Implementing RTSI using Signal Express

[BrianK]

Hi,

I'm interested in synchronizing the analog signal recording (start time, sample acquisition, stop time) of 3 PCI-6289 cards, and possibly syncing them with 3 to 5 more PXI-6221 cards, which are connected to the same host PC via a PCIe-8361 to PXI-8360 interface.  For the moment, I'll concentrate on the 3 PCI-6289 cards.  I have used an RTSI cable with VI Logger, no longer supported, and assumed that it would work well with Signal Express.  I set up a NI-RTSI Cable device in Measurement & Automation, which includes the 3 PCI cards.  The 3 PCI cards are the only devices connected to the RTSI cable.  When checking the properties of the RTSI Cable device, I have left all of the boxes unchecked, which I believe is the proper setting for my application.  In Signal Express, when adding a device to the left hand panel to acquire data (analog voltage signals), I still see each individual card (Dev1, Dev2, Dev3, etc), but there is no RTSI device.  If I look into the advanced Timing features of each of the three DAQAcquire panels, I can apparently use one card as the master clock, and the other two cards will take timing signals from the master card (is this automatically via the RTSI cable, or will I need to make external connections from clock out on the master card to clock in on the slave cards?) but again, there is no mention of the RTSI cable, or the RTSI device setup in MAX.  Using the clock from one card in a master/slave situation helps, but will it start and stop all three cards simultaneously, or will any apparent synchronization between the cards simply be a coincidence? 

Any suggestions are appreciated.

regards,

Brian

[DFGray]

SignalExpress does not require the RTSI device to synchronize DAQ devices, but does require the cable to be present or the devices to be in the same PXI(e) rack.  If you set up multiple DAQ devices on the same bus in the same acquisition, SignalExpress will synchronize them for you using the RTSI lines.  This is not true of other device families (e.g. NI-SCOPE).  There is a limited ability to synchronize multiple devices on different buses by specifying clock and trigger lines.

As an historical side note, the original synchronization system of SignalExpress was written before the RTSI device code.

I am unsure whether I have answered your questions.  Please let me know if not.

[BrianK]

Hi, yes, that does help, thank you for your reply.  Signal express does make it easy to work with multiple devices, as I am, though something seems to be amiss with the timing between cards.  For example, setting Signal Express to do nothing more than log data at 100 pts/sec on all channels on all three cards (PCI-6289), and using the software trigger in Signal Express to start the digitization process, and the stop button, to stop data logging, certainly seems to be a very convenient solution.  However, the longer I record (an hour seems to be enough time), the more likely that a signal recorded on all three cards, will not synch, point to point, (comparing signals recorded on channel 0 on all three cards, which I believe should be captured simultaneously through out the recording process, for example) and data recording does not stop simultaneously on all three cards.  I can handle the data logging not stopping simultaneously, but my source signal apparently not being recorded synchronously is a whole other problem, and the impetus for my question.  Thanks for the tip about NI-SCOPE, I was going to start working with that on a different project.

I am using the most recent versions of SignalExpress, NIDAQmx drivers, etc., on a PC running 64-bit windows 7.

regards,

Brian

[DFGray]

I thought that setting up the acquisition in a single DAQmx Acquire block would synchronize the different boards on the hardware level, regardless of trigger type.  Your experience tells me that may not be true, since it sounds like your cards are drifting in time due to slightly different sample clock frequencies.  I will query the designers of that block to confirm what the expected behavior is.  If it is unsynchronized due to the software trigger, you can probably still do a good synchronization by using one board as a master and sharing its sample clock and trigger with the other two.

[BrianK]

I look forward to learning what the designers at NI think about this, and thanks for making the inquiries.  In the mean time, I'll be pursuing the master/slave scenario, or perhaps a separate external clock for all three cards to use.

Thanks,

Brian