Ria and I finished our new dining room table. We picked up a bunch of random reclaimed lumber from the ReUse People in Oakland. We got a bunch of cool old rough sawn 2x3s that were full of tons of little tiny nails. After a long time with a metal detector, vise-grips, and a dremel to remove the nails, we milled everything down to random widths, 6′ long about 2″ thick. After cutting down enough wood to make the table 3′ x 6′ we made a jig for the drill press and drilled 4 holes in all of the pieces of wood. We ran threaded rod through the holes and put nuts on each end to hold the wood together while we glued it up.

We added an additional ring of wood around the edge of the table to cover up the nuts on the threaded rod, and the end grain.

Once the glue dried the table top became incredibly rigid. Using a power hand planer, and a belt sander with 40 grit, we leveled the top surface. Next we filled all of the knot and nail holes with black epoxy. Once the epoxy was dried we ran a 80 grit belt over everything and then 100 and 220 grid sandpaper on the orbital sander. Robert sprayed the top with a 2-part conversion varnish which looks excellent. Ria and I scored a reproduction Herman Miller/Eames table base at a garage sale for $10 that works perfectly with the top.

We left the underside of the table rough. Now we just need some chairs.

I’ve been working on a quick prototyping board using the PIC18F4550. It’s crammed on a 5cm square board to take advantage of the Seeed Studio PCB service of 10, 5cm square PCBs shipped for $14. All of the PIC IO is brought out to headers on the sides of the board. Ideally I’ll make 5x5cm boards to attach to the top or bottom similar to Arduino shields.

The board is almost all surface mount components. One of the first projects will be to make a toaster over into a USB controlled reflow oven.

While I wait the 30 days or so for the PCBs to arrive, I’ve been playing around with EagleUp which exports boards from Eagle to SketchUp. It has a very small parts library, but I’ve been learning SketchUp to add some more parts.

“I fly like and airplane with unlimited fuel”

“Girls are right, the semicolon rules”

Be warned, it’s not rated PG. Visit the wrd

Do visit, twowines.com

Detailed Project Writeup

The exterior structure of the Breath of Fire is a ten foot tall black obelisk made of a wooden 2×2 frame covered with a thin plywood skin. The tip is covered with sheet metal, and has a crafty custom pilot light enclosure. the BOF can be set on wheels to be mobile or sit stationary on the ground. on the front of the structure at about 4′ from the ground is the main control interface with the LCD readout, the blowhole, start button and indicator LEDs. on the backside is an access panel for changing propane tanks, and accessing the manual valves and electronics. The BOF is transported partialy disassembled and is assembled in a cryptic order that somehow seems to work.

Overall (.pdf) Details general layout.
CAD Dimensions (.pdf) Measurements for one side of external structure.

fuel is propane fed from standard (BBQ) 20lb propane cylinders. the propane starts as liquid in the cylinder, and when the valve is opened the liquid boils into a gas and is piped through 3/4″ brass into an expansion chamber. At 70 degrees propane will expand to about 130 PSI, higher the temperature the higher the PSI. the BOF is equipped with quarter-turn manual shutoffs both before and after the expansion chamber. the output from the expansion chamber is controlled by an ASCO 8210 solenoid rated for propane well above 130 PSI. the pilot is fed from the main supply tank via 1/4″ copper tubing to a custom made pilot light just below the output at the top of the tower.

Propane MSDS (.pdf) Material Safety Data Sheet for Liquid and Gas Propane
Creating Dangerous Art Safely From burningman.com
NFPA 160 National Fire Protection Association, Standard for Flame Effects Before an Audience

The brains of the BOF is a Parallax Basic Stamp II (BS2). The input comes from a Figaro TGS2620 ethanol sensor with integrated heater. The sensor is interfaced with the BS2 via an National Semiconductor ADC0831 analog to digital converter. The LCD panel is a Matrix Orbital LK-202-25 2×20 serial LCD display with backlight. the BAC is calculated on µM-FPU math coprocessor. the formula to convert from the sensor output to blood alcohol content was borrowed from the breath-o-matic.
Flowchart (.pdf) Overview of electronics functionality.

Skot Croshere, Adam Bannister, Dave Croshere, Ria Blackburn, Katie Wilson, and more!

The Wickerpedia.org

It started when I found a ancient RV-M1 industrial robot sitting in the basement at CCSF. Apparently it had been donated a while back, but didn’t include the manual so nobody had really used it. After a little googling I was able to locate the manual and the service manual. The robot arm is so old that it uses a DB25 serial port, so I had to convert it to the not-quite-as-old DB9 port. After a little fun messing with the serial pinouts I was able to speak to the robot using a USB serial adapter and my macbook. The RV-M1 has a gigantic “computer” box attached to the robot arm that will do the inverse kinematics, you you can specify a 3D coordinate and press go. I quickly found out that the joint servos are quite strong and fast, and there is no protection to keep you from smashing the arm into the table. fun!

At some point I decided to get the robot arm to solve the Towers of Hanoi puzzle. After doing a little reading I was particularly intrigued by the legend behind the game;

There is a legend about an Indian temple which contains a large room with three time-worn posts in it surrounded by 64 golden disks. Brahmin priests, acting out the command of an ancient prophecy, have been moving these disks, in accordance with the rules of the puzzle, since that time. The puzzle is therefore also known as the Tower of Brahma puzzle. According to the legend, when the last move of the puzzle is completed, the world will end.

 

I had recently heard in an astronomy class that scientists predict that our sun will die in approximately 5 billion years, and in it’s final stages of life it will swell up and swallow the earth — surely this will be the end of the world. The problem is that with the original 64 disks, even if the priests moved one disk a second, it would take 585 billion years to finish. The robot arm is fast but it definitely can’t make a move in 1 second.

With the idea that the robot would go in a display case at school, I decided on one move every 70 seconds and 51 disks, for a total completion time of 4,992,815,678.847 Years, just in time for the end of the world.

The RVM1 has a pretty limited reach in any one dimension so in order for it to handle 51 disks they needed to be thin and each disk could only be 1/8″ larger than the previous. Using a threaded rod and a router table I built a jig to precisely cut round plastic disks.

The 5 different plastic colors made for some pretty saw dust.

Since the rules of the game only allow the player to handle the top most disk in any stack, I outfitted the robot arm with a magnet and countersunk steel washers in the bottom of each plastic disk. By just bringing the arm near the top disk, it grabs onto it, and then once the disk is on the pole, I can just swing the arm away to release it. It works really well.

Getting the algorithm for solving the puzzle into Processing was no problem, as this website has the problem solved in almost every language.

It hasn’t finished the puzzle yet… luckily. This video shows a much faster version.