Dear All,
I have set out to create a many channel system using the labview realtime operating system, and PXI6733 and PXI6133 cards. My experiment is very time critical, so i am keen to avoid any jittering if possible.
My plan was I thought quite simple, and I assumed similar to what other people would be doing. However I am stumbling because I don't know the correct APIs to achieve my desired design, so perhaps I am doing something crazy.
In pseudo code:
init
----
for each card
{
for each channel
{
initialiseChannel
(DAQmxCreateAOVoltageChan)
}
}
thread A
-----------
for each card
{
for each channel
{
populate a buffer with a wave function
write the buffer to the card
(DAQmxWriteAnalogF64)
}
}
Signal thread B that the data is ready
thread B
----------
DAQmxStartTask(handle)
Signal thread A to prepare the next lot of data...
Obviously then, the requirement is that thread A will create a buffer full of data in the time it takes thread B to output that data.
This all raises some questions:
1. Since I cannot use the same handle for all cards, how can I start all cards playing simultaneously?
2. When I write the data to the card: I intended as above, to write the data for each channel, but I obviously can only address the card handle, so do I have to write the data for all channels as a single memory block? This would be inconvient since I may not necessarily want to use all channels on all cards, so I was hoping to minimse the processing time (in thread A) by only calculating the used channels.
3. If the wave functions I want to write out of the card are longer than the buffer size of the card, then can I write a second buffer to the card, while it is playing the first, such that I can swap them instantaneously on completion, and hence maintain an uninterupted evolving signal? What is the API for that? Or is there an API for a FIFO buffer on the card in CVI?
I'd be grateful for any comments you may have.
James
Currently Confused.
