Precision of Sub-Micro Servo 3.7g (Generic)


Now my device is fully functional and works with a Sub-Micro Servo 3.7g (Generic) and maestro controller I have the problem that the resolution is bad and there’s slack on the servo when it starts in opposite direction of where it came from. Because of this the positioning is not correct.

The controller can be set with a resolution of 0.25 us. I measured that 13.3 us is 1 degree rotation. But the servo only starts moving at value 1040 when the position value is increased per 10 from a startvalue of 1000. It arrived from the opposite direction. meaning there’s a slack of about 30 us or something like 3° mechanical rotation. For my optical system it means double: 6° stepresolution. This is really bad: I aimed at 1° optical rotation, meaning 0.5°mechanical rotation. So the motor is not accurate (1000 and 1030 can be the same position, but not always) and not precise.

Which servo motor (given the fact I want to use my maestro controller with RC) does the job precise and accurate? This is a scientific device, so the resulting position MUST be the same if the set position value is the same.

Hobby grade servos are not meant for very accurate positioning, and especially not for scientific devices. Yours is functioning more poorly than I would expect, but you certainly cannot rely on 0.5 degree positioning. The feedback element that determines the position is a small potentiometer, which is noisy and not particularly linear or accurate.

You might try a digital servo with position feedback, like this one: … uator.aspx The manufacturer claims 0.29 degree resolution, but I would not count on it. The same manufacturer also offers digital servos with magnetic encoders, which are claimed to offer higher resolution and would be more reliable than potentiometers.

A high resolution stepper motor would be the best choice. It is hard to find one with 0.5 degree resolution or better, so a standard 200 step/revolution, 1.8 degree stepper with a toothed belt reduction drive would be the cheapest option, or use it in half step mode if 0.9 degree resolution is acceptable.

Note that although microstepping motor drivers are readily available, they are also not meant for accurate positioning, and often skip steps under load.

Hi Michiel,

Jim is right but I think the key is to approach every position you set from the same direction and if possible the same distance.

My chhhhheap generic sub micro servo has a repeatability of better than 0.5 degress. The non linearity is much worse, especially at both ends of its range. Varying load will have an adverse effect as will ageing of the servo. If you want (and leave your email) I can send you my results. Can’t upload a spreadsheet or photo as a comment.

I did it kind of scientific. Turning a long toothpick over a black cloth in steps of 100us. Taking photos from above and have some software calculate the angles. Repeated this 10 times.

I’m curious how you arrived at your results.



Another option is some of the microgear motors with the magnetic encoder, and you will need to processor the inputs and add some sort of control algorithm to turn the power to the motor into the target location .Can be very repeatable though.

But that’s only if you have someone or are yourself dedicated to spending a lot of time on this.