PDA

View Full Version : [Question(s)] Arbotix and nuke woes



Hanzuki
05-26-2012, 08:02 PM
So I have my new arbotix and I've gone through the nuke setup on my robot and that worked okay, however when I went to upload the program to my arbotix so I could try the test drive it gives me compiler errors. At first it said no such file "commander.h" so I moved copies of commander.h and commander.cpp into my robot folder and was able to move on. The next error was for no such file "Arduino.h". I moved about 20 copies of arduino.h into every folder and directory I could think of and it still won't compile.

I'm using arduino 0018 with all the libraries and arbotix and bioloid controller files and folders installed as per the getting started on the arbotix page.

lnxfergy
05-26-2012, 10:23 PM
Are you using the release zip? Where did commander.h come from (it should be from the robocontrollerlib zip, if you pulled anything from SVN trunk, it will not work with Arduino 0018, as trunk is aimed at an Arduino 1.0 release).

Bottom line: when using the released ZIP files with Arduino 0023 or older, you should never see "Arduino.h". Also, NUKE has not yet been fully tested with Arduino 1.0, so SVN trunk is probably not ideal for a walker.

-Fergs

Hanzuki
05-27-2012, 07:56 AM
I'll bet thats it then. I installed Arduino 1.0 first, then realized it wouldn't work so installed arduino 0018. Should I remove everything and start over?

Hanzuki
05-27-2012, 08:33 PM
Okay, good news and bad news. Did what fergs said and fixed my compiling errors. However after that I couldn't get arduino to find my avr isp on my pc. Spent a fair bit of time trying to figure it out and then tried on my laptop and got it working. Got my sketch uploaded to the arbotix and fired up test drive.

Now I'm getting an error that says "coxa failed xxx (xxx represents a number like 661)" and the turn and strafe functions just lift the legs but doesn't move the coxa servos. The tilt and rotate work okay. Any ideas?

shrub34
05-28-2012, 07:01 AM
You've hit the point I have with Nuke. The initial coax position is outside the min/max positions and is running into structure which is causing the servo to turn off due to error(s). I'm going to spend time debugging/reading through the generated code to see how to correct my current issue of 3 of the coax are eating the body.

Once I have a solution/recommendation I'll let you know what I find. My quick guess is that the positions are being fed to the wrong servo.

Hanzuki
05-28-2012, 12:19 PM
I did that the first time through. I had some of the neutral points at the extremes. In this case my neutral pose is 512 on all 4 coxa servos and none of them can move enough to hit the body and none of them have gone into failsafe. Both the min and max values are valid also. This is where I am confused.

Hanzuki
05-28-2012, 01:44 PM
Got my coxa errors figured out. I did my min and max setups on the safe side so as not to go beyond the servos travel but it was insufficient for ik engine. Edited the nuke.cpp file to change the min and max settings to closer to the max travel of the servo and no more errors.

Other question though, I have a power supply that outputs 13.8v constant. Think thats pushing the ax12's too much? I killed a lipo with all my tinkering by leaving it plugged in for too long.

tician
05-29-2012, 10:29 AM
Robotis tends to be quite strict on the 12V limit when advising on their QnA site, so I'm guessing the PWM values used to drive the H-Bridge could allow sufficient current to damage the motor and/or MOSFET driver when run at much more than 12V (there is no current or torque sensing to protect the servo's motor or driver).

Just as a reference though, the voltage regulator used in the AX-12+ appears to be capable of up to 30V, and the MOSFET driver should also safe to 30V for both the P-channel and N-channel halves of the half bridge as long as the current remains below 3.5A.

jwatte
05-29-2012, 02:35 PM
Note that rated voltage/current on components assume a certain thermal solution (often, fins with forced air.)
Once you put the components inside a small box with only the box itself as the heat sink, the specifications of the raw components may de-rate significantly.
This of course depends on how hard you push the components mechanically, too. If it's just dangling in air with no load, the load is different from when it's doing jumping jacks :-)