We’re trying to connect the Classmate, Dev Laptop, and Robot so we can get feedback from our sensors in WindRiver C++.
We’ve connected it with a setup like this: dev laptop and classmate to router, which wirelessly connects to the robot. We’ve also tried hooking everything up directly to an ethernet switch. Still don’t work.
When we try to connect with the console, the Classmate intermittently displays “Watchdog Not Fed”. The robot reaction works, but is delayed when we try the drives. The output on the console has all 0 values. Without the Dev Laptop in this network, the robot works fine.
We fed, disabled, and even tried to kill Watchdog. We’ve tried different laptops. It still does not work.
The Watchdog not fed message means that the running program enabled the watchdog timer and is not calling the Feed() method before the expiration time runs out. If you have autonomous code that does a long Wait() you are probably running out the watchdog timer. If you have a program loop in your teleop code then you might be missing a call to Feed().
The configuration of the programming laptop and driver station shouldn’t matter if they are all successfully talking to the robot.
Are you running some sample code or your own program?
We are using our own code. There are no loops in our teleop, and we haven’t used the autonomous yet. We had GetWatchdog().Feed() 's all over the teleop when we were trying to feed the watchdog.
That sets the expiration time of the watchdog timer. That means that it must be fed within 0.1 second (100 milliseconds) before it times out and the motors stop with the “Watchdog not fed” error message.
If there are any Wait() calls in your loop that are greater than 100ms, the the watchtog will also time out.
You could use GetWatchdog().SetEnabled(0) for turning it off right before your long wait and then use GetWatchdog().SetEnabled(1) for turning it back on again when you’re not initializing whatever you’re initializing (i’m assuming your long wait is due to setup of some component)
We tried setting the Watchdog at SetEnabled(0), but still get the same problem. We set it at 0 in the initializers, in autonomous, in the operator function, and in the operator loop. Setting SetExpiration() at an impossibly high number had no success, either.
We tried instigating the Watchdog error while watching the serial output, and found something very interesting.
We could get output from the serial output (meaning that we got the printf values from the code in the serial output) and the sensors were working fine! Outputting data! But once we turned on the target console on the programming laptop (the programming laptop and the serial output laptop are separate laptops) everything STOPPED. Came to a STANDSTILL. No more output, nothing. The serial output stream seemed to FREEZE. That may be why the dog was barking.
But the data we received from the sensors are somewhat strange. The different printf statements mixed and mashed with each other, seeming to overlap. I can’t make anything of the data because they overlap so often.
did you have
at the end of the printf’s? this is the new line sequence. if you have printf(“part 1”) (more code here) printf("part 2
"), it should output like this:
part 1 part 2
part 1 part 2
part 1 part 2
part 1 part 2
part 1 part 2
part 1 part 2
part 1 part 2
part 1 part 2
part 1 part 2
so you can see 1 and 2 as the numbers fly by instead of
part 1 part 2part 1 part 2part 1 part 2pa
rt 1 part 2part 1 part 2part 1 part 2part
1 part 2part 1 part 2part 1 part 2part 1
part 2part 1 part 2part 1 part 2part 1 pa
rt 2part 1 part 2part 1 part 2part 1 part
2part 1 part 2part 1 part 2part 1 part 2p
art 1 part 2part 1 part 2part 1 part 2par
t 1 part 2part 1 part 2part 1 part 2part
1 part 2part 1 part 2part 1 part 2part 1
part 2part 1 part 2part 1 part 2
We set it at 0 in the initializers, in autonomous, in the operator function, and in the operator loop. Setting SetExpiration() at an impossibly high number had no success, either.