fuzzycontrolfreak, on 22 November 2009 - 01:39 AM, said:
Hello All,
I am starting a project where deterministic real-time performance is obligatory. Unsurprisingly, we're on a tight budget too
What I wanna know is, should I stick to FPGA's for real-time applications, or is just using the real-time module to design an application that would run on any platform (M series PCI for instance) enough.
In the latter case, will running the code directly from PC hinder the real-time performance, or is it accessing a real-time processor on the PCI card.
P.S. Generally speaking, what makes a piece of HW real-time compatible or not !!
Thank you for your valuable response,
Omar
Like Ton said,
M series is straight up a
basic DAQ card
You can get
PCI with FPGA but it is R series.
The code you write for the FPGA will run on the FPGA, meaning it will be deterministic.
There is no OS on the FGPA, the code is burned onto the chip.
I am pretty sure you will only need LabVIEW + the FGPA module to program the R Series card.
Depending on you specifications requirements you may find using a target with a RTOS is more flexible. E.g
cFP or
cRIO (which has FPGA as well).
I.e you cannot do floating point maths on FGPA. You can now do fixed point which is cool tho.
You would use the LabVIEW + the Real Time module to program the RTOS.
Well written code will be deterministic.
E.g. A standard PC running Windows OS is not a real time target. If you wanted to turn a PC into a RT target see
here and
here to check if your PC meets the requirements.