In most cases, it is not necessary to use a counter to timestamp your analog input data. You can easily compute the time of each acquired point when you use hardware timing for your acquisition. In these cases, you can compute the time that a particular point was taken using (start_time + (1 / input_frequency) * index), where index is the number of the point since the beginning (starting with zero). However, if an external clock is used, or if the input is irregular, a counter can be used to count an accurate timebase (usually provided by the board) and provide the time that the point was taken.
The analog input circuitry generates a pulse on the STARTSCAN signal (PFI 7) each time that a scan of data (i.e. one point from each of the channels in the channe l list) is acquired. This signal can be routed to the gate of the counter, which, if it is doing buffered counting, will latch the counter's value at each rising edge of the gate, corresponding to the beginning of each scan. The time of the point will be (start_time + (counter_value / clock_frequency).
Please note that the accuracy of this method is limited by the accuracy with which the start time can be determined. In order to provide better resolution than the Windows clock, you will probably want to start your acquisition using a trigger. The time that the trigger occurred can be measured (or controlled) with an external device. Hope this helps. Filipe
Thank you for your advice, Filipe. I'm a total newbie in "data acquisition", so maybe you can tell me, what I can use as an external clock for triggering the acquisition (frequency-generator, DCF77-PCICard,....)?