A Homemade Servo Driver Input Request

[teilhardo]

To all the EE geniuses out there, I have a request for information,

Do any of you have any knowledge into circuit design of servo motor drivers. My dream is to have a system that uses servo motors instead of primitive stepper motors but the limiting factor is cost. Geckodrives are to expensive for my budget. I have been looking on their yahoo groups for a while now and finally understand part of what they are doing for the G2002 project but I was chagrined when I found out that you still need the drivers that cost >$100 each.

What is involved in the EE design of a closed loop driver? PICS? STAMPS? Something fancier?

I would never want to take any business away from geckodrive because Mariss is a real nice and knowleadgable guy and I would like to thank him for all the help he has given to this CNC community. I just can't afford his products and I would really be interested in how those little blue and black boxes work.

Thanks,
Tei

[ESjaavik]

What is involved in the EE design of a closed loop driver?

More than can be condensed into a few posts!
If you don't have a lot of time and effort to sink into this, I would advise you to save your money elsewhere. There are a lot of theoretical knowledge and practical experience behind those boxes. I have some years in electronics and microprocessors, and your thoughts have crossed me too. I found out it's nice to know what's going on inside. And in principle not very complicated. The more I learned, the more I realized it's best left to those willing to invest the time to make the theory work in practice.

But by all means seek the theoretical knowledge. It will help you applying them in your tool. And it's always fun to know what's happening. I would compare a servo to balancing a rod in your hand. Your eye detects any tendency for it to lean over, and you move your hand to counteract it. If it leans hard, you move your hand more. If you overreact, it will lean over to the other side. If you always overreact, you will eventually loose control (oscillation). Try it. Then when you think about how well tuned the human brain and movement is, you'll understand it's not easy to mimic using a microprocessor. In this example your hand and mucles are the servos, your eyes are the feedback device (encoder or resolver), and your brain is the processor.

A stepper system is more like gripping the rod. If you have enough muscle force, you can keep it where you decide without any need for constant adjustments.

[tpworks]

There is alot of info about motor control at this site.

http://www.electronics-ee.com/Electr...ical_motor.htm

[Stevie]

I don't know why you would not entertain steppers; I've just got my tiny lathe moving and I can't believe the repeatabilty; once I got my steps set just right to match the threaded rod leadscrew; I can move 2-4-6-8" and be dead on to my gauge block set every time; and with the teflon buttons to push the screw into the nut as it wears i can almost forget about backlash
I'll still be looking a a better leadscrew setup; but even making it with the lathe would be possible with a spindle encoder

[teilhardo]

Don't get me wrong. I am very happy with my NEMA 23's, but they do have their limitations. For example: if I want to make a large router, I would have to go with NEMA 34's and they are pretty large and grossly ineffeficient not to mention they get to be pretty expensive.
With servo motors, you can add your own encoder plus you have the motivation to learn about how to make your own driver! Thats mostly why I want to build one, to get the knowledge of the concepts and save some $$$ along the way.
Thanks for the help,
Tei

[vacpress]

tei - there was mention of a DIY servo controller in the painter threads.. sdfine i think was the user.


he said the algorithms are esay to implement, and well documented.. he has written code for an atmel and has offered to share it.

i was surprised when he said it was easy.. maybe hes some kind of guru. i was under the impression that accurate closed-loop motor control is difficult to implement.