Seeker III Autonomous Sumo

Seeker III is a full-sized autonomous sumo robot, being built to compete in the 2006 ECRG. I started construction work on September 23, 2006. The overall design has some obvious similarities to Seeker II, my autonomous mini-sumo.

Below are some pictures (which you can click on for full-sized versions):

November 16, 2006 - Here is Seeker 3, in all its finished glory...

I have to say, it matches the CAD model down below pretty well.

Seeker 3 has been sponsored (and paid for) by Hylands Underwater Vehicles, Inc.

Here's a rear-quarter view. Gotta love those wheels...

Seeker 3 weighs in at a trim 1.6 kg, which I'm hoping won't be too much of a disadvantage.


And finally, Seeker 3 and Seeker 2 together. I will be competing with both robots this coming Saturday, at the ECRG.
Here's a CAD model of the current design. The robot is 7.5" square, and is 2" high. According to the rules, it has to weigh less than or equal to 3 kg.


November 3, 2006 - I realized I didn't include a decent picture of the circuit board by itself. Here's the top of the circuit board, where you can see the PIC 16F877, the LM2940 voltage regulator, the MOSFET used to turn on and off the Sharp GP2D12 range-finder sensors (they consume about 50 mA each, and we only need them running when we're actually in a match).

There is also a power LED, and a few capacitors and resistors, and a pile of various connectors for the various sensors and other stuff the board is connected to.

The two resistors beside the voltage regulator form a divide-by-two voltage divider, so I can measure the battery voltage using one of the A/D converters in the PIC.

This is the bottom of the board. I cut the circuit board from a blank I picked up at Radio Shack, and cut the shape out. To do boards like this, I solder all the socket-like components on first, and then run "traces" between pins using small solid-core wire. In general, red is +5, black is ground, and white is used for I/O lines.

In general, to do boards like this I design the board using a PCB editor as though I was going to have it professionally printed, and then I print that on paper, and use it as a guide when I'm soldering everything together.

October 29, 2006 - Things are looking pretty good. I just got back from my trip to the US, and the new tires are looking nice.

The front scoop is not attached to the chassis yet - for now its just sitting there.

Total weight: 1.625 kg

Once I replace the front bar with a brass version, and a few other pieces, it will be up in the 2.3 - 2.5 kg range, which is where I want it to be.

Here you can see that I have a little trim work to do on the "fender" on the left side. The right side has good clearance already.
October 22, 2006 - The past few days have been spent molding tires. I've molded a total of four times, with two failures. In the picture to the left, the tire closest to the camera was successfully molded a few days ago, and it looks perfect. The second one came out of the mold this morning (about ten minutes ago), and it looks like it should be okay.

I'm heading down  to the US for a week (new contract), so I will be working on software over the next week. I'm going to leave these tires to sit and harden up a little bit, and will mount them to the robot when I get back.

Here's the tire material, bought from a place called Sculpture Supply Canada. Total cost was around $43 CAD plus shipping.

The mold release works wonderfully. Inside the carton are two bottles with the solutions you have to mix 1:1 in order to make the polyurethane. As the box says, these tires are 20 durometer, which is about as soft as you can get and still be firm enough to handle the rough and tumble of a sumo ring.

If you wait 24 hours before removing the object from the mold, it works a lot better...

October 17, 2006 - Here's a picture with the lid mounted. This thing is definitely starting to look like the CAD model...

Fortunately, the lid is 0.032" brass plate, so it adds a bunch of weight to the robot. I'm probably going to machine a new front bar from brass as well, since brass is a whole lot heavier than aluminum.

October 16, 2006 - The electronics are done now, and so far everything is working. I haven't written any of the code for this sumo yet, but it should be a very straightforward port from Seeker 2 to Seeker 3.

Seeker III weighs in at this point at a lightweight 1.1 kg, so I'm going to have to add somewhere around one to one and a half kilos to get the weight up to something useful.

Here's Seeker 3 next to his little brother.

Still pending: lid, scoop, motor cover, paint, software, and tires.

I've ordered the polyurethane, and should have the tires done later this week.

October 9, 2006 - I've added a few braces, and mounted the motor controller, the battery, and the front IR sensors. The motors are hooked up to the motor controller, and next up is the main board, which will hold the PIC and a voltage regulator.

Other stuff to do: the main board, the line sensors, the rear motor cover, the button and switch holders, the front scoop, and of course the tires. After all that is done, I'll paint it flat black (of course), and we'll get to see how it works.

Here's a rear quarter view. Decked out like this, the robot weighs in at a featherweight 1 kg. I'm going to have to add somewhere around a kilo and a half to the final design just to get the weight anywhere near the 3 kg allowed.
October 7, 2006 - Here it is, with the side panels cut and installed. It is looking quite nice at this point, I think.

I cut the side panels out with a hacksaw, and smoothed the edges with files.

Here's a different angle.
October 5, 2006 - This is the wheel assembly attached to the base plate, with the front bar in place. Things are coming along nicely.

There sure is a lot more room for stuff in here compared to a mini-sumo...

October 4, 2006 - This is the tire mold I just finished making. I'll be molding the tires in the same way I did for Seeker II, which is also the same way my brother Dave does his.

I didn't have thick enough pieces of plastic, so this mold is made from four pieces. Each half is made from two pieces, which aren't really ever supposed to come apart. The two halves held together by the big bolts will come apart once the tire is poured and dried, to allow removal of the wheel/tire from the mold.

Here's a view of the mold with the wheel mounted inside. The tire will end up being 1/8" thick, so the entire wheel/tire combination will end up 2" in diameter.
September 29, 2006 - This is the full rear assembly, with both wheels hooked up to the motors. The large aluminum cross-member connecting the two sides will also connect the entire rear end to the base plate.

This part was the hardest part of the whole thing, with respect to machining. The rest should go fairly quickly.

September 28, 2006 - This is the left wheel assembly, completely assembled. The right side is mostly done - all that I have left to do on it is the motor mount plate and the black plastic bearing support.

Everything here, with the exception of the machine screws and the motor, was machined on my Sherline lathe/mill.

Here I've taken off the wheel, so you can see how it mounts. The aluminum wheel is attached to a brass hub, which is attached to the brass axle.

The black plastic piece (machined from Delrin) has a ball bearing assembly just in behind the hub (you can see the outer edge of the bearing in this picture). The bearing is to help support the axle, so the motor shaft isn't taking the entire weight of the robot by itself.

The axle slips over the motor shaft, and is held in place with a machine screw.

September 24, 2006 - Here's both wheels done now (I machined the second one this morning). They each weigh about 55 grams.
This is the front of both wheels.
September 23, 2006 - This is a picture of the first wheel I have machined, from aluminum. For reference, the wheel is 1.75" in diameter, and 1.5" wide. The eight holes around the edge are 1/4" diameter, and are strictly for looks.

I'm going to add a 1/8" thick molded urethane tire, which will bring the total diameter to 2".

Here you can see the wheel is hollowed out, with the inside cut being 1.5" in diameter, and 1.25" deep.

Beside the wheel is the blank for the other wheel, which I haven't started machining yet.

Just for reference, in case anyone is wondering, boring a 1.5" diameter hole in a chunk of aluminum on a Sherline lathe is a lot of work...

Click here to go back to my main mini-sumo page...