Subject: Nashua Robot Builders's March Newsletter

Nashua Robot Builder's March '97 Newsletter

HI ALL!

N E W L O C A T I O N ! !
April 12th will see us at http://www.cyberjo.com/ here in Nashua, NH where you can surf the net and meet the 'bots. We'll still meet every second Saturday at 1pm but now its a (gasp!) public place! Jo has a bunch of comfy couches, nice chairs and a coffee and sweet bar so you can Jolt yourself into cyberspace.

Its free to sit and chat, food is reasonable and tasty.

N E X T M E E T I N G
Next meeting will finally be our IR beacon finder warm-up. I'll go over the rules again and have sample IR beacons to hand out for practice.

March was a good month for us, we got a new member and a couple of regluars showed up too. We all had a good time talking about robots as usual but I think Ron Beck really made the meeting memorable by getting me and Jay Francis excited about being engineers.

Ron has a lot of experience with being an engineer and also with running a company which he shared in quantity with Jay and I. Alot of what he had to say was something like:

"Theres the World guys; go get it; here's how I did."

Ron got us talking about how to run a business, how NOT to do it, and how to deal with people sucessfully, which is most of what it takes to run a business. Ron should know, he's started a few and done quite well.

I did a fair bit of talking too about speed control loops for 'bots. I made my bot do some speed control before and after _alot_ of tweaking got it working pretty well. The whole trick is how you get the speed feed-back sensors going and getting differential drive wheels syncronized. The trick is to get several things right:

   Lots of shaft encoder clicks per wheel revolution, the more the better.  Make one wheel the "speed control" wheel and make it match the desired velocity. Make the other wheel constantly match the existing velocity of the first wheel. Then it goes STRAIGHT and you've broken
the problem into two parts. Part one is getting one wheel going the right speed. Part two is matching that speed with the other.

PID loops dont work well near zero speed or at the top of their range. This is because they're constantly trying to overshoot or undershoot the setpoint and cant do that near zero or the top of their range.

Ok, enuf is enuf.

Next meeting is Sat, April 12 at http://www.cyberjo.com/ . mailto:cj@cyberjo.com or 603 889 7642.