Three-Channel High-Speed Data Acquisition and Pulse Compression System Implementation

Literature Overview

This paper by Wang Wei, Chen Xiaoying, and Han Yueqiu (Beijing Institute of Technology) published in Modern Radar (2005, Vol. 27, No. 6) describes the design and implementation of a three-channel high-speed data acquisition and digital pulse compression system. While this work falls outside the traditional domain of steel pipe manufacturing, the underlying signal processing and data acquisition principles have direct applicability to ultrasonic testing (UT), phased array ultrasonic testing (PAUT), and other non-destructive evaluation (NDE) methods used extensively in pipe and fitting quality control.

Core Technical Architecture

The system employs a three-channel parallel architecture designed for real-time signal processing of radar returns. The key components and their specifications are:

Signal Processing Methodology

The pulse compression module adopts a dual butterfly operation unit with parallel processing capability. The base-4 butterfly unit simultaneously completes FFT, complex multiplication, and IFFT operations, reducing hardware scale to approximately one-third of conventional implementations. The system employs block floating-point algorithms to improve dynamic range, which is critical when dealing with signals that have both strong direct returns and weak reflected components—a scenario analogous to ultrasonic testing of thick-walled pipes where near-surface reflections can mask deeper defects.

Relevance to Pipe and Fitting Inspection

Connection to Ultrasonic Testing Principles

The pulse compression technique described in this paper is directly analogous to the signal processing used in modern ultrasonic NDE equipment:

1. Wideband excitation and matched filtering: In UT, a broadband transducer is excited by a short electrical pulse, and the received echo is correlated with the transmit waveform to improve signal-to-noise ratio—essentially the same principle as radar pulse compression.
2. Dynamic range requirements: The 14-bit ADC with block floating-point processing provides approximately 84 dB of dynamic range, comparable to high-end UT systems that must detect small reflectors (such as lack of fusion in welds) against strong back-wall echoes in thick pipe sections.
3. Real-time processing: The 57.70 μs processing time for 1024-point compression is relevant to phased array scanning speeds. In PAUT inspection of girth welds on large-diameter pipes, the system must process multiple A-scans per millimeter of scan travel.

Application to Pipe Welding Quality Control

For engineers involved in welding quality assurance, the principles in this paper inform several practical considerations:

• Signal-to-noise ratio optimization: In UT of welded pipes, particularly for detecting lack of fusion defects in HFW or LSAW welds, the ability to compress noise while preserving defect signal is paramount. The dual butterfly architecture enables faster processing, allowing more sophisticated filtering algorithms to be applied within the same scan time budget.
• Variable resolution requirements: The ability to switch between 256-point and 1024-point processing corresponds to the need for different inspection sensitivities. Short 256-point windows are suitable for near-surface defect detection in thin-wall pipe, while 1024-point windows provide better discrimination for deep defects in thick-wall pipe or pipe fittings.
• Floating-point output: The 32-bit IEEE floating-point output format ensures that amplitude measurements maintain precision across the full dynamic range, which is essential for quantitative UT techniques such as TOFD (Time of Flight Diffraction) sizing.

Technical Analysis and Engineering Implications

Hardware Efficiency Considerations

The reduction of hardware scale to one-third through the base-4 butterfly architecture is significant for embedded NDE systems. In field-deployable pipe inspection equipment, power consumption, thermal management, and physical size are all constrained. A more compact processing architecture enables:

• Smaller, lighter inspection instruments for field use
• Lower power consumption for battery-operated portable equipment
• Reduced thermal issues in confined inspection environments (such as inside pipe during internal inspection)

Block Floating-Point Algorithm

The block floating-point approach addresses a fundamental challenge in signal processing: maintaining precision across signals with widely varying amplitude levels. In pipe inspection, this manifests as:

• Strong back-wall echoes in thick-wall pipe (potentially 40–60 dB above noise floor)
• Weak indications from small planar defects (such as 2 mm lack of fusion) that may be only 10–15 dB above noise
• The need to preserve both strong and weak signals simultaneously without clipping or quantization loss

Key Reflections

From a NDE engineering perspective, this paper demonstrates that the signal processing challenges in radar and ultrasonic testing are fundamentally identical. The three-channel architecture could be directly adapted for multi-element ultrasonic transducer arrays, where each channel corresponds to an individual transducer element in a phased array probe. The variable-point processing capability is particularly valuable for inspection systems that must handle both thin-wall and thick-wall pipe without changing hardware configuration.

The practical lesson for welding and pipe inspection engineers is that the quality of defect detection is ultimately limited by the signal processing chain, not just by the transducer or excitation source. Investing in advanced pulse compression and filtering algorithms can significantly improve detection sensitivity without requiring more powerful equipment. This is especially relevant for inspection of high-strength steel welds (X80 and above) where material attenuation is higher and defect signals are weaker.