James Cann's Posts

TRI Trial #1 - Successful enough for now

Posted by James Cann on January 30, 2011 at 2:00pm

I couldn't get the software serial to work properly, so I am sharing the serial.
This does mean I have to remove the signal wire from pin 0 every time I want to download a new program, but it's doing what I need.
1) Reading the ArduIMU+V2's data stream
2) Sending to the PC for a bit of feedback
I'll be picking up a MEGA soon...

Today was the first trial of the TRI, purely to see if it could lift off, cancel yaw, and maybe stabilize the platform a bit without any PID control.
All servos rotate based on ground_course, scaled and limited to +/-45degrees from vertical
Thrust 0 = (Global "lift" offset from RC throttle input) + (small correction from IMU pitch feedback)
Thrust 120 = (Global "lift" offset from RC throttle input) + (small correction from IMU roll feedback)
Thrust 240 = (Global "lift" offset from RC throttle input) - (small correction from IMU roll feedback)

I found these handy velcro tie-down things - This beast may be uglier than the Ugly Pink Quad could ever hope for.

In the video, when you hear me laugh it was because my finger slipped and ripped off one of the power capacitors on the RoMeo...............Quick check: Power still there? Yup! Servos glitchy? Nope! EDFs armed? Yup!
The show must go on!
Obviously the battery needs to be shifted a bit for the next shot. :-)

If at first Quad didn't succeed, TRI TRI again

Posted by James Cann on January 23, 2011 at 5:30pm

Following the death of the Ugly Pink Quad's 4th thruster, Entirely accidentally, I am going for a tri-rotor design, on

a 1/4" plexiglas base, with 3 Futaba S3003 servos for vectoring thrust.

The Nano that was originally being used for the Ugly Pink Quad, and now the Tri, well...it died, so in comes the DFRobot RoMeo.

My initial idea is going to be to use the thrusts of the three ducted fan units to stabilize horizontally and control altitude, while using vectoring to produce the X-Y translations and yaw motions as needed.

Given the wide, flat body and the ability to vector thrust, I might be able to use the plexiglas as a lift-generating device to aid battery life when not merely hovering.

Given that lift, and position can be somewhat decoupled, it may end up a very stable platform for cameras and such, even in breezy conditions.

I hope to have the IMU stream stabilizing the platform by mid-week.

First test with the AFSoftwareSerial for COMMs with the ArduIMU seem promising.

I can't help it...It's an illness I tell you...

Posted by James Cann on January 13, 2011 at 5:30pm

I have clinical multi-projectitis.

It's incurable.

After seeing the GRASP lab video sof quads literally jumping through hoops, I HAD to make one of my own.

The beginnings:

Although everyone else seems to be going open-rotor, I want zero collateral damage to people, pets, and the living

room walls, so I decided on a ducted-fan design.

4 EDF combos from Hobby King

https://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=11166

Combined thrust of 4400g.

Combined mass of 424g.

The 5Ah 4S 40C LiPo from my currently grounded Swift Heli:

https://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=10307

Mass: 578g.

For testing, the initial fuselage is a chunk of extruded poly insulation.

5 Deans connectors and a new soldering iron later (the old 30Watt one couldn't make a decent joint for the wiring harness), this is the result:

I christen her "The Ugly Pink Quad."

At present, she's brainless. I just did a test with a little solderless breadboard sending the throttle signal to all EDFs simultaneously and, not surprisingly, things were learned:

1) With 4400grams of max thrust pushing 1220grams of pinkness, it only took about 1/3rd throttle position to hand-hover.

2) It wants to rotate.

Next issues: Basic landing gear (The motors are outrunners), basic brains, and an IMU

I2C is delicate, I too see

Posted by James Cann on November 30, 2010 at 3:23pm

My first attempt at communications was going to be I2C, with the ArduIMU as a writing master, and the Nano as a reading slave. I had hoped that the 5V nano would be OK with the 3.3V magnetometer, but COMMs were never achieved, and it seems to have damaged the mag.

So, back to serial, while I wait for FedEx to bring my early xmas gift.

On the Nano is the ArduPilot 2.7.1 and COMMS are happening just fine.

Given that the 2.7.1 takes up 24886bytes of the Nano's 30720, the first step is extensive cutting back of any code that I know I won't use for now.

ArduIMU Crawler

Posted by James Cann on November 28, 2010 at 9:30pm

It's getting cold and I'm limited by space to fly and money to repair crashes, so I'm going to feed all my toy-wants by starting off with an AGV, and using the knowledge and experience to make liftoff next spring smoother than would otherwise be.

I already tried using a Traxxas Rustler, but keeping control at low speeds and preventing motor overheats was an issue, so...browse....browse...browse....ah-ha!