Hey NIck,
Sorry about the misinformation before, but as Sal stated, the USB-9215A
does support high-speed USB transfers. I think that upgrading to
a USB 2.0 interface will increase your performance. The operating
system will still have to service the device through interrupts, as
opposed to DMA transfers as your PCI card uses, and you will not gain
any performance increases in this area with a high-speed USB
port. However, only one interrupt is created for each data
transfer. Once the operating system services that interrupt, you
will be able to perform the data transfer much faster with the
high-speed interface, and you should see a performance increase.
With regards to your other questions, whenever you use the DAQmx Timing
> Sample Clock VI, or create a task in MAX and set the Task Timing
to N Samples or Continuous, you are creating a hardware timed buffered
acquisition. The sampling rate you set determines the rate of the
sample clock for how data is acquired on the device. The 'samples
to read' you specify in the Task Timing in MAX, or on the DAQmx Timing
VI, is just used to calculate the buffer size to allocate for the
task. The actual number of samples being read into LabVIEW at a
time is determined by the 'samples per channel' input of the DAQmx Read
VI. If you set this 'samples per channel' input to -1, you are
still performing a hardware timed buffered acquisition, with a buffer
size allocated based on your settings under Task Timing in MAX or in
the DAQmx Timing VI. The device will be continuously acquiring
the samples into a buffer in system memory, and all the interrupts and
data transfer from the USB device to system memory will occur behind
the scenes. However, when the DAQmx Read VI is called, it will
grab all available samples in the buffer. If the DAQmx Read VI is
within a while loop, the rate of the while loop will no longer be
controlled by the rate of acquisition and number of samples to
read. Instead, the loop will execute as fast as possible, and
whenever it begins another cycle and calls the DAQmx Read VI, the
program will not wait at that VI until a certain number of samples have
arrived, but will instead immeadiately grab all available samples
(which may be zero), and continue executing the rest of the code in the
loop.
From your straightforward requirements of what you are trying to do, I
would recommend upgrading to a USB 2.0 port. I would then perform
a continuous hardware timed acquisition, with your required sampling
rate and increase the 'samples to read' in the Task Timing in MAX and
the 'number of samples per channel' input of your DAQmx Read VI
(keeping both the same), until you no longer see buffer
overflows. If you want a fixed number of samples to average each
iteration, then don't use the -1 samples per channel method I
recommended above. As you increase the number of sample to read,
you will be reducing the number of interrupts the OS must service to
transfer data from the USB device. I would also recommend
reducing the amount of of other processing that is done in your while
loop as much as possible. Try to move as much of your analysis to
post acquisition as possilbe, so that LabVIEW is not tied up with
analysis tasks and has time to execute the DAQmx Reads.
I hope this helps to answer some of your questions and sorry if I complicated things by going into these advanced timing topics.
Regards,
Travis G.
