April « 2009 « Productivity Sync

April 11, 2009

Eagle cad is a pain to use

Filed under: Uncategorized — markgross @ 12:52 pm

“No forward-/backannotation will be performed!” is the curse of my day.

I played with the schematic, and reviewed / addressed some of the error tool warnings, and somehow this error went away!

At the dorkbot meeting one guy told me that when he gets the error its because there is a missmatch between the PCB and the schematic, and that the way to fix it was to nuke the PCB and auto generate a new one from the schematic. I don’t kow what I did to get around the issue but I’m happy its behind me now.

Thicker lines for current signals:

My current issue is now how to rout thicker signals for the 4A max current load from the H-bridge. I don’t yet know how thick the lines need to be to carry 4 amps. I do know that my current board that uses as large of line i could fit fails the default board rule checks and I need to rip up those lines and redo them.

After some playing around I found the rip up tool and a dialog box for setting the min line width. These allowed me to rip up the low current lines and re-auto rout new lines (thicker) where I needed. Doing this a few times allowed me to get the lines I need for the H-bridge output.

Work in progress:

My schematic and board layout can be seen at eagle.tar.bz2

I keep updating it so check back every once in a while.

Board Fabs for stuff like this:

Monty recommends www.4pcb.com for its first time user discount and they do good work for his bittybot

I know that www.sunstone.com, www.pcbexpress.com, are options and I’m still shopping around.

Comments Off

April 5, 2009

Line Maze Solving robot

Filed under: Uncategorized — markgross @ 12:11 pm

The Task:

build a <6″x6″x6″ robot to solve the line maze problem http://www.robothon.org/robothon/maze.php

The Plan:

Use the following bag of parts:

Teensy++ micro controler
2 Pololu QTR-8RC sensor arrays
L298N Dual Full-H bridge driver
2 Portescap gear head motors with encoders that where gifted to me (I’m afraid to look up how much they cost)
Assorted mechanical components (wheels, chassis)
Assorted discreet bits (resisters, capacitors and whatnot)
Custom PCB laid out in Eagle
Implement a Dijkstra route solver
Implement a map building algorithm
Implement PID / PWM motor speed control with odometery
Some UI code

My goal is to have this bag of parts assembled into a robot that will track lines well enough to start implementing the mapping logic and performance tune the moter speeds and line sensors before moving to the route solving. I’m currently thinking the project will take about 3 months. 1 month to build the platform and get signes of life. 1 or 2 months to get the odometry and PID working well, and to implement the mapping and route solving.

If I get mechanical signs of life by the May PARTS meeting I’ll be pretty happy.

TODO list: (last update April 18, 2009)

lay out schematic for PCB
- created library part for the teensy++ (done)
- PCB schematic and board layout review (mostly done)
- I have a needling feeling that I need to suffle a few signals around to free up the serial port pins.
- laid out schematic (done, see latest files here)
map out and size data structures for mapping and route solving (only have 4K RAM but lots of flash and 2K EEPROM I can use.)
Locate 2 cm weels or figure out how to rubberize the 1.746 cm mounting hubs into wheels (mostly done)
mount motors and sensors to chassis. (done)
bread board the encoder circuit to see if I can get away with using the IO pin pull ups (50Kohms) instead of the discrete 10K’s
Fabricate the PCB
Populate PCB and test.
Mount PCB and connect sensors and motors
Stub out program working top down mostly, then fill out the AVR specifics for PWM, timers, and banging the ports..

Comments Off