09-17-2008 10:21 AM
09-17-2008 02:49 PM
The answer is both can be used.
Software does the actual calculations and joins the separate records together to create the single record returned for RIS. However, the driver will use multiple records on the device to improve the performance of RIS if it is possible.
So, for example: Say you had an 8MB PXI-5152. This board has 8bit sampling, and has a normal maximum real-time sample rate of 1GS/s, with a maximum RIS rate of 20GS/s. This is an oversampling factor of 20, so the largest RIS record you could take would be approximately 160MS in length. This is because the driver can split the RIS acquisition into several different acquisitions of 8MB, then interleave and average them as appropriate. This would represent "single scan."
Anything smaller than that, and the driver may be able to take multiple records in a single acquisition to reduce the number of acquisitions required to fill all the bins. This is merely an optimization to enable the digitizer to find more qualifying records faster.
* The 5152 also supports TIS (time-interleaved sampling, also known as ping-pong) for 2GS/s on a single channel, but this doesn't apply in this example