From: Dave Coder (dave@nucleus.immunol.washington.edu)
Date: Fri Aug 27 1993 - 12:13:07 EST
Although the overhead of Windows may make flow data acquisition on a PC difficult, it is certainly possible to do with with a DOS machine. I and others have done this. Howard Shapiro, for example, has been doing it for quite a while. Several years ago, I put together an acquisition system that collected data from FACS sorters that employ the universal data lister. A '386 PC running Quarterdeck's DESQView was used. The digital data were run into a single digital I/O card, and using double buffering the data were checked for consistency and written directly to disk as FCS 2.0 files. Data acquisition could proceed at up to 10,000 events per second of four parameter data. Data analysis could run at the same time in another DESQView Window. Though not recommended, data could be written directly to a remote server disk over an ethernet connection. (DESQView has quite good multitasking. To see how far we could push the envelope we had a 20MHz '386 with two ethernet cards (one running Coulter's EPINET, and the other running PC-NFS) and two serial ports all operating simultaneously. This allowed downloading data from an MDADS II via EPINET, data from a pdp/11-73 via a 9600 baud serial line, data from a pdp/11-23 via another 9600 baud serial line, and data transfers over the network to a Sun server. All transfers ran without problems.) Bidirectional operations (data acquisition and instrument control) add further levels of complexity that a more robust (and appropriately more complex) operating system such as OS/2 may handle more easily. Dave Coder dcoder@u.washington.edu
This archive was generated by hypermail 2.1.6 : Thu Jan 01 2004 - 17:25:58 EST
