07-22-2008 10:56 AM
07-22-2008 01:39 PM
07-22-2008 01:59 PM
I have gotten a number of different answers to this question. I have been told by one NI rep. that I can load a LabView 8.5.1 Developers Suite exe. on the cFP. Another strategy maybe to run the cFP program on a PC attached to a cFP via ethernet. I have been looking for guidance on NI's website, but there appears to be no clear and concise discussion of how a cFP is programmed or if there are options. My question concerns the need to spend an additional $3k or more on another LV software package when we already have LV 8.5.1.
Thanks for your response. I always seem to encounter problems using NI's website, for it appears to be written by experts for reading by other experts. Thanks again!
07-22-2008 02:52 PM - edited 07-22-2008 02:55 PM
I've deployed 6 cFP systems to date and last I knew, it was still necessary to have the Real-Time Module loaded in order to create and deploy .exe's to cFP. I don't think you can even load the cFP drivers and libraries until the RT module is installed.
I'll admit that the things seem a bit muddy on the front end. Can I ask what kind of application you have for cFP?
07-22-2008 05:13 PM - edited 07-22-2008 05:14 PM
07-23-2008 07:49 AM
CENTERBOLT, I have found some items buried in the NI maze of information which supports your comments. I am upgrading four test stands which require PID control of RPM and applied force. Also, a count of revolutions, temperature, RPM, and other attributes is required. Typical control circuits which affect power, start, stop, etc are needed also. I have concerns about accomplishing PID control over an ethernet buss between a PC and a cFP controller. We have recently been successful in controlling four test stations employing a PXI and straight LV 8.5.1 programming. This latter system is far more complex and demanding than the former under consideration. We had no significant WINDOWS OS problems. I am not familiar with the term "determinism" used in the context of a Real-Time controller. I have some idea of what it might mean, but can not reduce it down to the level of how it affects my application and NI's cFP information just assumes the reader knows what they are talking about.
I have looked at various cFP modules and understand some of their speed and resolution and bandwidth limitations as compared to what I have been using in my earlier PXI system. I think I can work with the cFP modules available. But, it begins to appear as if the whole concept of cFP relies on a significant software upgrade of an additional costly software module and whatever else I am not aware of yet such as installation fees and R-T toolkits. I understand that if I just want to monitor cFP module outputs and change inputs on a very slow time scale, then I can use a PC with LV 8.5.1 and a non-R-T cFP controller. All of this drives me back to using LV exe. programming and 5B modules and SSR switches along with a large number of cables and attendant interconnections. I would really like to drastically reduce the cabling needed to implement the 5B module approach and go to ethernet, but it appears that the cost impact is very significant.
07-23-2008 07:53 AM
07-23-2008 08:02 AM
07-23-2008 09:56 AM
The term determinism is used in reference to how well a system executes as compared to when you want it to execute. Real Time systems have a high degree of determinism because if they have a loop that is set to run at a specific rate, it will do everything possible to run the loop at that rate. Widows is not a deterministic operating system because cannot determine if your code will be interrupted by some other process. Windows may be fully capable of running a loop at say 50ms, (with plenty of processor to spare) but if another application is started, that application may cause a delay in that 50ms loop rate. For some control applications, this interruption could be a real problem. In Real time OSs, like what runs on the cFP controller, you can specifiy loop rates and priorities. The "time critical" priority, if properly coded, will not allow anything to interrupt it.
07-23-2008 11:01 AM
Any chance of a hybrid combination? Use dedicted off-the-shelf stand-alone PID controllers for loop control and use (c)Fieldpoint for the other I/O functions?
Just a thought.