I am communicating across USB, using a proprietary protocol, with some custom hardware I've built. I have a GUI that handles all the communications/interaction with that hardware and a (C#) DLL which exposes all the relevant USB functionality. I need to write a LabVIEW driver (VI) for communicating with the hardware. My thought is that I just use LabVIEW to open up my GUI and have a socket with which I expose all the relevant control to LabVIEW with. Is it possible to open a socket in LabVIEW and communicate with my GUI? Is this a bad approach or should I just try and make LabVIEW invoke the DLL and handle the hardware control instead of my GUI (polled communications, solicited/unsolicited commands, etc)?
IS there a reason you want to use your GUI only? In terms of time, I would say build a good front panel in LabVIEW and just communicate to the hardware directly using the DLL. Adding the GUI is just an added layer of complexity which might be difficult to maintain later on? Why not do everything in LabVIEW if you can?
Yes, LabVIEW supports sockets using both TCP/IP and UDP.
You should be able to create a program/service that continually runs acting as TCP/IP server. You can send commands and receive responses as strings. If you need to pack data, you can use the flatten to string command.
Essentially, your application should be structured as a loop running the TCP/IP server, and another loop that actually communicates with the instrument. This might change if you need to get data back from the devices to your TCP client. A producer consumer model, if you will :)
To get you started off, open up the NI Example Finder (Help -> Find Examples) and browse to Networking->TCP and UDP-> Simple Data Server.vi
It depends who is going to be using the LabVIEW driver and for what. If you're handing over this hardware to someone else who is going to want to create their own application(s) for it, they would probably prefer to talk directly to the DLL rather than go through your GUI. If it's more about automating your existing software from LabVIEW to do testing or repetitive tasks on the hardware, for example, then driving your GUI from LabVIEW might be less work.
Related
I am using Ethernet to feed data into my LabVIEW system from a custom FPGA hardware. I wanted to know if there is some vi functionality or front panel functionality to display the throughput of the data received from the Ethernet port into the LabVIEW system or the PC.
Scouting through the National Instruments forums, I did not find a solution for showing this statistic in LabVIEW.
This level of diagnostic is not something that is typically available from a LabVIEW function.
If you want to determine this data rate, you have a couple of options depending on your implementation:
Monitor the data rate yourself: Presumably you are catching and processing this data somewhere so write your own wrapper that counts the number of cycles and packets it is processing
Look for a .NET implementation: LabVIEW has fairly nice integration with Windows .NET libraries, and if you are able to find an appropriate library that implements this functionality you should be able to work with it.
Run a monitoring tool like iPerf via the command line periodically. This may not return as accurate data since it is external, but it should be reasonably straight-forward for testing.
Is it possible to access external hardware without using a driver, i.e. not having the driver abstraction layer in between program and external device?
Can you use a device by implementing your own driver-like controlling/handling directly in your program code?
I'm trying to understand a program that implements a Modbus protocol and some very specific Modbus configurations. Now I don't know how exactly it communicates with the Modbus devices.
It looks to me that this is very similar to what a driver does.
But can it even communicate DIRECTLY with the device without having a driver installed?
Yes, there are several micro-kernel OS's that always configure this way -- drivers are entirely implemented outside of the kernel.
The first thing you likely need is to get access to the device's registers; typically performed with mmap(), you may need to dig around a bit to find the right settings for cacheability, etc...
Second problem is interrupts. Unless you are running something like QNX, you won't have a way to have interrupts signal your program directly. You will probably have to turn them off and poll the device periodically.
If you are using linux and need io ports (inb, outb, etc...) man ioperm for more information.
It's about a mechatronical system that needs to be controlled via software. It is not yet clear in which language it will be written, but since it is not important, let's just say in Java.
The 1. thing is that we will need to send messages via CAN. So we have the control software, some event happens and we send a message via CAN. The mechatronical system will react.
Now the 2. thing is that obviously it would be good to be able to test the software without a real mechatronical system, since it reduces effort. So I thought about writing another program, a simulation program.
So I imagine that the simulation program notices when a CAN message is sent and reacts to it.
How is a good approach to accomplish that?
I mean for the real mechatronical system the control software needs to send a CAN message directly on the bus(, maybe via some native code). For the simulation program some kind of interprocess communication is needed. How must the control software be designed that it doesn't care if there is some simulation program that is listening or a real mechatronical system that gets the CAN messages?
My first thought was that the control software always sends "CAN messages" via an interprocess communication approach. Let's say for the sake of simplicity it is RMI. Then to send real CAN messages via the bus there is some module in the same control software that gets the "CAN messages" via RMI and forwarding them to the real CAN bus.
Now the simulation program is able to receive the "CAN messages" via RMI, too, and can react to it.
Is that a good approach? Because I see that there is some overhead in the control software by communicating to itself via interprocess communication, which is not neccessary in principle. But I see no other possibility to have an abstraction layer, such that I have no special code for the simulation program in the control software.
Thank you for feedback!
You're describing one aspect of Hardware-in-the-loop testing. It's a standard approach for developing mechatronic systems that combine software and hardware.
In a software setting one way to solve this problem is to provide an interface (as in a Java interface, rather than a physical one). You end up with two concrete implementations of that interface, one for your real hardware, and one for your test version. Because the real and test versions provide the same interface, they should be interchangeable.
Once you've got your interfaces described how you implement them should be irrelevant (ie/ you could use a scripting language to develop the test code more quickly or cheaply) - so RPC may be a possibility, but there are certainly other choices.
I'm using a DSP to control a sensorless brushless DC motor, The DSP is on a board which has a parallel port and a jtag connection (it's an eZdspTMS320F2812). What would be the best way to communicate between a PC application and the DSP as it was running? Ideally I'd like to have a GUI program with buttons like start, stop, accelerate, decelerate... but I've never done anything like that before. Which ports and method would be easiest to use?
Thanks
You can also use simple RS232 communications. I use always because it`s cheap and easy to implement.
The RS232 transceivers are very cheap (like MAX232 from Maxim-IC), and easy to use. Also they come in many packages like DIP or SOIC for example and can be found almost every electronic shop.
You can use any USART from your microcontroller to link with MAX232. Then, using a PC serial-usb converter (or if your PC does have a serial port it`s easier), you can use serial port programming from any programming language to develop your desktop application.
After that, all you have to do is create a protocol to exchange data between your PC programm and your DSP (some simple commands to start, stop and change motor direction for example).
Good luck in your project.
The parallel port is probably the easiest route. Depending on what OS and programming language you are using you should be able to find example code or libraries to support bi-directional communication via the parallel port. Since you have a small set of commands that you might want to send to the DSP board then you can probably just send a single character to the board for each command, e.g. 'R' = start, 'S' = stop, etc.
I want to write a program from scratch to see the sockets activity, what they send, what they receive, etc. I don't want to use a Library because it's more for academic purposes than anything else.
Where should I start?
Just to be clear: my program won't be connecting to anything or creating any socket, it just wants to listen to the activity in a computer.
Any thoughts (in any OS) about where to start will be appreciated.
Thanks in advance.
Take a look at libpcap/WinPcap.
You are out of luck as far as portability goes.
Although the application APIs are the same or similar for UNIX and Windows, and, both implementations are historically based on the same Berkeley Sockets BSD code, the operating system architecture and hence the ways you access low level OS functions such as network IO are completely different.
Linux has a number of 'network sniffing' tools(tcpdump, ethereal etc.) easily available. Not so sure about windows, MS provides a Network Monitor and there are some tools available -- Google 'Sysinternals TDImon' for the MS tools.