Monday, August 12, 2013

Yay progress!

   Huzzah! I swear, after figuring out the snafu with the arduino loaders (hint: use 1.0.4 if you can for now), everything seems to be running smoothly and quickly. It's 3:37am EST by my phone's clock, and despite my other job looming right around the corner, I managed to get the software issue fully resolved. Yes, the code is completely finished as far as I am concerned. I do however need some extra hardware as the tilt servos have taken a bit of a beating since I started this project. Back to ServoCity to get two more servo-blocks. If they can't take the downward strain off the tilt servos... Not sure what could. In any case, each servo is rated to carry the full weight of the camera on its own, so having two just eases the load. Hopefully the servo blocks will help that much more. After that, I'm pretty much can say 1.0 is fully born.
   I do have a debate broiling in my head though, and it's whether or not to mount the arduino on the slider dolly itself, or to run Ethernet between the arduino/control box and the dolly. Basically works like this: eight wires in a cat5/5e cable. As such, I divide up the wires to certain outputs. Signal line for each of four servos, motor positive, motor negative, and the 12v positive and negative. The reason for 12v is for two reasons: one is that higher voltage can push farther down a wire easier than lower voltage. The second it that the 5v regulator I built is currently mounted on the dolly with no issue, and I don't want to have to move it. That regulator powers the four servos. So yeah, basically I'm too lazy to move it. Besides, the heat sink on it will get more cooling airflow this way (and it needs it, I nearly burnt my finger touching it once!). So yeah, I'll keep you all updated in my overly verbose blog on how this is going. Have fun, and see you next time!

Sean Watkins
Apokalypsis: (Greek) to unveil, to make known.

Saturday, August 10, 2013

The Slider Dolly

   Progress has been much slower that I would have liked on this Arduino monstrosity. Since last time I posted (yes, that long ago!) I have tried several controller ideas, including a homemade box with several potentiometers and rocker switches (anyone have 4 hands?), a Wii classic controller, a knock off PS3 controller, a Wii nunchuck, and I even tried an Arduino Esplora. None of them would work, and for no discernible reason. That last one would have been amazing had I been able to get the Esplora talking to the Uno, since is has every input one could ever want in a robot controller. I even fried a nunchuck trying to get things to work!
   Alas, I had figured it out, and it wasn't TOTALLY user error... A large part of it was updating to Arduino 1.0.5. As soon as I tried 1.0.4, everything works gloriously and perfectly. At least with the nunchuck. By simply tilting the nunchuck up and down and rolling it right and left, I could easily control the pan and tilt functions. After applying a mathematical filter to the servos, 95% of the jitter from the nunchuck disappeared. It was amazing and glorious, and all one night's progress!
   But I noticed a couple other issues, and I'm not fully convinced it's the coding that is the issue: I can only get the motor to move one direction. Pushing left or right, the motor only spins right, but it ramps the speed correctly. Did I get a(nother) faulty motor shield? This one I'm using is the Arduino brand one, due to the fact it can reassign the pins for one motor driver (there are two built in) to allow a forth pwm open for servo use. I would seriously like to keep that last pwm slot, as that would be my follow focus control on the camera while it's moving on the rig.
   One last thing: since I'm typing this all on my phone I can't really check to see if I mentioned this before, but I have the rail system done. And it is amazing in and of itself. Thanks to the huge array of parts at ServoCity.com (no, not getting paid by them, though I certainly would mind!) I was able to make a simple rail system that the two main rails fold together when the locking screw is not in place. By pure happenstance I realized that if I turned it on it's side, this rail works EXACTLY like a camera jib. Two awesome tools in one! Well, that's all for me, I'll see if if I can get some video posted for you guys soon!