Started construction this weekend on a new robot project.
This one has an on-board PC that will talk to an Atom Bot Board which in turn will talk to an SSC-32.
Phase one of the hardware assembly (read: the easy part 
) is about done. Once I have the rest of that done, I can start the hard part - the coding.
The initial build report and some pictures are up on my site:
A big thanks to everyone here whoâs been answering my various questions the last couple of weeks!
Mike,
One thing I noticed about your motor in the picture below are the 2 tiny screws holding the entire assembly. I think you should put some kind of support bracket on the motor housing to help those screws.
http://www.madmaker.com/images/pcbot/drill-motor-small.JPG
Just a suggestion.
Theyâre actually fairly substantial - they run the full depth of the nose of the drill.
Also the nose of the motor assembly is flat on two sides, with the flat portion at the bottom held squarely against the aluminum angle.
But yeah, Iâm waiting to get all of the electronics situated on the lower level first, then Iâll fit a brace on the motors from below.
I have to say again your workmanship is awsome.
Thanks!
If I had Photoshop, itâd look even better! 
It occured to me that you might also have meant those little tiny silver screws. Those actually are my crude solution to the fact that the drills came with a cheapo version of a torque clutch on them.
(Iâd have taken a picture of that, but I was chasing the BB-sized ball bearings around on the garage floor after I took the first drill apart and they shot out all over the place. An audio recording would have been pretty amusing. )
I pulled all of the âtorque adjustmentâ stuff off the front of each motor, flipped the original retainer ring around backwards, and then used a couple of servo mounting screws on each one to force it down against the ball bearings, effectively locking the clutch. So those screws fortunately arenât providing anything but a bit of clamping force to the clutch.
I only noticed the black screws screwed in to the aluminum angle framework. I trust your design, itâs just the photo look as though the weight of the motor was free to bounce around.
I canât wait to see it running off the itx board, this will be neat to see. With your design, you can put so much on it with all that room.
you might want to take a look at gimp, it is a free program that works under linux and windows. Does a lot of things that photoshop does as well.
if its just light balancing and croping of images ⊠picasa.google.com/ works good for that⊠I havenât used it for anything besides that.
Got the speed controllers, the SSC-32 and the ABB installed into the bot last night.
For now, because I had them left over from a 12lb âcombatâ bot, Iâm using two inexpensive RC speed controllers. After my hobby budget recovers a little, Iâd like to go with a pair of Jimâs HB-25s. (BTW Jim, if I get a Lynxmotion tattoo, can I get a discount? )
Made up the power harnesses for all that. Iâve decided to run three batteries: the SLA for the PC, a battery pack for the drive motors, and a battery pack for the microcontrollers and servos.
I have a feeling that, if I donât undervolt those drill motors, Iâm going to have the worldâs first robotic squirrel-impersonating bumper car. This way I can make up a smaller voltage pack just for the drive if I want.
You wouldnât happen to be referring to my motherâs driving practices, would you?
About the tatoo⊠I should think that the mandatory soldering-iron branding of âLynxmotionâ across your forehead would suffice.
Wait a sec⊠did you forget to read about having to do that in the fine print?
Now thatâs just weird⊠LOL
Note to self:
When doing soldering-iron tattoos, donât try to do it yourself in the mirror.
Now Iâm hooked on⊠noitomxnyL!
Apologies for the above post, but I couldnât resist.
Weâve had a âsans-childrenâ weekend, so I havenât spent much time on the computer to post an update.
Iâve been doing the code testing with just my desktop PC to save on the battery of the onboard PC. (I can see that getting a better battery or a more sophisticated charger is going to be a focus later on in the project)
All Iâve done so far is a basic proof-of-concept. I took the SerialPort code that Andy posted, and some serial test code from the Atom manual, and set up a simple feedback test - type in a âbâ on the PC, and the ABB resonds with a tone and a Serout text response back to the PC.
Next step will be to modify one of the rover .bas files to get serial data instead of PS2 data, for example.
It drives!
More pictures are up on my build report page:
I added aluminum T-Track from Rockler to the outside of the chassis. The idea is that I can mount âstuffâ pretty much wherever there is track using t-nuts.
Also, I was right - it drives like a hyperactive squirrel. Iâm hoping that changing the crappy RC speedcontrollers out for something decent will help with that.
Wow, very nice, I love what you have done with the chassis! You do fantastic work. That panel looks very good for a first attempt. Iâm not sure how much better you could make it, but then again, a true artist is never satisfied with his work.
I think quite a few people have had problem trying to run their logic boards and the motors off the same supply. If you find that the logic supplied of the same supply start resetting or youâre seeing brownouts when the motors kick on, this is likely your problem and youâll want to segregate them. The issue is that when the motors kick on, they can draw a large surge of current. This can cause the voltage of the supply to drop below that required by the regulators on the logic boards. Just something to watch out for.
Some of that problem can be prevented by ramping the motors not just slamming them on full field. Lead dress is also very important to controlling where the significant currents flow. Is there any additional input filtering at the ATX DC-DC power supply input? A couple of ohms in series with a several thousand uF of capacitor across the DC-DC input will do wonders in helping prevent motor and servo introduced dips from tripping the power monitoring circuits into resetting your board. Fun stuff. 
The several thousand uF idea seems to be the easiest way to prevent this sort of thing. Thatâs something I did not think of.
What doesnât show in the block diagram is that I now have three seperate batteries - the SLA for the PC, a 7.2 volt pack to power the ABB and the SSC-32, and finally a seperate pack (of whatever size I want to try - last night it was 7.2 volts) for the drive motors themselves.
The RC speed controllers were about the cheapest ones I could find at Tower a couple of years ago when I was playing with 12lb combat bots. (Traxxas something-or-others)
Iâm not sure exactly the difference in the control methodology, but the response âfeelsâ nothing like a couple of motors connected to something like a Scorpion. Even with verrrrrry gradual application of âthrottle,â the motors come on pretty much all at once. That and theyâre just not as responsive. When I tested this setup (Joystick>PC>LAN>PC>Serial Port>SSC-32>drive) with the hacked servos, it behaved much better, so Iâm hopeful that decent speed control will help.
BTW, any recommendations for cost-effective speed control that can handle those drill motors?
Perhaps MOSFETS are for you.
homepages.which.net/~paul.hills/ ⊠llers.html
Thatâs an incredibly comprehensive site made by a Battle Bot guy.
Building your speed controller from scratch is but the tip of that siteâs iceberg.
I took the lazy way out and just ordered two HB-25s from Jim.
With those controllers, I could probably upgrade to automotive starter motors if I eventually want to turn PC Bot into a tank-tread outdoor bot.
I wish I could have seen where technology would end up back when I was little and my dad helped me build a ârobotâ out of tin cans, solder and welding-rods.
