earlier this month i attended robotic2012 in Birmingham to watch the UK micro-mouse competition. After building a micro-mouse last year for a project it was really interesting to speak to some of the best in the country. whilst there we also got to see some of the other competitions they were running such as drag and robot tracer.
Robot tracer is an event where a robot has to follow a white line on a black surface around a course, you usually get up to 5 runs to set a best time. the robot who completes the course in the quickest time wins. many of the robots that entered this year seamed relatively slow and many were made from “off the shelf” robots, some even used stepper motors to drive the wheels. having built many simple line followers it seamed like it would be relatively easy to build a small line follower that could easily outrun the slower bots.
so what sets this competition apart from previous line followers i have built?
the interesting thing about robot tracer is that the course is made up of straight sections and curves of set radius. on the track there is a marker at the beginning and end of every curve or straight section as well as markers for the begging and end of the circuit. this means that if you can read the markers you could easily learn the course on your first run and then adjust speeds for your other runs (faster on straights and slower on tighter curves). most of the robots that entered this years competition were not able to read these markers and did not learn the course, limiting top speed.
i have decided to try and build a line follower to enter next years competition. i intend on building a small, light robot capable of learning the course. this will require the robot to have some sort of encoders to record position as well as accurate control of wheel speeds.
after doing some research i have purchased some small pololu gear-motors for the robot as well as an atmega 328 based microcontroler with a motor driver. at the moment this has been mounted to a breadboard for easy prototyping and is running off a spare 7.4v battery from my micro-mouse (far to big). i am currently using tires from some Solarbotics RW2i wheels however i have printed my own hubs for these, along with a prototype chassis to hold everything together. this is only a rough design and i intend to make it allot smaller eventually however this will make testing ideas very easy.
plan now is to work out some sort of control for the robot (either Bluetooth or small RF transmitter) and also work out how i can add encoders to the motors.
So today I decided to upgrade my print bed to something a bit stronger. I have read that many people use a glass printbed so decided this would be a good idea.
firstly I found a piece of mirror that was the correct size. I realised that I needed to drill holes in the corner to screw it down. 4 mm in and 1 drill battery later I realise that you cant drill glass without a glass drilling bit. TO PLAN B, I cut the corners off and mounted this to my print bed with the protective frame corners that the mirror came with. whilst I had the print bed off I also added some cardboard to the under side to provide some insulation as I have been having problems getting the bed to temperature. When I powered it up I instantly noticed it was heating up much quicker and when it reached temperature it was much more stable.
It seamed to be going well but then as I started printing I realised that nothing was sticking to the bed. the surface of the glass is to smooth for the ABS to grip to. I tried to rough it up with some sand paper but this didn’t have much effect. after doing some research it turns out that glass is only suitable for PLA. some have had success with sand blasted glass but as I don’t have a sandblaster to hand I cant really try it.
LESSONS LEARNED :
1. insulating the underside of your heat bed makes a big difference.
2. glass print beds should not be used when printing ABS
3. don’t try to drill holes in glass, if you don’t have the correct equipment its just not going to happen.