NI Help

Since I work a lot with different instruments and don't have the luxury of completely learning the ins-and-out of one product, I rely on a companies help functions to get my job done.

Lately I've been integrating some Agilent RF equipment into a test set. I now have a much better appreciation of NI help. I've always felt like NI has some of the best technical help around, examples, and documentation but I've come to appreciate it even more.

I'm trying to use TestStand to control an automated test on an RF Signal Generator, RF Power Meter, and an RF Switch, all from Agilent but having trouble. Agilent really is trying to help but still have a really long way to go to get close to NI help. When I had problems I entered a couple of on-line help requests but received no responses. I had to call the local sales rep to get the e-mail of a tech rep. I have been e-mailing him and he has been helpful in a limited way. I did get the manual web sequence to control the RF Switch, but for automated tests, that doesn't mater.

The RF signal generator was easy because I've used it before and figured out it's quirks. The examples for the RF Power meter had nothing to do with actually reading power, hmmm. When I made calls to the driver functions, they came back with error messages that gave no hint at what was needed. The funcion panel help gave no hint at what the parameters were looking for. I wound up using some low level SCPI commands in combination with the driver functions. I couldn't get it to work with all SCPI or all driver functions.

As for the RF Switch, the examples seem to include calls to driver functions for many different switch types. For my RF switch, the driver functions generate a lot of "Not supported" error messages. But it was very hard to figure out what is supported or what calls I could make. Also, once I loaded the Agilent IO Drivers (which are required for the system to even see their switch) the Pickering switches disappeared from the PXI chassis, at least from a logical standpoint. I still haven't figured that one out.

It's been a week and a half trying to get the RF interconnections to work. I think I'm close but compared to the NI instruments, it's taken way to long.

Using Agilent instruments has given me more of an appreciation for how good NI help, discussion boards, phone help, and e-mail help really are. I especially want to thank Joel Garner, NI Sales Engineer extraordinaire, for his help.

Misunderstood

I've found that the Test Engineering group is misunderstood, at least where I work. Most of the engineering groups have an identity. The Electrical Engineering group builds the hardware, the Mechanical Engineers builds the moving parts, Software Engineers write the software to control the hardware, and the Systems Engineers glue it all together.

It seems what Test Engineering does is a mystery to the other groups. We are called in late in projects because Project Managers don't know why they need us until they realize they need to make more than one widget (or whatever it is they're making). Then they realize their widget (or whatever it is they're building) was designed in a way that makes it incredibly hard to test.

It's too late but that's when they figure out they needed us to begin with.

We're trying to educate the programs on what we can do for them. We can do is:
- We do hardware and we do software.
- We develop systems (Test Systems)
- We can help them design their widget so it can be tested.
- We can make their tests automatic and repeatable.
- We can help with developmental testing
- We can make sure the widget is built correctly
- We can test your system in the field or on the production line

...Bottom line is we can help.

CVI Rules!

And LabVIEW Drools. It had to be said.

I finally got away from doing documentation and processes for a day, so I was able to do some fun stuff...Technical work. I'm currently working on a Test Station Self Test. Some parts of the Self Test are done in LabVIEW, partially because others did those parts with LabVIEW, partially because I broke down and wrote some LabVIEW code.

Well, I'm updating all the LabVIEW code to fit it into TestStand for the Self Test. It's frustrating, hard to find the vi's I need and it does things for me...I'm not sure what...that I don't think I want it to. For example, I opened one of the vi's to see the picture on one of the DAQmx self test vi it contained (in order to find it's equivalent for NI Sync) with out changing a thing. When I closed the vi, with no changes, it asked me if I wanted to save. I didn't change anything! What was different where it needed to change?

As I was working in LabVIEW, it took quite a while to update the code, but I was learning. I was figuring out what the microscopic pictures were on the vi's, although I still have no idea why they chose the pictures they did for some of them. Things were moving along smoothly with only a low constant pain in my wrist. I was finally done with the LabVIEW part.

Then I needed to develop some CVI instrument wrappers. And BAM! They were done! The text based code just flowed from my finger tips. I plugged it into TestStand and BAM BAM! It worked...first try! And I wasn't even entirely sure how the RF Signal Generator worked. Compared to the struggle I had getting the LabVIEW stuff out, it was a breeze. It was like night and day! It was slicker than deer guts on a door knob! (A good old southern saying)

All I could say was CVI Rules! and LabVIEW drools!

Music and Teams

I was hanging out at the book store, as I tend to do since they have almost all that is needed in life; Coffee, lots of books to read, and free wifi. Some how, I accidentally wander out of the Computer book section, somehow missed the Science and Engineering section and wound up in the music book section. I browsed a book about Zappa for a while but stumbled a book call "Comfortably Numb: the inside story of Pink Floyd". It's a great book!

There were several points that seemed counter intuitive about teams until you realize that Pink Floyd produced many great albums (for the young people, an album is a vinyl disc that music was recorded on and played on a record player...with a needle) including The Dark side of the Moon, one of the best selling albums of all time.

First point, hiring people based on best team fit is over rated. Sometimes the band members were at each others throats. The point was they were all smart, all creative, and they had the drive to get the job done. They didn't let creative differences or personal difference get in the way of there goal.

Second point, some teams fall apart with dissension between team members, some teams refocus their experiences into creativity. Pink Floyd re-focused into writing great music while software teams can refocus into great software.

Third point, Creative people push themselves, and others, hard. Creative people will push themselves to bring their idea's to the forefront and sometimes will push others to get their idea's out there. They seem to push each other, sometimes in not so nice ways, to get people to complete there own ideas.

Fourth point, don't let technology overwhelm the ideas. This one may be a stretch but it has a good point. With Pink Floyd all the technology behind the shows sometimes got in the way of the music. Today, all the e-mail, IM, blogs, distract us from reaching the goal of developing our ideas.

Overall, the parts that I read were very enjoyable. I'm going to have to go back to the bookstore, hangout some more, and read the book. Or, hey, I could actually buy the book! We'll have to see.

LabVIEWism

-ism: Docturine; Theory; System of principles; Distinctive or Character trait.

I've noticed some traits of LabVIEW programmers and the word "LabVIEWism" came to mind. Sort of like Hinduism or materialism. It seems to be more than just a programming language, it's seems to be a way of life...programming life that is. Similar to Protestantism is to Protestants or Communism is to Communist. (No association with LabVIEW implied)

To me, it seems that LabVIEW is ingrained in the LabVIEWist being, it flows in their blood. You cut them and they bleed hemoglobin vi's. (I'm sure the icons are red)

LabVIEWism has a lot of followers that are dedicated to the LabVIEW way of programming, focused on converting us lowly CVI programmers. NI Week is a tribute to LabVIEWism, they inundate everyone with LabVIEW with only a token acknowledgment to CVI.

I don't bleed vi's, or dream graphical code. I still dream in C and preach the virtues the code editor and the power of the command line. LabVIEWism may take over our companies test group, but there will always be holdouts.

LabVIEW...

I have been making a concerted effort to use LabVIEW lately, trying to learn the "Graphical" way of programing. Due to Carpel Tunnel Syndrome I use my left hand to mouse a lot. I've tried the, so called, "Keyboard" shortcuts ("Keyboard" as defined by NI), but since they still involve the mouse, they aren't as helpful as I need them to be.

But continuing to work with the pain, my left hand is just not as dexterous as my right. It's hard to hit the small wire Connections. It's annoying trying to figure out which vi to use for what functions.

Some of the features that would be VERY helpful to get us LabVIEW handicapped people working in LabVIEW is:
- a zoom feature. I could make a vi bigger and actually hit the terminal.
- Or have it where, when I miss hitting a terminal, I can use tab or shift tab to move the connection from the current terminal to the next or previous terminal.
- Have LabVIEW be able to select the end of a wire and then drag it around using the arrow keys instead of the mouse.
- Have the pallet navigable with the arrow and tab keys. Make it where you can select the pallet, then with the arrow and tab keys move about in the pallets to the vi you want. From there, be able to select it and drag and drop with the arrow keys.

Have the LabVIEW developers ever heard about Test Driven Development (TDD) or automated Builds? Here's some more features.
- For automatic builds, have VI's able to be Compiled (or whatever happens to them) from the command line and verified none of them are broken (with the broken arrow for the run button)
- For Test Driven Development, have an automated test frame work. Frame works similar to the NUnit test development suite.

One more suggestion...have a cheaper copy for use by anyone. MS have copies of things like Visual Studio C++ for $100 or so at computer stores.

It's still easier to think in C (LabWindows CVI) than it is in LabView and until I can develop LabVIEW as I can develop C code, I'm sure I'll still be using CVI. But I will always keep trying to learn new stuff.

Now as soon as NI comes up comes up with telepathic programming interface, I'll be there!

OK, Now that we know the Super Intelligent Shades of the Color Green are going to take over the world, lets get on with less important things... like testing.

I've been working on processes and metrics for developing test sets for quite a while now (a couple of years). I love doing technical work so it has been tough doing mainly documentation. I have worked on some test programs part time while working on the processes, so it hasn't been totally tortuous.

I believe there is always a process of one type or another in developing test stations. The problem is some times the process is the POOMA (Pull Out Of My A**) process or maybe the "wing it" process. But then there's the other end of the process universe. This is where the process makes it nearly impossible to do a small project. Then, in the middle, there's the agile process which is suppose to be the processes that allow tasks to be done quickly and with the ability to easily adapt to changes. It's still not a cure all.

Even people who say they have no process, have a process, it's just not as straight forward as most or even seems coherent.

After working on the processes for a while I know that one process does not fit all projects even though that's what we try to do. I was suppose to document the process we go through to build test stations. Each project has been unique in it's own way, with some process steps that aren't in the process and skipping other steps that are in the process. Some projects that try to follow the process finds that outside groups don't follow the process which makes our process fail. There's many reason's process fail.

If the processes aren't working (not that people just don't want to do it), they need to be adjusted , in real-time rather than going through the year long process change process.

The project leads, the one's actually have to make the process happen, have to be the ones to adjust the process as needed. They need the lead way to adjust it as needed. The Process police also need to understand a single written process is not a panacea, it won't cure all. The processes need to be adjusted.

While I've been working on the processes, this is the dilemma I've been dealing with, having processes that don't imped getting things done but satisfies the Process Police.