We have an application involving 16-24 channels being sampled at 1KHz in LabVIEW RT and being streamed to the Host PC via Network Streams. As the data come in, we display the data by averaging 50 points at a time (for an effective Chart Display rate of 20 Hz), and saving the "data we want to save".
How do we signal which data are to be saved and which are not? The RT side, which is generating the data and "knows" if it is supposed to be saved, uses one of two mechanisms (we've used both mechanisms at different times). One is to add a "Save Me" flag to the data, for example by adding an additional Flag channel that is "1" if this point is to be saved and "0" if not, with logic for stripping off this channel at the Host side and managing the opening and closing of data files, or using another communication channel to send "Start/Stop Saving" signals to the Host.
This puts all of the responsibility for synchronizing Start and Stop of saving with the acquisition of the data on one (Real-Time) system and creates a parallel "Consumer" loop in the Host that is told when to start and stop streaming the data. Works like a charm. All of the timing stays with the Remote Acquisition system (which has the accurate clock). If it turns out that you want a single "Start" signal and want to save data for 0.1 seconds, then you can (at the Host end) translate the "time" criterion into a "Number of saved Points" criterion based on the desired Time-of-Saving and Sampling Rate.
Bob Schor
