Ever since I saw a clip of Robot Wars 1994 on Discover channel's "Next Step" I have dreampt of being able to compete in this great event. Living in Indiana made this dream a little difficult, but now that I had graduated and was moving to San Jose I knew it was time. I already had a base for a robot from the "Wasteland" competition and figured that I could add weapons to it for Robot Wars. I sent in my entry fee and registered before I graduated. I will make an attempt at describing the event and my own personal robot.

Before I could begin building, actually modifying, my robot, I had to get into the proper mood. I surfed the web reading up about people's robots and their stories from the past 2 years. I also watched the tape of the Next Step episode about 50 times. I watched that thing in slow motion, fast, backwards, and frame by frame checking out all the details.

I did a wallet check to see how padded the bank account is. I don't really want to add it all up, but these projects really cost a lot of money. I mean a lot of money! I also browsed the surplus world to see what was available. I tend to build my robots using surplus parts or I make the stuff myself. I can't afford to pay for the good stuff. After getting a few ideas in my head I began to order parts and tinker around with some ideas.

Did I mention that I didn't have much time because I had just moved from Indiana to California? When I got to CA I had to live in hotels and a room in someone's house until an apartment became available for me to move into. It wasn't until I moved into my apartment that I even unloaded the trailer that I dragged here. I managed to find an apartment with a garage space. If they knew what I was doing in that garage......... After moving in I had to quickly acquire some equipment. I purchased a metal lathe and floor standing drill press. I also had the luxury of using the shop room where I work.

OK enough wasting your time. Here it is. The pictures of the finished product taken after Robot Wars 1996.
You are now looking at the front of Pokey. The board below is the controller. Like I said, if it wasn't surplus I made it. Pokey was built using a robot that had already proven to be tough. It was completely built from aluminum for strength, flexibility, and light weight. Because of its previous life, it is shaped like a giant toaster. It measures about 28" long, 18" wide and 20" tall. I decided to add 2 weapons to the bot. The back (pictured here) has a spike ram, the front has a blade for lifting.
Now you can see the front with the lift. The weapons were chosen to be effective for flipping the opponents over and to poke sensitive electrical and mechanical equipment. The weapons were also the easiest to implement. Well, sort of easy! I ended up using CO2 to opperate the pneumatics for the weapons. Oh! I used pneumatics because I wanted them to be fast, hit hard, and able to lift a lot of weight. There also happens to be surplus pneumatic stuff available. Another important detail was that I only had about 40 lbs to add the weapons because the robot already weighted about 60 lbs by itself.
The front lift was built using a ball-bearing drawer slide. It was made from machined aluminum which allowed me to modify it. I simply turned the center part of the assembly over so that the back side which is flat faced out. This allowed me to attach things to it like the blade and the top piece. Also this gave enough clearance under the slide to attach it to the frame using button-head screws. For a gas CO2 Turned out to be a pretty good choice. The rules do not allow gas pressures above 1000 psi. If you turn to the correct page in your CRC handbook (thanks Larry) you will see that CO2 has a vapor pressure of ~900 psi at room temp. This meant that I could have a tank of liquid CO2 and remain within the rules. It is critical to have liquid so that you get enough volume of CO2 that your weapons will last.
The cylinder for the lift had a 12" stroke and a 1.5" bore. This meant that it could lift about 250 pounds.
The cylinder for the spike had an 8" stroke and a 1.75" bore. It could generate about 350 pounds of force. Because of the big bores I would use a large volume of gas if the weapons were fired several times. I didn't want to worry about running out so I used a 2.5 lbs CO2 tank.
You might have noticed the springs that connect to the tip of the ram. Those pull the spike back. I used a double acting cylinder, BUT I could not connect anything to the return side. To get the ram to hit as hard as possible, I had to eliminate anything that restricted airflow. Believe me this made a big difference! I also did this with the lift but gravity caused it to lower.
Let us proceed by removing the side pannels. The pannels were made of 6061 T6 aluminum about 50 mil thick. You would be surprised what it takes to punch through this aluminum. The robot is built kind of strange for this type of application, but it worked! You can see all the pneumatic related stuff sits on the outer most layer. This is because it is big and, well, the last thing to be added. The bot had to move first before it got its weapons.
Looking down from the top you can see the following items. Starting on the right side: The 2.5 lbs. CO2 tank, directly attached to the tank is a pressure regulator, the regulator feeds the valve block, the valve block has 2 double acting and 2 single acting valves, an accumulator tank is attached to the end of the valve block.
The tubing coming out of the face of the valve block connects to the cylinders. I used flexible copper tubing because it is cheap and absorbs heat well. I did not want the valves to freeze so I used the copper lines to warm the CO2. In the bottom left corner is the receiver taped in place and behind the valve block is the driver board for the valves. This is a small board with FETs and diodes. Pretty simple, thats the way I like it.
Looking into the vehicle from the side one can see the battery, wheels, and the second air-cylinder. There are battery compartments between the wheels on both sides. The battery was lowered slightly below the main frame for increased stability. To shield the batteries from potential puncture they were covered with a piece of conveyor belt and 1/8 inch thick Lexan. It would take a while to puncture the Lexan!
The front cylinder was mounted in a piece of Aluminum channel. This is very thick to prevent it from being bent. If the cylinder were pushed back into the robot it would probably break the micro-controller. That would be bad. The diagonal brace also helped stabalize the upper frame and provide backing for the sides.
I removed the upper level so that you can see whats underneath it. The front ram sits directly above the microcontroller (the green board) and the motor controller.
A close-up of the regulator and rear cylinder. The regulator I got from Grainger. It is made for CO2. It is the cheapest Victor I could find with an output range up to 300 psi. Nice regulator, but it cost a lot and weighs about 5 lbs!
I removed the cylinder and cables for a better view. The microcontroller board has an Intel 80C196KB with 16KB of RAM and 16KB of FLASH. It also has a monitor program in the eprom. The board was developed for use in Labs at Purdue, but I have heavily modified it to hold FLASH, RAM and ROM in 3 memory sockets. It also has a bus for external memory-mapped IO. At the right edge of the board you can see the 2 serial ports. This is to acces the monitor program. The board works great, but I think I had a problem with the power supply underneath it.
The motor driver board I built using relays. I tried my luck at building a digital controller using FETs but it blew up. My friend tried to build one and his blew up. Wanting this to be reliable I went back to the old relay. To drive the relays I used a logic level FET. A simple buffer off my micro board drove the FETs that switched the relays. I sacrificed speed control, but I did have direction and breaking. I think the noise from the motors and relays caused some problems with my micro.
A close-up of the micro and power supply underneath. It was a good plan!
A straight shot at the rear shows the drive motors and gearboxes. The motor and gears came as an assembly. They were originally designed for those electric cars that kids can drive. I used four to simplify the drive train.
The wheels coupled to the gearbox and were mounted on 1/2 inch shafts. The wheels had pressed precision bearings for low friction and load capacity. The shaft passed trough the bearings and into the frame. A spring and cotter-pin held the axle in place. You can also see the brass sleeve the kept the wheel against the gearbox.
Finally, A shot of the tiny garage I worked in. I live in an apartment that has enclosed garage spaces. If they knew what I did in mine.........
A close-up of my work area. I have a 10 inch Logan lathe and a 17 inch drill-press. On the floor are a few pieces I worked on for the bot but didn't finish. One is a 2-cycle engine to power an airpump. The airpump would have replaced the CO2 tank. Good thing I stopped, I think the CO2 worked much better.

Overall the robot weighted 88 lbs. I was sturdy enough that my girlfriend and I could sit on it and ride. The center of gravity was very low. The bot could be tipped 45 degrees before it would begin to fall. I could stand on the lift and it would pick me up and throw me. The spike could small holes in the T6 Aluminum I had. The robot had enough power and traction to easily push a box of books I had. The box probably weighed 50 lbs. The performance of the robot was very good, but let me tell you what happened at Robot Wars.

In the first head-to-head battle, I was up against a big piece of wood. It didn't really do anything since it was built about 30 seconds before the round started. I whent over to it and tried out the spike. Because of the sloping sides and thick plywood I wouldn't have done any damage. I turned around to use the lift. I made several attempts to lift the dead wood and turn it over. Unfortunately it was really wide and just wouldn't go over all the way. Well, everything is pretty much a blur, but I remember hitting the lift button the last time and a split second later Pokey was on its side. It seems that the power of the lift caused Pokey to throw itself over on its side. I was stuck, I lost. I also found out that the rounds were single elimination. That is a crappy way to run a contest! OJ got a second chance, I want mine!

Well all hope wasn't lost. I still had the Melee. The next day I came back with the new improved Pokey. I wasn't going out the same way twice. I added stabilizer bars that stuck out the sides 10 inches. There was no way I was tipping over again. Anticipating something else would fail, I tested all day long. Everything seemed ready. It was time for the Melee and there were 5 left for the middleweight class. The Melee started and I was pumped. I planned on killing everyone else except the Ag first. After they were out of the way, the Ag and I would battle to the end. I started by pushing two of my least favorite robots into a corner. I had plenty of power to hold them there for the pin. I got them into the corner and was trying to flip one over when my car stopped responding. I checked the transmitter, but the problem seemed to be in the robot. Pokey was dead and I held 2 others trapped in the corner. The Ag got off easy! He pinned the one robot left and was declared the winner. I am sure I could have defeated him, if my bot would not have failed. It turns out the the micro froze. I simply pressed the reset button and it was alive again. Hmmmm, guess I should have put the two lines of code in that would have enabled the on-board watchdog timer. Don't worry Pokey fans, I plan on bringing him back next year!