Intermittent CODE ERROR, followed by uncommanded relay switching

[tcj103]

We've encountered an intermittent problem with this years robot controller. We have the 2007 streamlined camera code on the robot, as provided on Kevin Watsons web site. Once in a while, we'll power up the robot, which will set a "code error" on the controller. Then, if you hit the reset on the controller, everything seems to reset OK, the code error light goes out, but then a spike relay will start cycling at about 1 Hz, sounding like an old car turn signal. Turning off the robot, and starting back up, sometimes will clear the condition, but most times it wouldn't. Moving the joystick all the way to one side (I think to the left) makes the relay stop. I'm not sure if the trigger still operates the relay (we modified the Kevin.org code for relay controls and some PWM control shaping)

The other symptom during this time that prior to the reset, the main battery voltage indicates OK on the controller (over 12 VDC). But when the relay is clicking, the battery power display shows about 8.2 VDC, even though during that condition, we metered the battery terminals, as well as the power going into the controller, and got over 12 VDC.

So, can anyone explain the incorrect battery voltage? I'm looking through IFI reference documents, but haven't found anything yet. We have since reloaded the streamlined code, with fewer changes, and are waiting to see what happens. To my knowledge, we're using the firmware on the controller that came with it, no updates have been loaded that I know of.

Also, I found one explanation about code error, is there a more detailed description of it anywhere?


Thanks
Ted

[Goldeye]

Are you using timers or interrupts in any unusual or extensive way?

Also, the 8.2-8.3v problem is one that showed up last year and everyone hoped was gone this year. Guess not.
Check out http://www.ifirobotics.com/docs/memory_problem_8722.pdf

[tcj103]

Thanks for the tip!

We are not using any interupts or timers, beyond what the camera code uses from Kevin Watson.

The 8.2-8.3 V problem sounds more like firmware/hardware. Can operator software cause this problem?

I'll check to see if they've complied with the memory problem update.

[ericand]

The key to soliving the problem that was showing up last year (with symptoms similar to yours) was the change in the linker script. The 8.2-8.3 problem was found by our team to be temperature sensitive (occuring more often when the processor was cold, and less when it was warm).

The problem seemed to occur when the memory used by txdata and rxdata (the memory used to communicate between the master controller and the slave controller) was in the memory section adjacent to the hardware registers. The linker script patch makes that piece of RAM protected so it is not used (and the memory used for communication is moved away from the hardware registers).

[Kevin Watson]

Originally Posted by tcj103

We've encountered an intermittent problem with this years robot controller. We have the 2007 streamlined camera code on the robot, as provided on Kevin Watsons web site. Once in a while, we'll power up the robot, which will set a "code error" on the controller. Then, if you hit the reset on the controller, everything seems to reset OK, the code error light goes out, but then a spike relay will start cycling at about 1 Hz, sounding like an old car turn signal. Turning off the robot, and starting back up, sometimes will clear the condition, but most times it wouldn't. Moving the joystick all the way to one side (I think to the left) makes the relay stop. I'm not sure if the trigger still operates the relay (we modified the Kevin.org code for relay controls and some PWM control shaping)

The other symptom during this time that prior to the reset, the main battery voltage indicates OK on the controller (over 12 VDC). But when the relay is clicking, the battery power display shows about 8.2 VDC, even though during that condition, we metered the battery terminals, as well as the power going into the controller, and got over 12 VDC.

So, can anyone explain the incorrect battery voltage? I'm looking through IFI reference documents, but haven't found anything yet. We have since reloaded the streamlined code, with fewer changes, and are waiting to see what happens. To my knowledge, we're using the firmware on the controller that came with it, no updates have been loaded that I know of.

Also, I found one explanation about code error, is there a more detailed description of it anywhere?


Thanks
Ted

As others have mentioned, this is a nasty bug that some folks have had problems with. Grab a copy of the 2.1 camera code and extract the *lkr and *.lib files into your build directory. Make sure you use the new library when you re-build.

-Kevin