Choice of motor - application query

davezawadi (David Stone):

I think you have never tried to design a servo system Z, the problems very quickly can overwhelm the concept. Let's say an encoder with 100 positions around the shaft and your motor as advertised. How long do you have to stop the drum at the right position before it rotates past it and overshoots? You have some proportion of 3.6 degrees, which is not a lot given any inertia. If it does overshoot can it turn back? How do you stop it oscillating back and forth? A real feedback system will have several controls, the position (proportional to the rotation required), an integral term to make the servo get to the stop position if it "stops" a little distance away, and a differential term that controls the rate of speed change to be within the mechanical limits of the system. Each of these interacts, and getting optimum control from PID as it is called needs some measurements and a lot of maths, or a lot of luck!


The reason I suggested a stepper motor is that the mechanical movement exactly follows the pulses applied to the coils, and then it stops even if a tiny bit out of position. It has a large braking force when stopped, and can be made to have very similar characteristics even when getting smooth rotation. It does not need any feedback for this application and is therefore simple to implement, and therefore will probably work first time it is powered up. All the drivers and components for software drive are readily available.


Even mass-market products like laser and inkjet printers often use stepper motors, it is partly because there is little to go wrong, whereas a stiff bearing can make a full servo very badly behaved. I have had an experience of trying to find an electronic servo fault in a film camera with a duff bearing, the speed simply was not as stable as it should be,


The O.P says, 

1. The motion  profile is likely to just be one full rotation of the cylindrical object at a fixed speed, but there may be a requirement to perform the rotation in increments with pauses in between (i.e. one full rotation split into 6x moves). The duty will be low, i.e. 5-10 such operations per day.

Well, why would I design such a simple thing when one can be bought off the shelf.  The cylinder is to be up to 20kg in weight and rotate at a snail's pace for "inspection" whatever that means. Inspection of the cylinder I presume, or the contents, we have not been told.


It can be manually controlled as it rotates so slowly while it is being "INSPECTED".

The drive system must be powerful enough and durable enough to provide steady rotation of the cylinder which has a varying unstable load as its contents.


Just turn it by hand. Printer and camera motors are very tiny and will be unsuitable.


Z.

davezawadi (David Stone):

I think you have never tried to design a servo system Z, the problems very quickly can overwhelm the concept. Let's say an encoder with 100 positions around the shaft and your motor as advertised. How long do you have to stop the drum at the right position before it rotates past it and overshoots? You have some proportion of 3.6 degrees, which is not a lot given any inertia. If it does overshoot can it turn back? How do you stop it oscillating back and forth? A real feedback system will have several controls, the position (proportional to the rotation required), an integral term to make the servo get to the stop position if it "stops" a little distance away, and a differential term that controls the rate of speed change to be within the mechanical limits of the system. Each of these interacts, and getting optimum control from PID as it is called needs some measurements and a lot of maths, or a lot of luck!


The reason I suggested a stepper motor is that the mechanical movement exactly follows the pulses applied to the coils, and then it stops even if a tiny bit out of position. It has a large braking force when stopped, and can be made to have very similar characteristics even when getting smooth rotation. It does not need any feedback for this application and is therefore simple to implement, and therefore will probably work first time it is powered up. All the drivers and components for software drive are readily available.


Even mass-market products like laser and inkjet printers often use stepper motors, it is partly because there is little to go wrong, whereas a stiff bearing can make a full servo very badly behaved. I have had an experience of trying to find an electronic servo fault in a film camera with a duff bearing, the speed simply was not as stable as it should be,


The O.P says, 

1. The motion  profile is likely to just be one full rotation of the cylindrical object at a fixed speed, but there may be a requirement to perform the rotation in increments with pauses in between (i.e. one full rotation split into 6x moves). The duty will be low, i.e. 5-10 such operations per day.

Well, why would I design such a simple thing when one can be bought off the shelf.  The cylinder is to be up to 20kg in weight and rotate at a snail's pace for "inspection" whatever that means. Inspection of the cylinder I presume, or the contents, we have not been told.


It can be manually controlled as it rotates so slowly while it is being "INSPECTED".

The drive system must be powerful enough and durable enough to provide steady rotation of the cylinder which has a varying unstable load as its contents.


Just turn it by hand. Printer and camera motors are very tiny and will be unsuitable.


Z.