06-23-2007 09:57 AM
06-23-2007
10:22 AM
- last edited on
03-14-2025
01:03 PM
by
Content Cleaner
Raghunathan,
I would look at the M-Series cards for you application: http://www.ni.com/dataacquisition/mseries.htm. For straight control with no analysis maybe a PLC is a better choice because of cost. For any real analysis LabVIEW would be the choice.
For deterministic you should be using LV RealTime if timing is critical. It is not LV's fault and is a function of the operating system. LV RT: https://www.ni.com/en-us/shop/product/labview-real-time-module.html. I find LV to work quite well and maybe a function on how you have your code structured.
LV's strong point is analysis and is MUCH better than anything you will find in the PLC software.
- Documentation of all the graphic code is very, very difficult if not impossible. Thus offline study of code from a printed copy ( like we do in C ) is not easy.
Solution: Experience. LV code properly written is very easy to read, analyze, and review. If you make a hair-ball out of the code then yup, can't read it.
- Nested structures are a big pain to troubleshoot and debug.
Solution: Use more subroutines and limit nesting to at most 3 or 4 levels.. LV has excellent debugging capabilities far beyond any other programming languages
- As the control complexity increases, wiring becomes a nightmare.
Solution: See hair-ball above 😉
- Even with the best of intentions and efforts, presenting a neat wiring with comments is a pipe dream for complex controls...
- When working with sub panels, setting break points and single stepping are almost a joke.. particularly with time critical functions.
Solution: Strong Disagree, anyone can use any software package and make a mess. Take a NI class or two, watch some WEB videos on NI's site, and give it a chance. Experience is the only solution for your problem. Hire a certified developer to work with you and you will see the light. I have created code in LabVIEW to control submersible cutters, smoke chambers, multiple DAQ applications, Robotic Controlled Laser Measurement Systems, and detailed analysis of parts in conjunction with a PLC when it was far beyond the abilities of the PLC to analyze the data. LabVIEW is up for the task and can take anything I throw at it.
http://sine.ni.com/apps/we/nievn.ni?action=display_weod_main
Too many resources to list.
06-24-2007 12:48 AM
06-24-2007 02:03 AM
@Raghunathan wrote:
To make my query more clear, let me eloborate :
If there is a machine with following requirements :
- Handle some 10 user push buttons / switches.
- Read in some 28 field level digital sensors like fluid level, limit switches, proximity switches, safety doors etc.
- Control about 14 solenoid valves based on the above inputs .
- Read in some 7 analog channels and plot / save the values.
- Generate 0-10V command signals for motor rpm and system pressure.
- Generate reports after each test and print it.
I'm currently working on something not dissimilar in terms of equipment (although there is more I/O) and we're using cFP-180x (which is only a remote I/O module). If we needed it to be more RT, we could basically take the control code and download it to a cFP-2xxx (although it might have to be adapted). These are expensive compared to PLCs, but they are more powerful and they run LV. We have done similar systems in the past (with more and less I/O and with various combinations of PLCs and FPs) and LV can definitely handle it, just as long as you don't need too much speed and that you accept that you are running on a non-RT system which could theoretically fail.
As Matt suggested, writing a complex program in LV is no different from writing a complex program in C or ladder logic - it's just that you would get farther without training because LV is more forgiving. The key at these level is to optimize - you have to know how to avoid LV allocating more memory than you like and how you can make it faster. This is all assuming that you are allowing your machine to potentially fail and miss sometimes, because even though XP is reliable when it's clean, it could still get stuck. If the machine can absolutely never fail, you need to go RT.
06-24-2007 02:04 PM
06-25-2007 01:35 AM
That would require two VI's you can't have two SDs with the State Diagram in one VI.
@unclebump wrote:
Maybe the state diagram toolkit would help. Is there an option with this to have parallel independent state diagrams running at the same time??
06-25-2007 06:53 AM
Raghunathan,
One nice thing about NI and lab is you have many options. Choosing the right solution depends on several factors:
- Is this going to be running for months / years at a time? If yes, then consider a Real-Time solution Windows good for a few months then needs a reboot.
- What is the sampling speed needed? Microseconds, milliseconds, seconds?
Slow but easy: FieldPoint using Ethernet or DAQ Card to lab running on a Windows OS
Faster but more complex: FieldPoint using a Real-Time controller but need lab Real-Time
Fastest but most complex: cRIO running Real-Time but needs lab FPGA & lab Real-Time
If you want monitor and control of Real-Time application then you need to create a lab application running on a PC
Solution: National Instruments provides many choices at price and performance levels that you just can't find in a PLC system
Ton,
You can hve more than ore if you disconnect from the state editor.
06-25-2007 10:35 AM
Translation to English....
Ton,
You can have more than one State Diagram if you disconnect from the state editor. Recommend technique is to build up the two individual state diagrams individually then copy the cde into the main application.