Practical Project Deployment

[leblanc]

Hello,

I'm new to labview deployment.

I'm working on my senior project and noticed a team implementing their project completely with labview.

They plan on deploying their labview application to the client once complete. I assume the user needs to install a client similar to java's virtual machine. So their customer can run the application. Is this correct?

So does this mean the customer needs to buy the same daq equipment used by the development team? won't that be expensive? How much extra cost is this adding?

I always pictured labview as a test/development platform and never a production line item. The only place I can see this beneficial is in a large control system where it will greatly reduce the # of custom controls that need to be created in c# or qt. But most control systems i've seen are developed inhouse and not sold to a mass market. So is their solution viable for mass production?

From my experience I've used serial or com port to interface hardware with a pc. My senior project interfaces hardware with usb using a usb to uart bridge and c#. (my project is a transmission controller designed for gear heads that want to fully control shiftpoint of an automatic transmission. They use my gui to load/save user preferences and flash them to the controller.)

In my application controls don't even exist for what i need. Example i created a graph where the user can drag the lines on the y axis to modify default operation. I'm guessing this is where Measurement Studio for Visual Studio .NET seems to fit in. I can create my custom controls and use labview's dll for data aquisitions - that would normally be received through serial/com/usb .

Anyone with any thoughts? List of successful projects that integrated labview in the real world.

Leblanc Meneses

[Chris Davis]

I read this post and your other one, and I thought I would try to answer some of your questions...

QUOTE(leblanc @ Mar 14 2007, 03:27 PM)

They plan on deploying their labview application to the client once complete. I assume the user needs to install a client similar to java's virtual machine. So their customer can run the application. Is this correct?

Your question seems to be, "Does LabView require a run-time install?". The answer is yes. No different from any of the other higher level computer languages out there.

QUOTE

So does this mean the customer needs to buy the same daq equipment used by the development team? won't that be expensive, because only high end boards seem to be the only ones advertised with SignalExpress and Measurement Studio for Visual Studio .NET, or is it possible to buy a cheap terminal block and still use vs.net?

If this "team" is developing a piece of labview code that will interface to some hardware, then yes, any deployment, large or small scale will need to have that hardware as well. As for SignalExpress and / or Measurement Studio for VS.NET, they are two different products, designed for different users. Measurement Studio is designed to provide the VS.NET programmer an easier way to interface with the wide range of National Instruments hardware. SignalExpress is designed to help non-programmers interface to data acquisition devices and perform tasks.

QUOTE

I always pictured labview as a test/development platform and never a production line item. The only place I can see this beneficial is in a large control system where it will greatly reduce the # of custom controls that need to be created in c# or qt. But most control systems i've seen are developed inhouse and not sold to a mass market. So is their solution viable for mass production?

Its hard to determine what you are really asking here, since the sentences don't make much sense to me. I use LabView for just about everything I do, from automating my source code control's (subversion) backups, to high level image processing, to taking 10 megasample / second data displaying it to the user and allowing him to analysize it anyway he can think of. Controls that represent real-world type items, such as knobs, dials, tanks, and switches are just one of LabView's features. Is this "team's" solution viable for the mass market, I would say that it depends the market, and what it is willing to bear. I've seen presentations featuring Labview's ability to run a DARPA unmanned vehicle race, provide feedback from oil well's in remote locations, and power a LEGO robot to react like a scorpion.

QUOTE

From my experience I've used serial or com port to interface hardware with a pc. My senior project interfaces hardware with usb using a usb to uart bridge and c#. (my project is a transmission controller designed for gear heads that want to fully control shiftpoint of an automatic transmission. They use my gui to load/save user preferences and flash them to the controller.)

In my application controls don't even exist for what i need. Example i created a graph where the user can drag the lines on the y axis to modify default operation. I'm guessing this is where Measurement Studio for Visual Studio .NET seems to fit in. I can create my custom controls and use labview's dll for data aquisitions - that would normally be received through serial/com/usb .

Your project sounds interesting, if you are asking could it have been accomplished with LabView, the answer is probably yes. Your USB to UART / serial port bridge could be accessed using an API known as VISA. The GUI and all the code needed to successfully convey your information to the user could have been developed in LabView, and you could have recorded voltages and or current readings using a National Instruments daq device.

QUOTE

Anyone with any thoughts? List of successful projects that integrated labview in the real world.

There are a lot of successful projects that have been done with LabView that integrate into the real world, NI has a http://www.ni.com/news/webreply3.htm' target="_blank">good instrumentation newsletter that has new ones each issue. Check them out, one from this month had to do with an underground "excevator" that was powered with a compactRIO that communicated with a computer via fiberoptics.

[leblanc]

chrisdavis,

Thanks for sharing some of your insights.

I still need some things clarified.

Note: i'm moving some of the relavent information into my other post.

QUOTE(chrisdavis)

large or small scale will need to have that hardware as well.

Would it require the EXACT daq hardware? Example in lab i may have an expensive bnc daq board but in production i may be able to use screw type terminal blocks.

QUOTE(chrisdavis)

Is this "team's" solution viable for the mass market, I would say that it depends the market, and what it is willing to bear. I've seen presentations featuring Labview's ability to run a DARPA unmanned vehicle race, provide feedback from oil well's in remote locations, and power a LEGO robot to react like a scorpion.

Sound like interesting projects. I just didn't think a daq board as a customer deliverable, but i guess it can be. I guess I just approach the problem differently, most freescale microcontrollers have adc on them. I can see the development to deliverables shortening by using ni tools.

So here is my current belief, --[do you agree?]

For every item sold:

1) it will be more expensive

2) the packaging is larger

3) its brought to market the fastest

This is good to know, most of my work has been in small embedded projects. If i ever get a commercial account with a large budget and needs something developed asap, Ni with visual studio.net might be the best choice. Beats having to wire it all yourself with surface mount components.

QUOTE(chrisdavis)

Your USB to UART / serial port bridge could be accessed using an API known as VISA.

thanks for the tip, I'll take a look at it.

QUOTE(chrisdavis)

NI has a

good instrumentation newsletter

that has new ones each issue.

another great tip, thanks.