Michael Kohn - linksys copter

CONTENTS

Linksys Coptor

Introduction

So using my Atmel Pilot firmware and software around it, I wanted to build something that could fly from it, so I stole my friend Oliver Hillmann's idea of building a quadcopter :(. Sorry Oli :(. Anyway I did deviate from his idea in that I wanted to use a Linksys router instead of a radio kit of some kind or Fonera. Using the Linksys router all the commands to the micro can be done over TCP/IP with the Naken Web webserver. For more information on how this system works, take a look at the Atmel Pilot web page. There is an explanation of the system with a diagram and a video of it working.

Explanation

So basically what this is is 4 DC motors with propellers controlled by the Atmel Pilot system. The throttle control of the joystick will control the speed of all the motors together. When the joystick is tilted the speed of the motors will become uneven causing the device to tilt.

Latest News

January 3, 2010 Yesterday I finally ordered a LiPo battery (and got a neato charger) from HobbyTown USA in St. Louis. This battery run at 14.8v and can put out 30A. If this battery doesn't lift the router, then the motors/propellers need to be changed.

August 26, 2009 Just wanted anyone interested to know I haven't given up. So earlier this year, I hooked up my oscilloscope to test some things (I've had a horrible voltage drop across the motors with the batteries I had.. actually with any supply I put on the circuit) and I guess I hooked something wrong and ended up smoking a transistor. There was literally smoke rising from this thing :(. So I remade the circuit, which is probably a good thing cause I did it so much nicer this time (pictures below). So anyway around May I read that NiMH batteries need to be charged and discharged a couple times before they hold their full charge so I decided to try it. I charged the two battery packs up (my 20A @ 14.4v power source) and ran the battery down (playing with the throttle a bit to test a firmware fix) until the motors just quit (the router stayed on). I did it a second time but this time just leaving it run without touching it. This time when the motors shut off, the router shut off too. I thought this was kind of weird, but what was even weirder was the hissing sound and then the horrible smell. I carried the thing to the kitchen to see that one of the battery cells had popped and wouldn't stop smoking. Being afraid the other cells would explode I took the whole thing outside and let it smoke out. Took a while. So after this incident, I cancelled the project. For about 2 hours. So I was thinking about it and wondered if my voltage drop was due to the small wires I used to connect the motors to the PC board. So I first unhooked one motor, placed a 600gram guitar pedal on top of the router to help hold it down, hooked the motor to the 15A @ 13v power supply I have in series with an ammeter and measured 4.5A starting and 4A while running. I then connected the same circuit except direct to the motor (not through the small wires) and measured the same thing. By the way, according to Radio Shack, these motors draw like 2A max :(. So I hooked up the motor with the small wires back to the power supply and measured voltage across the motor and got 13v. No more voltage drop! And btw, with just this 1 motor running at 13v, that edge of the router was picking up off the carpet. If it wasn't for the shape of the base and the 600g guitar pedal, it probably would have flipped over. Looks like I just have to figure out the reason for the voltage drop (which I'm 99% sure I know what it is now) and buy a new battery, which I found already a LiPo battery capable of 40A which I plan to buy in a couple weeks. This project may finally be flying soon :).

Materials

4 SuperSpeed 9-18VDC Hobby motors Model 273-256 from Radio Shack

MJ4701 Collet Prop Adaptor for 2.3mm Shaft From Hobby-Lobby
PE06055E 6x5.5E Thin Electric Prop From Hobby-Lobby
Linksys WRT54GL with OpenWRT Firmware
Atmel ATMega168
Electronic parts (read the schematic after it's posted)
Wood, paint, wood glue, epoxy. Lots of epoxy :(.

Weights

Router 350g
1 Motor 50g * 4
Wood 25g
Microcontroller circuit 75g
Battery:
- 8 AA Rechargable or Throw away 200g
- 12 cell NiMH R/C airplane batteries (I think 400g)
- LiPo ?

Weighing the entire system together, it appears to currently be around 1150g. (January 3, 2010: Some of the above changed. I have a new motor control circuit with big big big resistors and I ordered the LiPo battery).

Possible Problems

My biggest concerns are:

The Atmel firmware has an echo to the main computer but the main computer doesn't check to make sure it's the same. If some bits are lost the router could be in deep @#!#@
If the wifi connection is lost the motors will stay in the last position they were told to be in.
If the webserver drops the connection it's a pretty long delay to re-establish a TCP connection.
If the web server crashes. Byebye router.
Delay from the computer to the motors (due to network data transfer, UART speed, and the 20ms update time of the router) may make it hard to control.

Status

So all the circuitry and software is working. I can control all 4 motors on the router with my Logitech joystick. The problem appears to be the weight and probably the motors. I put the device on a scale (pictures below) and with the motors turned off it appears to weight around 1150g. When i turn the motors on full blast the weight drops to around 700g. So it looks like I'm able to push about 500g with those motors/propellers. So I'm kind of wondering if using motors actually made for R/C airplanes would fix it. The problem is an R/C airplane electric motor can cost between $80 and $200 (while these cheapo radio shack motors I used were $6 each) and I've already sunk quite a bit of money (for me at least) into this thing, so I'm not sure if I really want to try to sink another $320 to $800 into it. If anyone would like to donate 4 identical motors to me, let me know. Also, if anyone has experience with motors/propellers like this, feel free to email me any suggestions

I have a YouTube video of the thing almost working here: http://www.youtube.com/watch?v=eYOMHge0_yk

Pictures

	A brand new Linksys WRT54GL router with a fresh install of OpenWRT Linuxsnapped apart. I think my warranty may be voided.
	A view of the wires I soldered to the UART on the motherboard of the router.
	All put back together with its new dual serial ports out the bottom.
	Here's the first propeller I got and motor mounted to a candle holder sitting on top of a scale. I was trying to use the scale to measure how much thrust this prop/motor combo would give using different power supplies, altho because of the way the motor was shaking, I think it may be a false test. Either way it only seemed to want to push about 25 to 50gm, which isn't enough for what I need.
	Here is the circuit board sitting next to the router and one of the motors while I test to make sure I didn't mess something up. This is one of the most annoying circuits I've ever had to make. So many parts and not enough space the board.
	Here is the final product minus the 8 AA batteries that will go at the bottom. The frame is made out of all wood with lots of wood glue and epoxy to hold it together. Since the router was already mostly blue, it was pretty obvious I needed to paint the structure yellow so I could have my favorite color pattern: blå och gul.
	This shows the router sitting idle on a kitchen scale with the newest propellers I've been trying (bigger than the original). It's hard to tell from this picture, but the whole thing with these R/C airplane batteries (12 NiMH cells instead of the original 8 AA batteries I was using) weighs almost 1200g.
	Here is the router again, zoomed in on the scale with the motors running at full blast. The scale now reads 700g which means I'm only pushing around 500g.
	(June 28, 2009) -- So I goofed up while probing the original board with a 'scope and shorted something out which caused it to catch fire, so I redid the board a couple months ago and finally installed it yesterday trying to do a better job with the wiring and such. This circuit board has much better part placement and organization (as can be compared to the board above) so it's maybe not a bad thing this happened :).
	(Sptember 20, 2009) -- So I swapped out the 2N3904's for PN2222's and swapped out the 220ohm resistors for these big freakin' 47ohm/5 watt resistors. This should allow enough current to flow through the motors so it should actually work now. Just need to buy a battery now :).

Schematic

Download

Coming Soon

This page was designed to work with all standard HTML compatible web browsers including Firefox, IE, Safari, and Links.