Labview Watchdog issues

[jhersh]

Originally Posted by isaacdl

We are using labview and our robot is up on blocks. We started a default project that uses the arcade settings. Teleoperated works fine but when we put code in the autonomous iterative vi it goes into watchdog whenever autonomus is enabled. We have tried feeding the watchdog as well as just killing it but to no avail. Suggestions on how to make the watchdog happy.

Make sure you haven't overlooked the false constant on the autonomous independent VI that the template generates. It disables all autonomous code by default. That way you don't accidentally run autonomous.

[BLProgram2010]

I'm having a problem with Watchdog. Two days ago I try to run code that was working just fine but every time I hit enable on the dashboard there is a message of Watchdog not feed, and joystick won't work, The robot won't respon. I read that their was a 'bug' in NI labview, what are they talking about? How can I fix the Watchdog problem?

[Greg McKaskle]

The bug related to watchdogs is that the driver station would occasionally be late sending the control packet and the cRIO now prints messages to log the watchdog glitches, no matter how short they are.

The watchdog not being fed message tends to occur when the watchdog is enabled and actually not fed.

Greg McKaskle

[BLProgram2010]

Would deleting the watchdog completely from the Labview project help any? how do I go about repairing the problem (stop the watchdog glitches)?

[Greg McKaskle]

You can disable the watchdog or you can delete the watchdog VIs. However, you may be better off determining why the watchdog is firing.

Greg McKaskle

[Vikesrock]

Originally Posted by jhersh

Make sure you haven't overlooked the false constant on the autonomous independent VI that the template generates. It disables all autonomous code by default. That way you don't accidentally run autonomous.

I am trying to help someone on another team debug a watchdog error on entering autonomous, and I cannot locate this false constant you are referring to.

I have looked in Begin.VI where the Auto mode is set, in Robot Main in the Autonomous Enable case and in Autonomous Independent.VI itself and have not been able to find a false constant that looks like it would disable autonomous code. I even checked Autonomous Iterative just to be sure.

Any hints on where I should be looking for this?

[Joe Ross]

Originally Posted by Vikesrock

I am trying to help someone on another team debug a watchdog error on entering autonomous, and I cannot locate this false constant you are referring to.

I have looked in Begin.VI where the Auto mode is set, in Robot Main in the Autonomous Enable case and in Autonomous Independent.VI itself and have not been able to find a false constant that looks like it would disable autonomous code. I even checked Autonomous Iterative just to be sure.

Any hints on where I should be looking for this?

I think Joe was referring to the constant that was in the code last year, and during the beta this year. Since Autonomous was completely rewritten with the tracking code, it looks like that enable constant was removed and doesn't apply anymore.

[Vikesrock]

Originally Posted by Joe Ross

I think Joe was referring to the constant that was in the code last year, and during the beta this year. Since Autonomous was completely rewritten with the tracking code, it looks like that enable constant was removed and doesn't apply anymore.

Alright that's good and bad news I guess. The good news is that I'm not blind or crazy, the bad news is the watchdog error still eludes me (I'm not a fan of remote debugging)

[Greg McKaskle]

I'm not a fan of remote debugging

Tell me about it.

Perhaps it would help to verify in Begin whether they are using the Auto Independent or Auto Iterative. I'd suggest the independent.

Open the VI and reproduce the issue. You can put down probes, turn on exec hilighting, or pause from time to time to see where the code is. Is there a loop that doesn't feed or delay & feed?

If you get it down to that VI, post it and I suspect someone will spot the issue.

Finally, I did run into teams who simply press the Run button on auto and were surprised it didn't work. Make sure they are running auto by running main and putting the DS into auto mode.

Greg McKaskle

[Vikesrock]

Tracking down the watchdog error isn't actually a big deal as it isn't their code causing the issue.

They put some code in Iterative, but Begin still pointed to Independent. Independent still had the default code in it so it appears that it was the default code triggering the watchdog error. While this isn't a huge deal, I am curious as to what is happening.

Taking a look at that code it seems to me like if there is no target on screen then the loop will exit the first time through (Rotate to target tries to turn to Gyro=0 and is already there) and the code will move to the arcade drive command with the 0 input.

At this point the watchdog will time out right?

[Greg McKaskle]

Technically, yes. If their auto finishes, they've said they don't want to do anything else for auto and the robot will spend that time disabled due to no watchdog.

Greg McKaskle

[kws4000]

what would happen if you deleted the watchdog vi's, like Greg McKaskle suggested?

[jhersh]

Originally Posted by kws4000

what would happen if you deleted the watchdog vi's, like Greg McKaskle suggested?

As long as the user watchdog is disabled, then you simply reduce the safety of your robot. It if happend to be enabled, then your robot will not work. I recommend at least opening it and disabling it so that you are sure that it is not dead.