans are produced in two different ways. The second technique is based on a spectroscopic analysis of the differential signature of a sample, which is obtained by subtracting from the reference signal, reflected by a flawless region, the pulse received at any arbitrary point and presentation of two-dimensional images at various frequencies. The experimental results indicate that both techniques enable one to determine between which layers delamination took place.

LED brightness control for video display application

The technique is suggested for a kind of pulse-width-modulation LED dimming in LED video displays. The significant increase of the gray levels number and image refresh frequency can be achieved when the large number of LEDs is controlled serially. The need for a large number of grayscale levels and high refresh rate is explained. The controlling data flow and the amount of buffer memory are the same as with binary-weighted pulse-width-modulation. Spreading the switching pulses in time reduces the electromagnetic interference.

Application of orthogonal ultrasonic signals and binaural processing for imaging of the environment

Data acquisition rates in ultrasonic imaging systems are limited by the finite value of the speed of ultrasonic waves. In order to improve the imaging speed, it is proposed to perform simultaneous scanning of the environment in different directions. In order to avoid cross-talk between adjacent channels in different directions, different orthogonal signals are transmitted. Application of cross-correlation processing and non-linear iterative deconvolution enables the reliable separation of signals transmitted by different sources and reflected by multiple targets. The spatial positions of the targets are found using the data obtained after the non-linear deconvolution as the initial data for binaural or tri-aural processing. This approach has been exploited in ultrasonic sonar used for navigation of mobile robots.

LED Far Field Pattern Approximation Performance Study

Some factors influence on light emitting diode (LED) far-field pattern (FFP) approximation performance is analyzed. The mean square root intensity error, mis correlation factor and half power and peak angles variance have been suggested for performance evaluation. The polynomial order, range and the noise influence have been analyzed. The best function candidate has been named.

Considerations of the Driving Electronics of LED Video Display

An overview of the electronics used to drive the light emitting diodes (LEDs) displays is presented. The common problems and the solutions are given. The driver technologies overview indicates the remaining challenges: the high current distribution and the LED dimming range. New driver topology suggested could solve the mentioned shortcomings.

Optimization of the ultrasonic excitation stage

The ultrasonic transducer excitation circuits have been investigated. The system structure is presented where excitation circuit is planned to be used. The possible excitation solutions are presented and discussed. The result of impedance measurement of ultrasonic transducer has been used to derive the Butterworth-Van Dyke electrical model.Then this model was used for PSpice modeling of candidate topologies. The chosen circuit was simulated and experimentally investigated. In this article a structure of ultrasound generator is analyzed.

Analysis of adaptive imaging algorithms for ultrasonic non-destructive testing

Abstract An adaptive imaging algorithm for ultrasonic non-destructive testing (NDT), which enables one to compress the dynamic range of the received ultrasonic signals, is presented. For this purpose, the data obtained in previous steps during B or C-scans are exploited to predict the system gain necessary at the next scanning position. Such an approach allows one to collect data during a single scan without primary knowledge about the object under investigation. This increases the reliability and the speed of NDT. The reduced dynamic range allows one to obtain high quality acoustic images with lower resolution high-speed analogue-to-digital (A/D) converters because the initial full-range signal can be restored after A/D conversion using collected gain values. The results of an experimental investigation of the measurement errors for various gain adjustment methods are given. The method described has additional advances in sophisticated signal post-processing. The images obtained using the synthetic aperture focussing technique (SAFT) are presented to demonstrate the differences in image quality. The sources of possible artefacts are analysed.

Half bridge topology 500 V pulser for ultrasonic transducer excitation.

Application of half bridge topology for ultrasonic transducer excitation using long pulse trains is presented. The novelty of the approach is the high speed solution for a high side drive. A commercially available high speed digital isolator and a high speed MOSFET driver were combined to give the possibility to deliver fast driving signals to a high side N-channel MOSFET. The experimental investigation indicates that the output amplitude of the fundamental harmonic can reach 624 Vp-p for light loads and 552 Vp-p when driving 50 Ω loads. The operation frequency at such voltages can reach 10 MHz for unloaded or 50 Ω load condition and 6 MHz when driving capacitive 3000 pF loads. The output impedance is 13 Ω for voltages below 500 Vp-p and 16-26 Ω for voltages 500 Vp-p and above.

Pulser for arbitrary width and position square pulse trains generation

Special kind of Spread Spectrum excitation is analyzed: trains of arbitrary position and width pulses (APWP). Such signals are easy to generate and to handle in digital domain. They offer an advantage over a single pulse: spectral response has no dips and posses higher that single pulse energy, yet being capable of short duration after matched filer compression. Pulser design for such signals generation is described. In order to have the electrical power consumption economy, both high and low - pulling elements are active. Such decision would allow to save the energy when pulser is loaded with capacitive (piezoceramic transducers, capacitive micromachined ultrasonic transducers, CMUT) or inductive load (electromagnetic acoustic transducer, EMAT). In order to evaluate the amount of energy consumed theoretically and real generator performance, energy per pulse was suggested as comparison parameter. Reasoning for output topology, element parameters is given. Pulser structure is described. Investigation of prototype performance is given. Output signal waveforms, energy per pulse performance is presented.

The influence of digital domain on time of flight estimation performance

Analysis of the time of flight estimation performance when signal digitization is used is presented. The crosscorrelation function peak position was used as time-of-flight estimator. In order to cope with the discrete nature of the cross-correlation the time of flight subsample estimates used interpolation techniques: parabolic peak, derivatives zero crossing, zero crossing of imaginary of Hilbert transform, double zero crossing and frequency domain phase. Noise density component of the Cramer-Rao lower error bound was modified to include the jittering and quantization noise. Basing the real amplifier noise density theoretical analysis concluded that taking the 180 nV/√Hz as lowest noise density at ADC input, 10 bit is needed for 100 MHz sampling. Numerical experiments have been conducted using Matlab to investigate the standard deviation of ToF in presence of additive white Gaussian noise and signals digitization. It has been concluded that systematic errors are related to subsample interpolation and a...