Robot will Run but Not deploy

[663.keith]

I know there have been other post about being unable to deploy, and I have read those posts. We have been able to deploy our code as a startup application in the past, but today it seemed to no longer work. When we select "deploy all", it compiles the code on to the robot, just like in previous times, but now when we restart our DS battery status reads "no code" and our robot does nothing. We feel pretty stuck, is there anything that we can do that can verify the code is working; or is there any other procedure we can do to run the code on startup?

[airnate]

I imagine that you have already done this since you indicate that previous code deployed okay, but just in case...

Make sure that you wait for a while after the reboot. It will typically take 30 seconds or more for the cRIO to boot and get the code up and running. So if it's not working, the first thing I would try is to wait a couple of minutes after booting to make sure it is not really already deployed.

[663.keith]

Originally Posted by airnate

I imagine that you have already done this since you indicate that previous code deployed okay, but just in case...

Make sure that you wait for a while after the reboot. It will typically take 30 seconds or more for the cRIO to boot and get the code up and running. So if it's not working, the first thing I would try is to wait a couple of minutes after booting to make sure it is not really already deployed.

we did wait at least 30 seconds for the "no code" to switch to the battery voltage, but probably not much longer than a minute. Would there be any reason for an especially long delay during startup? also would a re-image of the cRIO be any help at all?

[Greg McKaskle]

Be sure to not only deploy, but to set the code as startup app. You can also choose Run as Startup and it will mark the code, then reboot for you.

Greg McKaskle

[Hatefulhippo]

We are having the same problem. How do you set the code as a start up app.?

[airnate]

Right click on your build specification and select Set as startup (if it says Unset at startup, you've already set up as startup).

For more details look in the FRC Programming Guide, in the section entitled Creating a Stand-Alone FRC Application on p. 5-5 (p 61 in pdf).

[Hatefulhippo]

We are still having problems... our code deploys but the robot wont run weve tried the set as startup and the run as startup and yet nothing.... im begining to feel hopeless and recently instead of enable we have been getting a Watchdog instead.. does this have to do with anything

[airnate]

So are you saying:

1. You have code that operates correctly (e.g. the robot is responsive) when you press the Run button (i.e. you are running interactively with LabVIEW code on your laptop).

2. That you created a build specification using the same code and did a "Build" with it.

3. That you deployed this build specification as a stand-alone executable and after the cRIO boots, the robot is unresponsive?

[theCoder]

We are having the same problem. Yesterday we updated the cRio and Driver code to the most current versions from FIRST. We are at a scrimmage match and we toggle 4 seconds on "Watchdog" then one second on "Enable". We get this problem when using the scrimmage SSID as well as stand-alone. We have followed the update instructions very carefully, at least 4 or 5 times. No luck. Is anyone else having problems with the new code? Any suggestions would be greatly appreciated.

<< edit >>
We found the problem. Some code that worked with the older firmware caused problems running with the new firmware. One problem we figured out, the other went away when we commented out a large block of code. The first involved the compressor. We were directly controlling the spike, instead of using the compressor object. This code worked fine before, but with the new firmware caused the cRio to die with a "kernel" error. It took the better part of 4 hours to find. Unfortunately the compiler did not catch the problem. The second was buried somewhere in a large block of under-tested autonomous code. Yesterday at the scrimmage was the first time that code was fully exercised. As mentioned above, the controller was between watchdog and enabled. Commenting out a large block of code made the problem go away.