I have the high-torque servo hooked up to something that requires a bit of torque (microscope coarse focus knob). When I send a move command, if I move it more than 5 units or so, it works correctly. However, if I move it, say, 1 or 2 units, it doesn’t work and it sits there making a noise like it is trying to move, but can’t. I can move it slightly in the direction that it should move, and the noise goes away.
However, if I take it off of the focus knob, it moves 1 or 2 units fine (with no load). So, obviously, there is almost no torque available when moving only a small amount. Is this normal? If I want fine movement (like move forward 2 units) do I need to move forward 7 and then back 5? Is there a better way around this?
That is normal operation for a standard servo. You may want to look at high torque servos. They cost more but will do what you need. Another option may be to use a stepper motor.
chaz:
However, if I take it off of the focus knob, it moves 1 or 2 units fine (with no load). So, obviously, there is almost no torque available when moving only a small amount. Is this normal? If I want fine movement (like move forward 2 units) do I need to move forward 7 and then back 5? Is there a better way around this?
What is a unit ? Servo's respond to pulsewidths but usually have some deadband so as to stop it from continually hunting for position. You may also be finding that the loop gain is too low for your application. Some (digital) servos allow you to tweak these parameters.
Are you looking to automate the focusing or just have remote control ? How much slop/backlash is there in the system now ? Do you have an estimate of the torque needed to move the assembly ?
I wanted to automate focusing with stops at repeatable values. I thought a servo could accomplish this, but it looks like maybe a stepper motor is more of what I need.
codlink:
You are referring to degrees when you say unit.
I quickly read the manual and I still am unsure. I'll guess that the OP is running some program that talks to the controller which then commands the servo via the normal pulsewidth method. The spec for the controller mentions 7 bit and 8 bit commands but with variable scaling and no default that I could find. So I might ASSume that the default resolution is the listed 0.5 us and then ASSume that corresponds 1 "unit" on his control program. If that's the case then he might be running into a problem with the deadband of the servo, given it's listed as 5 us. But his symptoms don't follow that exactly. Then again I'm going by his verbal description and not running the test myself.
Without some details re: the above settings (PW, scaling, servo #, 7/8 bits, ?) and the mechanical interface between the servo and 'scope, I’m not sure if there’s much I can do. A thought that occurs to me would be to put the feedback onto the scope motion and not the in the servo. Depending of the mechanics and slop in the present system, that may (or may not) help.