Journal98 aug
From Simreal
August 1998
Fri, 7 Aug 98
The SORC (Society Of Robotic Combat) has a web page now! It's just a start, but check it out anyway. Eventually, this should be the premier web site for people who want to build killer robots.
On Tuesday I received some samples of the parts to be machined for Boris -- I was thinking the whole order was coming, but it makes sense that I would receive some single pieces first to check them out and confirm their useability. Charles has now assigned me an apprentice machinist (Hi Shane! Charles runs a training shop, where he creates the great machinists of tomorrow.) and I've sent over the modified drawings for him to hem and haw over. So now I'm waiting for parts again... dang. Anyway, I've got my own metal shop (for what it is worth) set up now and I can start cutting and assembling the six legs' frames at least. Can't finish until I get parts, of course, but I can get a good head start. There are pictures of the new shop and machined parts on the picture pages.
Normally, I would list the Boris tasks I plan on doing this weekend here... but this weekend I have an important non-Boris project that will probably take every spare second. We are doing integration of hardware, communications, and software on my robot contract -- which will be a good workout of the Brain software (which Boris uses), but won't allow for any time cutting Boris metal. I might be able to run errands today, though, and fill my gas tanks (if I'm really lucky).
Oh, and I found another gas-powered compressor at the local hardware store -- It's a Coleman Contractor, and seems to have roughly the same spec's as the Grainger job I've been oggling but at $700 instead of $1,000 -- and it looks slightly smaller, and seems to be lighter weight too!
Now that we've been using the Brains software (in simulation mode) for a month or three... and I've had another programmer using it, and getting some really great insights no how to make it better (Hi Seth!), I realize there is a fair amount of work I need to do to it before it is really great Brain software.
It's pretty spiffy now, but greatness is achievable and therefore I must reach for it. My plans also include making a "small" version of it for running on Microcontrollers -- a version that will eventually be made available for other robot builders. Don't know if it will be free (as a teaser; gotta buy the full development environment set to get the tools we should be making soon) or paid -- it isn't entirely mine to make that choice with -- but I do want it to be out there for other robot people.
One of the hardest tasks in robot work is getting the Brains working (that, and getting the motors running). Most non-programmers are stuck with the Basic Stamp or some other form of highly-limited programming system; very little memory, and not a bunch of speed either. My goal in life is to make the Brains aspect of robotics more powerful, easier to use, and available to many people. That will probably also require that I provide hardware kits -- but that is easy enough. I want to design a simplified Boris Board (with less stuff, but still lot's of I/O) for each of the major MCUs... several PIC versions, Atmel, and HC11. Not this year, but eventually.
This Robot stuff is way too fun, and needs to spread around as a hobby! Plus, there must be more people who want to run Autonomous robots but just can't get around the software barrier. Watch the SORC website for details, when they come... again, probably not this year, but 1999 is just around the corner.
Mon, 17 Aug 98
Setbacks. Robotica has been cancelled (and would have happened by now anyway), because Profile (of Robot Wars) placed an injunction. Some excellent discussions (and some not-so-excellent ravings) are on the Robot Wars Forum. And now, a new face is stepping into the ring -- Dave Cole, ready to redefine robot combat as we know it, and kick sand in Profile's face if need be.
Most of the SORC boardmembers have finished their moves, so we should be able to get incorporated soon and start being useful.
And finally, Boris. Some of these setbacks are from before my move to Texas, some are recent. In spite of it all, I will have a running robot by 1999. Let's see -- Motorola discontinued not only the pressure sensor I had designed the Boris electronics for, but several in the same class. They left me with one available -- the expensive one, of course -- and it has a completely different signal output. Of course, it is a 4.5V signal, so I don't need all of the pre-amplifiers. Nice piece, really. But it obsoletes the Boris Boards, or at least that corner of them.
My contract job also pulled back some of my PIC development hardware -- mostly the UV eraser, so I would need to get a new one of those. But thinking on it, I really don't like the UV erase cycle on PICS -- it slows me down far too much, even with a half-dozen chips to run through. And the ICE systems for PIC cost more than I want to spend.
The PIC16C74 doesn't work with the MAX3100 UART... the SPI modes don't mesh or something, so I am left bit-banging the SPI port... and that is just so lame. But I definitely want the FIFO of the 3100, so I can ignore the UART for long periods of time.
So... hmmm... tough choice. I could jumper some stuff on the Boris Board to bypass the A/D pre-amp. I could buy another UV erase. I could keep banging the UART. Or I could spend another $100 to $200 and make some new Boris Boards with a different MCU. I went overboard last time, ordering all the boards I thought I needed; next time I should just make Two and test it out (cost more per board, but saves money overall if they don't work out).
And for the MCU? I have some HC11s, but I'm thinking I want to play with the Atmel series some. They have Flash programming, and seem to have good specifications. Today, I am ordering their development kit (ATMCU00100 at $49), plus a few of their high-end AVR chips (AT90S8515 at $10.55 each), and start re-designing the Boris Board. Heck, I should be able to make it way better the second time anyway!
My metal shop is still in Georgetown, a half-hour away. The last week or so I have been totally swamped; I'm hoping to start cutting metal again before the end of the month though. Gotta get the hardware in place. Haven't heard about my machined parts yet this week... I'll have to send e-mail.
Some good news for a change... we sold our house in Oregon at a profit, and we were going to have to use that profit as down-payment on a new house here in Texas. It seems that my wife is still eligable for a VA loan at zero percent down and a good interest rate! Egads! That means I have a stack of money in savings that I could actually spend if I had to... well, I won't spend it all because that would be stupid (it's nice having a few month's income buffered in savings), but I can spend the $4,000 or $5,000 I need to complete Boris... that is, when I am sure it is safe. I'm thinking September I buy the compressor and pneumatics and put it all together. Then, Boris should begin operating in October! Cross your fingers, out there -- we may make it yet.
I also have a friend who wants to help me make the vision system for Boris work, for the autonomous mode.
Wed, 26 Aug 98
That stack of money in savings is shrinking... I put in my pneumatics order for the Boris actuators this week. Got some excellent deals from my local Parker-Hannifin dealer, Wilson Company. I'm also about to dive in and get a portable compressor... maybe even this weekend. My eye is on the Coleman Powermate Contractor... it has good weight, and good stats, and isn't as expensive as the Grainger options.
Last week I received my Atmel development kit. I've been playing with the Atmel a little bit now, and I'm happy with it so far -- especially about the rapid turnaround on code testing. I need to do a few of my "benchmark" functions on it to get fully up to speed, but I should be creating a new Boris Board design off of it soon.
Since the Atmel can use external SRAM without pain, I am going to add some so I will have a HUGE (relatively, at least) progamming space to work in. Probably I'll use FRAM, though it's a bit slower than other static RAMs... but it's cool, and holds its memory for ten years.
I'm also probably going to memory-map some I/O off of the system (with external RAM activated, most of the ports get taken over by memory management)... but I'll have to see what's what. Of course, I'll still have a couple of UARTs, a couple of SPIs, and the serial A/D... unless I memory map the A/D and use parallel chips. Dang, more decisions. I did this all once before, too -- it seems like a big setback, but I should end up ahead of the game for it in the long run. With what I learned on the PIC boards, I hope to be able to make this one rather a bit cheaper (read: smaller).
The SORC is moving right along... with rules for Competition, Arenas, and even some independent submissions for Autonomous Beacons coming together nicely. It looks like there will be a big event early in 1999 on the West Coast, with more coming up through 1999. There has been some event noises from the East Coast, and overtures from the UK about starting SORC branches there! Sounds like Boris might get a bigger audience than I had anticipated...
And finally, I hope to cut some metal this weekend... no welding, but if I can make a pile of parts I should be able to start assembly of everything early September aiming at a complete system starting testing in October.... which means I may be scaring trick-or-treaters with the heavy iron this year ;-}
