Using induction motors is tricky and potentially expensive as the drives are not cheap. You can get VFDs that are specifically designed to work as servo drives. They are usually called phase vector drives. Ordinary VFDs have very poor low speed response and it is difficult to get accurate positioning. Generally it only makes sense for the big stuff. If you need to move a 10 ton satellite dish then induction motors work well (I have the bits that came out of a setup like this when it was decommissioned).
For the small stuff I would suggest that you stick to DC servos. If you don't need the ultimate in speed and acceleration then an ordinary DC motor with an encoder attached works well. Add a DC drive such as a Gecko G320X and you have a surprisingly cheap and effective servo setup. When looking at DC motors the main criteria is smooth running. If you can get them, treadmill motors work really well. Les On 19/01/12 00:40, Leonardo Marsaglia wrote: > Hello to everybody on this list. > > I've been working for almost two years with LinuxCNC software (the old EMC > hehe) and steppers driving an old Schaudt cam/lobe grinding machine. Now > it's time to improve the system and change the steppers for something with > more speed, specially for the Z axis. Anyway, the idea is to keep the costs > as low as possible. That's the main reason to not to use servo motors. > > A year ago more or less, Andy Pugh showed that it's not a crazy idea to use > a conventional three phase induction motor and a VFD in closed loop with > LinuxCNC software for positioning. Anyway, Andy told that his VFD was old, > and it couldn't respond well below 200 rpm. Nevertheless, the experiment > that Andy did, was really good and impressive. Here is a video of that > http://www.youtube.com/watch?v=3oTJNEVpvYY. > > My question is, what do you guys think, about using an induction motor, > reduced with worm and gear and driven by a vectorial VFD, (also there are > newer ones that apparently can keep full torque at zero speed), to drive a > joint and control position with minimum increments of about 100 pulses of > encoder (I'm planning to use an encoder of about 1000 pulses per turn). So > this would be 1/10 of a motor turn. > > Do you think this is achievable taking all of this in consideration?. I > mean, the reduction with worm and gear reduces the inertia factor almost to > nothing, so the deceleration of the motor would be practically > instantaneous. And also the motor would need to turn mostly at high speeds > because of the worm, even better if a use full torque at zero speed VFD. > > I hope you can help me with this, since the idea is not to risk too much > spending money on solutions that can't work at all. I'll ve very thanked if > you could guide me with this replacement. > > Thanks in advance as always for your help! > ------------------------------------------------------------------------------ Keep Your Developer Skills Current with LearnDevNow! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-d2d _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
