在本論文中針對神經電生理領域，發展一種架構在IBM PC/AT的間隔式資料擷取系統。這個系統同時採用雙DMA與雙記憶體緩衝區交換的方法，以及神經判別的技術，以記錄高頻的神經電訊號長達數小時。在此系統中，根據神經電訊號的種類，適應性地設定陣列區間的大小並且間隔地擷取神經電訊號，以僅記錄一個完整的神經電訊號，並且避免遺失任何的資料。如此，既可節省大量的硬碟空間。這個間隔式資料擷取系統具有多頻道(16-channels)，高解析度(12-bit)以及高速(200 KHz)的基本需求，並且同時應用DMA與中斷的PC介面技術。此外，針對實驗的需要，可以即時地計算神經訊號波形的各種參數。所有的驅動程式以高階的C語言撰寫並且簡化成一副程式以利呼叫與使用。而訊號的顯示與處理架構在LabVIEW for Windows下。
A simplified interval data acquisition system based on the IBM PC/AT was developed in this paper for neuroelectric research. A dual DMA, dual banks exchange method and a neuroelectric signal detecting technique were adopted to continuously record long period of high frequency neuroelectric signals for up to several hours. The data acquisition system had the basic requirements of multi-channel (16-channel), high resolution (12-bit), and high-speed (200 KHz). A suitable bank size can be -set adaptively depending on the type of acquired neuroelectric signal to record an actual neuroelectric signal completely and without losing any neuroelectric data. In addition, a speak sorting technique was employed and the neuroelectric data was intervally acquired to save a large amount harddisk space. The interrupt and DMA techniques were simultaneously employed, and incorporated into our design. The data acquisition system can compute the spike waveform parameters on-line depending on the needs of experiment. All driver programs were written in highlevel C language as subroutines for easy call and use. The displaying and analysis program were written by using the LabVIEW for Windows.
Chinese Journal of Medical and Biological Engineering 16(3):282-296