Arduino and LabVIEW

[Andy A.]

Originally Posted by EricVanWyk

I'm not sure what David is talking about...

You can use I2C, SPI, or serial. Or, you could add an ethernet shield to the Arduino and use that.

For I2C and SPI, the cRIO wants to be the master. I'm not sure if the arduino easily supports slaving in these modes, you will have to check it out.

For serial, the Uno consumes its native serial port with its USB connection. Instead, you will have to use the software serial port library.

There is also a CAN bus shield available, although I know little about it. It has a DB9 connector, which I guess means it's really intended to interface with automotive diagnostics stuff (complete with a sd card slot for data logging!). Still, I'm sure a clever team could handle the wiring and code required to interface it on a FIRST bot.

There are, of course, legal issues regarding custom circuits being in the CAN bus loop. Without carefully looking I think using the bus just to pass data is ok, as long as it doesn't try to control motors. I'm not sure exactly what practical advantage it would give you over SPI or I2C, but it would be pretty neat.

CAN bus shield on Sparkfun. It might be avalible elsewhere.

[WizenedEE]

Custom Circuits are explicatively allowed to be on the CAN bus. They are also explicatively illegal to control any non-decorations (and I'm not sure about the rules on decorations). Unless, of course, you email FIRST and get something as an approved motor controller.

I'd most likely try to communicate using I2C, but if you can get the C-RIO to send out serial commands, that's the arduino's native protocall.

[Roger]

Well, we might be but not all that clever, especially when we're short on time. I'll have to check those possibilities, Eric. Where does the data wire(s) get connected? Right now I'm in the "one step back" mode (computer doesn't want to download to the Arduino), so it'll be a bit.

All I need is to send a single number to the Arduino, and it'll turn on some LEDs. Can I do it through a Digital IO (which sounds the simpliest method)? WPI_PWMSetValue.vi?

Just saw Adam's post: Yeah, it's not quite decorative LEDs, so thats out. Any of the LabVIEW I2C examples are good to look at? Or the serial?

[Mark McLeod]

How big a number?

You can simply use the digital outputs on a sidecar to digital inputs on the Arduino to send a binary number.
You don't need a communications protocol to do that.

Maybe use one DIO to signal when the number has changed and should be read by the Arduino.

[Roger]

Do you mean (for example) 4 DIOs sending a binary signal for 16 numbers? (And no, I don't have many numbers, but I can think of ways to expand that.) Just using the signal portion of the PWM wire, right? (The Keeper of the Numbers says we have 5 DIOs free for use.) Heh -- I had thought of that too, when we originally thought of using DIOs. It seemed too obvious to be true. Could it really be that simple? Thanks, Mark!

Of course, I spent the day trying to make connections on an older robot to test all this, and now discovered (after connecting all the dots, so to speak) that LabVIEW and the Arduino programming port don't play nice together. I am making a nice collection of laptops around me just to test this.

[kamocat]

Originally Posted by Roger

Do you mean (for example) 4 DIOs sending a binary signal for 16 numbers? (And no, I don't have many numbers, but I can think of ways to expand that.) Just using the signal portion of the PWM wire, right? (The Keeper of the Numbers says we have 5 DIOs free for use.) Heh -- I had thought of that too, when we originally thought of using DIOs. It seemed too obvious to be true. Could it really be that simple? Thanks, Mark!

Of course, I spent the day trying to make connections on an older robot to test all this, and now discovered (after connecting all the dots, so to speak) that LabVIEW and the Arduino programming port don't play nice together. I am making a nice collection of laptops around me just to test this.

Yes, if you're not doing a high volume of data, this is an easy and viable method.

[Roger]

Well we are talking about an Arduino. But no, just a signal to change lights.

Howsoever, I did see at FLR a team using an Arduino for sonar processing, but I don't know what kind of data they were sending back to the cRIO -- much more than I'm asking for, I'm sure. These little boards are fun to work with.

[Roger]

An update -- I have some LabVIEW code that sends a DIO true/false value depending on a button press, recieved by the Arduino on a digital pin, and blinks the light on the Arduino and also prints out the high/low value (simple test), slowed down to 1/2 second cycle. Currently I'm getting a semi-random highs/lows, even without pressing the button, which I'm guessing is LV's issue, not continually sending the true/false signal? Unfortunately, I lost my LV computer so I can't check it -- maybe I can schedule for a 3 am session?

[WizenedEE]

If you haven't opened the IO port in your code, I think it may leave the signal pin floating. Does it work when you press the button, or does it ignore the button as well?

[Roger]

I think it's opened once in the Begin.vi (the computer got grabbed so I can't look at the code at the moment), and just referenced during the button pressing, and not open/send/closed each time.

There was no discernable difference when the button got pressed.

[davidthefat]

Consider the fact that a boolean on the PPC is 4 bytes. Also consider the fact that the Arduino uses little endianness. I am not sure about the PPC on the cRio, it might be a big endian, that might bring up some troubles.

[Roger]

But aren't they both using an on/off electical signal to talk between them? Not code.

(Pay Per Click?)

[davidthefat]

Originally Posted by Roger

But aren't they both using an on/off electical signal to talk between them? Not code.

(Pay Per Click?)

Well, Boolean only takes up 1 byte (8 bits) in Arduino, the PPC uses 4 Bytes instead (the PPC registers are 32bits). Sending over a boolean from the PPC sends 32 bits of data. Now, considering that you can't fit 32 bits in a 8 bit register, you have to cut the extra data off using some bit manipulation. And the other problem I was pointing to was that the PPC usually are Big endian, but they can be bi (so it can be both, just not at the same time) that means the bits read from left to right, but the little endian reads right to left. So you need to do some switcharoos there.

So on the cRio a "true" would be represented as:
0000_0000_0000_0001B (if big endian)
0001_0000_0000_0000B (If little endian)
Note: I might be wrong about the above regarding the binary

on the Arduino, you need to just take the last 4 bits so it becomes:
0001B

Every microprocessors have millions and billions of transistors, they all use electrical signals.


Now, if there are any veterans here, feel free to rebuke me. I am a student, I make mistakes and then learn from them.

[Roger]

It was this that got a train of thought starting:

Originally Posted by Mark McLeod

Maybe use one DIO to signal when the number has changed and should be read by the Arduino.

Have one of the four lines be a "control" bit: start with 0 (no info forthcoming), switch to 1 for first group of data (the 3 other lines), pause, switch to 0 for second group of data, pause, switch to 1 for third group, pause, switch to 0 to end. (Obviously you could extend or contract this, as long as the last group is a 1, so to loop back to 0.) I could get 9 bits of info, but that didn't seem right. I may make that last bit a parity or check bit "some day". The "pause" would be long enough to read the pins and do one loop of code, but I was trying not to be too cross-dependent with timing. Even at 1/4 second per group, it would be only a second to transmit the info.

I had scribbled the Arduino and LabVIEW code in bed last night, and got half way thru it by the time I got back to the robot, but my lead programmer had different ideas (a piddly 16 numbers - pft! ), so we went his way. Still haven't tested that it really works, either way.

[RoboMaster]

Originally Posted by Roger

Currently I'm getting a semi-random highs/lows, even without pressing the button...

This sounds VERY much like your pin is floating. This is not really a programming issue, but a wiring/electrical issue.

Watch this video, note why he uses the resistor. If you get it, good, otherwise, I'll explain:
Originally Posted by Roger
But aren't they both using an on/off electical signal to talk between them? Not code.
Yes and no. The cRIO will send out an on/off electrical signal, but the Arduino DIO pin does not "sense" "on or off." It senses 5v or Ground. If you hook it up to nothing (what you may call "off"), it will basically sense the voltage of the wire/air that it is connected to. Ever looked at a volt meter when you haven't yet hooked it up to a power source? It wanders all over the place. This is what your Arduino DIO pin is doing. It could be anything, high or low. You need to use a resistor to specifically connect the DIO pin to Ground when it isn't sensing a 5v from the cRIO DIO pin. This is called using a pull-down resistor (not to be confused with a pull-up resistor, but similar idea).

Do you have a mentor who is an electrical engineer or knows electronics? You should have him help you out. This concept is a little tricky, it took me a while to figure out. Before that I was thinking the same as you.