Tryclops

A Trinity College Fire-Fighting Robot

Okay, for you slashdotters you can look at this movie of Tryclops in motion. Look at the bottom of this page for more (older) pictures of tryclops. Here is a 7 MegaByte mpg movie of tryclops in motion

Okay, Tryclops made the cover of Nuts & Volts Magazine , and was choosen as 'Robot of the Month'
Tryclops took fifth place in the First Internet Robot Competition
You can see the schematic of the dual switching supply module that is in it's final design stages
You can see JPEG of the schematic of the dual switching supply module that is in it's final design stages
I decided that I didn't have time to finish the generic dual switching module before Trinity. It is on hold.
You can see the optics pix 2
You can see the optics pix 5
You can see the funky plotter thing for squrting goo, with a Z-axis

Monday March 5 photos
You can see the lathe I used for the mirror holder
You can see me boring out the hole
You can see the finished product, on the spindle motor

Wednesday March 7 photos
You can see the mirror holder before balancing
You can see me hogging out material on a big vertical mill
You can see Joe and the finished product, on the spindle motor

later that same Wednesday March 7 photos
spindle motor mounted in Tryclops
with the mirror mount mounted
underside of the spindle motor
Height checking the Tryclops, using a wooden mockup instead of a mirror

Early in the morning, Thursday March 8 photos
Tryclops. Still more work to do. But it might meet the height spec now

Friday March 9 quicktime movie. Sorry, I would rather do mpg's, but my still camera will do a quicktime movie.
This is a what the spinning mirror looks like. This is a wooden mirror simulation

This is a what my lab/shop looked like before I made room for the fire fighting floorplan
This is a what my computer room looks like

Here is the schematic of my PIN reciever
Here is the laid out board of my PIN reciever

Tuesday March 13 photos
You can see the schematic of the new motor driver board
I still have to finish off the 3-phase motor driver yet.

Monday March 19 photos
You can see the schematic of the new motor driver board with added switching supplies It also has a PLD to commutate the motors, and the latest version which is not yet posted has the inverters and xor gates plugged into the PLD, to save on board space.

Monday March 26 photos
You can see the triplets painting the fire-fighting house
You can see the laser driver chips that MAXIM expects me to be able to use
(I am getting the same thing in a more usable package)

And you can drop me some email at brynn@tiny.net

And if you want to see it move, Here is a 7 MegaByte mpg movie of tryclops in motion

Tryclops is my third robot. My first robot, Spot, used 4 fixed wheels in a skid-steer configuration. I was unsatisfied with the skid-steer, because turning was very difficult for my stepper motors. Also, to do any kind of dead rekoning was very tough because the number of steps needed to turn varied a great deal depending on the surface you were on. It was also slow, heavy and power hungry. My second robot, which is unamed, was put on the back burner so that I could devote all my robot time to building a robot to win the Trinity college fire fighting robot competition. The unamed bot uses two bicycle wheels oriented side by side, and the motors and electronics are up above making an unstable platform that has to be balanced at all times.

Doug at the Twin Cities Robotics Group brought his Killough-drive lego robot ( www.tcrobots.org/mtgpics/0006_05.jpg ), and after some bull sessions the merits of such a platform were clear. Doug's main site is http://www.visi.com/~dc/ Unfortunatly, to do it 'right' involved machining up a complicated drive train like Doug's Lego implementation. I would still like to do that someday, but I need to spend many hours with a mill to get what I would want. The palm-pilot robot appeared on the net www.cs.cmu.edu/~pprk/index.html and had the name and address of the company that sold the wheels it used, which would take hundreds of hours off the task of making a Killough-drive robot. I got together a group buy of wheels, and I am using a wheel that is much nicer than those on the palm pilot robot. Instead of 40mm diameter and nylon (HARD) rollers, I have 60mm diameter polyurathene wheels. I bought some 80mm diameter wheels and will be move up to them, mainly for more speed.

Here is a 7 MegaByte mpg movie of tryclops in motion The motors I am using are beautiful micro-mo 2233V018S with a 40.5:1 gear head and a 16 pulse per revolution quadrature encoder. This gives me 2592 counts per wheel revolution and a speed of over a foot per second if I don't over voltage the motors.

I am using two custom motor driver boards I designed using EAGLE, a very nice inexpensive schematic capture, PCB layout, and autorouting package available at www.cadsoftusa.com and then plotted directly to a blank PC board using an old HP7475A pen plotter and a ultra-fine point sharpie pen. The drive boards use an Allegro A3971SLB 2.5A 60V dual H-bridge, which is in a 24 pin SO surface mount package.

My CPU is a motorola 68332, which I bought from www.robominds.com from Mark Castelluccio. It is 2.9 inches square and has 512KB of flash, a 7 channel 1MSPS 8-bit AD, and 1MB of SRAM which I added to it. The CPU runs 25 Mhz and should be plenty fast enough for my needs. The great part about a 332 is the TPU. The TPU is a 16 channel timer coprocessor that does many various functions like quadrature decoding and counting and PWM, all with minimal CPU intervention.

The Robot is constructed out of 1/16 inch aluminum. There is a hexagonal plate top and bottom, and there is a 1x1x1/16 angle aluminum that is shaped into a hexagon and is bolted to the top and bottom plates. The motors are bolted to three of the hexagons sides. The batterys I am using are 3 3-subC cell NiCd 1800 mAh batterys, that give me a total of 10.8 Volts. I am stepping this up with a switching supply to 24 volts to run my motors. The Switcher is visible in many early photos riding on top of tryclops. Since I will need that space for the laser range finder I am working on, I am designing another board using EAGLE that has both a 10.8 to 5v buck switcher, and a 10.8 to 24v boost switcher and fits inside. The CPU board has a 7805 linear regulator, but that gets pretty hot and wastes a lot of power when I run the CPU at it's intended 25Mhz.