September 26, 2013 at 12:08 am #648
I have been looking at building an XYZ table for various purposes for years and something I have always struggled to deal with is open loop stepper motors on the axes.
At uni I had it beaten into my head that open loop is bad, closed loop good. Yet it seems most people are using open loop stepper motors in their “precision” CNC gear, fewer are using steppers with feedback encoders and even fewer are using servos. Now I understand that servos especially BLDC are expensive but feedback from steppers seems like a good idea. Is there a reason why people are not using these in their systems and are there any good closed loop motor controllers available? Or were my lecturers living in what i call “Happy physics land” where gravity is always 9.81, friction and wind resistance are negligible and everyone uses closed loop control.September 27, 2013 at 5:08 am #651
I can’t really talk to much about CNC specifically, but my thoughts are it probably comes down to the cost of an accurate, low noise encoder with the resolution required and to be able to survive in an industrial environment.
Stepper motors, from a controls point of view are a mechanical closed loop system unto themselves (i.e. within bounds they will move to and hold a position with external disturbance). However, obviously there’s no global feedback should the limits of the steppers electromechanical capabilities be exceeded.
The reality is steppers and a stepper driver are a pretty robust and cheap option these days. For a few hundred you can get three NEMA23 motors and a driver board. By the time you add encoders and feedback processing I can see that cost increasing significantly.September 27, 2013 at 11:54 am #652
Derryn, post: 506, member: 116 wrote: The reality is steppers and a stepper driver are a pretty robust and cheap option these days. For a few hundred you can get three NEMA23 motors and a driver board. By the time you add encoders and feedback processing I can see that cost increasing significantly.
Derryn pretty much nailed it from a stepper perspective.
The thing to remember is that, for a given job, a stepper system will either NEVER lose positional accuracy, or ALWAYS lose positional accuracy. A servo system only reacts when there’s a position error. If you set your stepper system up properly it can (on a small machine) be more accurate than a servo system.
The flipside is that servo systems require tuning. On a small machine where the mass of the workpiece is an appreciable fraction of the mass of the moving axii, this becomes problematic and you see problems with instability. If we put a 500g roll of solder on the end of the Arm-o-matic, then we have to retune it..
We’ve built quite a lot of CNC machines (ie >10) and a good rule of thumb is: For axii requiring <100W of mechanical power, steppers are fine. Above 100W you should look to servo's. But even this is flexible. Our big router runs about 300W mechanical on Y with two steppers…
Happy to give you the tour of the home shed..
PKOctober 10, 2013 at 11:24 pm #722
Very interesting to hear.
I have been looking at options for axes as well lately and I must admit i had always held the position that ball screw > rack and pinion however for a machine capable of taking a full size sheet there is apparantly a lot of problem with the ball screws “whipping” and forces you to slow down. Although one interesting idea is to use a fixed axis and rotate the nut instead. A prospect that normally appeals to my sense of do everything differently for completely pointless reasons.
It certainly sounds like steppers and missed steps are not as much of a problem as I had previously thought, thanks for the great info guys.
You must be logged in to reply to this topic.