UPDATE: Jan. 21, 2012
I decided to update this blog right now mostly to summarize for myself what needs to be done, not so much a proclaimation of any sort of victory. Taking this project back fully into my soul, I found it interesting that this robot has been almost a year in its creation. Amazing how long projects take for me to complete. To my credit, there has been a lot on my project timeline for me to take care of and a lot of worthy distractions. That and skills that are necessary to build this robot are on the fringe of my knowledge base. A lot of theory implemented by unsure means.
However there is a new challenge presented by DFrobot that requires the use of one of their mobile robot platforms. Since the RP-5 chassis I'm using qualifies I thought I had a good chance at this challenge. We'll see! I have till the 14th of next month.
Recent activity:
"Eye thing:" I got to a point where I needed to fully build the final version before I moved onto testing other systems. Only the final build could test all the subassemblies efficiently. There was the eight-LED angel eyes that will flash to the audio playback, the bellcranks and their articulation, and the "headboard" PCB that would fit between all the bellcranks and wiring. After much deliberation and patience in design and fabrication the "eye thing" unit is complete, save for the rear cover which I'll make if I have enough time, or after the challenge is over.
Blinking brain in a jar: In The Beginning, I had visions of a brain that would blink, pulsate, and glow. I had to back off those dreams and just get a brain that would light up enough. I went through several iterations of how to do it. From onboard processing for the 28X2 to take care of, to driving an I/O expander that would power the LEDs, to having a separate uC to handle its own I/O expander. I settled on a 08M2 driving the PCFxxxx I/O expander I show in some of my videos. The 28X2 will send an interrupt signal and a data byte. The 08M2 will run a lighting scheme based on the data byte received. The 08M2, PCFxxxx, and LEDs will be buried in the brain itself with four wires leading back into the chassis.
There were several issues I needed to consider with this part of the build. I made the solid pink silicone brain some time ago but was having problems getting the LEDs to shine through the material. Also I'd not really done any research on the densities of the materials I was using. I'm going to suspend the brain in a mix of water and glycerol (for viscosity) and a touch of blue food coloring. It turns out the density of cured silicone is greater than either. I don't want to have a solid strut holding the brain stationary in the brain jar. I want the wiring from the main PCB to the brain circuit to act as a tether. I think some movement of the brain when the juice is sloshing around will give it a more natural (?) look. So it became clear I would need to hollow out some of the brain if I want it to float. The brain mold is a two-parter so it will be easy to carve out some material from the upper half of the brain. I'm still trying to figure out what is necessary to give the brain an even buoyancy that allows the brain to sit level in the solution; i.e. not nose up/nose down. Being hollowed out should also allow the LEDs to penetrate the remaining pigmented silicone better. There were also only going to be two LEDs of each color (red, blue, green, yellow), one for each pin on the expander. I've decided to use the pins to drive PNP transistors instead, each of which will drive two to four LEDs apiece.
I also was not quite sure how to add depth to the outside of the brain. I'd tried painting darker red "veins" on the inside of the mold before pouring the pink silicone but the results were terrible looking. I've decided on partially filling the cracks of the brain with red colored silicone. Fortunately I've got a couple pieces of scrap brain material from earlier test molds that I will try it on first.
Brain jar/cerebral fluid design: For the brain jar itself, I want the fluid to be removeable, I want the jar to be detachable, and I want a solid connection between the brain jar and the bot chassis when connected. Flimsiness is not acceptable in this situation. I will custom mold a pigmented silicone grommet that will act as both the passage for the wiring and the plug for the fluid. The base of the jar will be a 1/8" clear acrylic disk glued over the opening of the jar. The jar will sit in a slight recess routed out of the expanded PVC chassis cover. Holes for mounting bolts will be drilled and tapped from the inside of the bot. There will also be flush mounted nuts in the PVC to give the mounting bolts a solid thread to grip onto.
Custom yellow tread difficulties: Still debating on whether or not to get some Shore 75A hardness polyurethane to replace the softer 60A tracks I have now. They can be streched out with a little effort and sometimes slip during usage. Comparing the stock tracks and the ones I made I'm guessing the orignal tracks are hardness 75A.
Colour is more important than color even: The gray wheels have proven to be a challenge. I received a yellow chassis to my surprise and tried to paint the wheels the appropriate gray but it rubs off where the tracks make contact. Sparing no expense, I went and spent $50 on a new gray chassis just to get the proper colored wheels! Oh, the madness. I like the size of the RP-5 so I'll use the other bits in another project for sure, no waste here.
New amplifier for MP3 player: I got the 3W amp that Chris the Carpenter is hot on and tested it out. It is an awesome little amplifier on a tiny little PCB. I found it interesting that the TDA7052 amp was actually louder when I had it on full gain. There was a little clipping for sure but it was still clear enough to understand. Design limitations dictate I use this 3W amp so I'm stuck with the voulme I have. I'm a little concerned it won't be loud enough, oh well.
Code, oh, the code: I've got my "complete functional" code written but many things leading up to this point have not been done. I'm having a lot of trust in my programming to be spot-on right out of the gate. I've been through so many code iterations due to the many methods I've conceived for the LED scheme.
I've got almost half the program memory left over and am considering how to have the bot wander about and how to determine what objects are people. As it is the bot will track and follow you but I want a little more character. Maybe add some sass too, who knows. I need to do more voice tracks to reduce the repeat playbacks. That will take a good bit of time recording, converting, and adding the data to the program. Not too bad though.
Alright, I'm off to go put the eyes back in the bot, slap the new programs in, and give it a whirl.
Artifact from previous blog version -
Going to convert this into my typical blog. I don't have anything to say right now as I just ported everything, including my thoughts, over to the robot project page. Eventually something will be added here. I'll keep the viddy.
https://www.youtube.com/watch?v=2jfhkkdrRe0