Blue was intended to be a solar powered, two motor - four legged walker with reverse, and phototropic behavior.

Intro: Blue is based on the Miller type configuration, with a multiplexer added (as defined at BEAM Tec.). I had planed to have him roving the backyard this summer (2001), but although operational, there is no chance of him walking through grass (he can not even handle thick carpet).

Circuit: I first tried a free form of the circuit, but troubleshooting soon became more that I could bear. Utilizing a blue PCB not only gave him a name, but also provided a nice layout that can be easily (relative to freeform) trouble shot. From right to left on the board, there is an 8-position dip socket, with the resistors defining the Nu core speed. Next is a 74HC14 with five of its six inverters utilized. Four inverters are used for the motor control (two motors X two directions) and one for the reverse timer. Then there is the 74HC244 "multiplexer", with enable channeling motor controls in reverse. Last is the 74AC245 motor driver (stacked two high to increase power output). Schematic by request. Although the reverse is operational, its function is less than optimal. He dose not turn upon reverse, so he will back up only to hit the obstacle once again when moving forward. Although the reverse feature didn't achieve the desired potential, utilization of the 74**244 as a multiplexer opens some interesting options, and was a very educational implementation.

Sensors: There is no need to rush to add sensors to a walker with a gait that dose not work in the first place. I intend to make him a shadow avoider, instead of a light seeker. To do this, I locate the "eyes" high, looking down. This will protect them from the sun (avoiding following the sun into a shadow), and instead he should go till a shadow or dark area is seen, and then moves to avoid it. Testing with large photoresistors (5k ohm in light and 20M in dark) it was obvious that the turning capability was virtually useless for obstacle avoidance due to the extremely large turn radius (about 20'). It is no surprise to me that the "Bio Bugs" use a little dance to turn, rather than turn while in motion. I believed that the photoresistor would create a gradual change in resistance to handle varying light levels, but this is simply not the case, and for the most part the influence to the gait was all or nothing (fast shift from low to high resistance). Photodiodes gave a much more consistent and influential change to the walking gait, but still had the turn radius issue. For the time being blue will need to survive without sensors, but his availability as a test platform is indispensable.

Mechanics: His legs have been reformed all over the place. I used stainless steel welding rods, that allowed the high strength/flexibility to weight ratio. There was allot of manipulation to the leg configuration to determine the best position relative to the center of gravity. He would go left, then right, dance, walk and finally trot, all due to the position and angle of the feet relative to the center of gravity. I was very surprised by they influence the leg configuration had over the walking performance. To determine the optomum walking gait, I routed the micocore resistors to an eight-pin dip receptacle. This allows for a very versatile microcore, and makes it very easy to try out many different resistor values rapidly. This was the smartest thing I did with the whole design. I can also try out photodiodes, photoresistors, plain resistors and combinations of the bunch. I found the fastest travel using 1M resistors for the front motor, and 2M on the back, but he would flip over when reverse was activated. With 1M resistors on front and back motors he was not as fast but would transition to reverse walking without issue.

Power: For trouble shooting, solar power was not an option. I used a lithium camera flash battery, light with good steady power and duration. By far the best battery choice I have come across.

Solar: Solar power was a real trouble for this guy (me and the bot). Failing at all attempts to regulate a stow/dump approach, I instead went for a direct surplus approach. I used a total of thirty-three $0.5 solar cells from allelectronics (part number #SPL-22). I have three groups in parallel, of eleven cells in series. Cells are sold as 0.45V 100mA, so with the configuration I have, the output should be 4.95V 300mA. This is large and bulky, even with the minimalist approach to materials I used. The cells double as a splash guard, and provide and element of redundancy due to the parallel nature of the three series groups. At present, he is still trying to walk. Due to the proximity of the solar cells, he seems top heavy, and rolls over any chance he gets.

Cost: I spent $10 a piece on the motors, using BGmicro lens motors. The solar cells worked out to be $16.50 in all, but I'm not utilizing them at the present. The battery was $10, so a little expensive but since most of my work is done at night, with short run times to evaluate different configurations, I felt it was the best choice. The circuit, wires and other stuff were under $5. So for now, work continues with a robot for about $35… But what did it cost, time wise?? Well, I had the concept completed in about 35 minutes, and you know what they say about inspiration vs. perspiration. Then I had to trouble shoot it. Then rebuild. Then I hooked the battery up backwards and fried the PCB. Then rebuild. Then the legs broke. Then I rebuilt it again. Other than that, it was little effort, and Blues operation continues to supply much enjoyment and inspiration.