Very high repeatability and steps


I want to make a pan/tilt system of sorts in which I can have saved positions and revisit them with absolute repeatability.

Initially I turned to servo motors but I need 360° rotating and from what I could understand on the shopping sites the steps are not that great for high accuracy, then I found a 30:1 worm drive gearbox which looked awesome but then the repeatability goes down the drain since from what I understand repeatability in DC/gear? motors are pretty difficult and even then with errors, although I considered adding a gyroscope to it to save the positions since the system will be stationary but it seems a bit raw and too complex for now, so I'm turning to help.

Is there a solution for absolute repeatability on 360° rotation and with steps less or equal to 0.1°? I'll be dealing with fractions of degrees so the more steps the better. I don't care a lot about speed, minimum speed could be a full 360° rotation in 2 minutes.

I appreciate all the help and guidance,

Have you look into using stepping motors?

What about using a limit switch?

Hey Larry, I previously briefly read about them and one thing that turned me down was that on microstepping their accuracy may suffer and their steps are 0.9°/1.8° but I redid a search just now and found something interesting.

A geared bipolar stepper but not sure if it can handle 2kg of vertical load. Also, even though the step angle is awesome 0.018° (although curious since any other stepmotors I found had 1.8° and 0.9° steps) like all other stepmotors they have step accuracy error of 5%, on a single move that maybe quite low, but won't this error going to build up if the motor keeps running for a long time?

@sakkie, didn't know about such thing, will do a search now :slight_smile:


Not sure where you’re getting that “step accuracy 5%”. A 1.8o motor is going to be 200 steps per revolution – and always 200 steps per revolution – so long as it is within its torque limits (nobody jammed it and caused it to lose a step or manually turned it). Perhaps if you were running it in a half stepping (400 steps/rev) or quarter stepping (800 steps/rev) mode you might get a 5% accuracy, but then you’re talking .045o or .022o (1/2 or 1/4 of 5% of 1.8o). Still it will come back to zero each full rotation.

When you’re building a pan/tilt mechanism you typically do not put the load directly on the stepper’s shaft but rather use a belt or gear arrangement to transmit the power, and when doing that you have your opportunity to adjust the ratio of rotations of the motor vs rotations of the platform. Stepper motors typically have bronze bearing shafts and aren’t designed for sustained load. has lots of examples of how it can be done.

A gyroscope tells you rate of rotation and never reports an actual position. For that you need a rotary encoder.

Micro stepping dual shaft steppers + rotary encoders should give you the accuracy and repeatability you are looking for.

CNC machines have used steppers, lead screws, backlash nuts etc. for some time.
They are capable of absolute repeatability.
Well I guess you define what is absolute :wink: .

CNC machines have used steppers

Industrial CNC machines typically use DC servo motors with high-resolution encoders to count shaft rotations - you only typically see steppers on hobbyist CNC setups (it mainly has to do with the high torque you can get from DC gear motors, along with the higher speed - both musts for industrial fabrication, not as much for hobbyist setups).

you only typically see steppers on hobbyist CNC setups

That’s what mine is, .0005" resolution.

Like its mentioned above "Absolute" accuracy is relative term. You need to determine what is acceptable especially within your price range, since no machine will be absolute. However, by using a closed loop system, you can remove a lot of the error from a system.
Another option would be to have an tilt sensor on board. Use an accelerometer to measure your current tilt, and use this input information as feedback for the motors.