Choosing a stepper or servo for a small positioner, low vibration

Reddinr

H-M Supporter - Diamond Member
H-M Lifetime Diamond Member
Joined
Jan 5, 2014
Messages
994
Attention servo/stepper experts.

My wife has gotten into macro photography. One of the things she's been thinking about getting is a stepper-based focus rail*. They can cost as much as a decent hobby lathe... Well, I happen to have a small 1mm pitch linear stage that should be just the ticket and I know how to make a controller/user interface for it. So, I'm wondering what sort of servo to put on it. Most commercial rails use steppers but from what I have read, they can suffer from vibration if you try to take shots while it is moving. I need to be able to control the rotation within 1/20,000 of a rotation. A 200 step x 128 or 256 micro-step stepper/driver could work. But, I also have a small BLDC motor that I could rig up but I don't know what sort of precision positioning control I can expect with that. My thinking is that it would be less vibration. I'm thinking ~Nema 17 size give or take. The positioner stage has about 100mm movement.

What kind of motor should I be looking at? Are there types of steppers that inherently vibrate less?

The other question I have is about the coupler. I want to couple the least amount of vibration from the motor. The torque will be pretty low but I need zero backlash. What are good choices here?

* Used to improve the images by automatically capturing several (or many) images at different focus points then using software to create a really sharp image with a big depth of field.
 
For no vibration smooth movement, the choice is servo. In this application, you only need a low power type.
Steppers are just fine once they have stopped moving, which may be OK for you. Steppers are, of course, way cheaper!

On 1mm pitch leadscrew, and 1/20000 rotation, that is a focus setting tolerance of 50 microns. I know the field of view for microscope optics is narrow, but that is extreme!

There are lead-screw and ball-screw solutions that can provide zero backlash. The tricky part here is the feedbacks. A motor shaft can be provided with (say) a 2048 pulse encoder. The A and B outputs are 90° out of phase, giving 4096 states. Then, using both the rising edge, and falling edge of the encoder outputs, the resolution is 1/8192 of circle. This is not enough for you. The next step is to use the motor shaft encoder only for acceleration, deceleration, and feedback stability, and rely on a direct outer loop encoder.

The direct and low cost outer loop positional feedback can be had from a small length linear DRO glass grating. (eBay - and inevitably Chinese). Even the lower resolution 5 microns type is a factor of 10 more than you need, and is way better than setting a start point calibration each run, then counting motor shaft pulses up and down for positional control. If the glass scale is too long, then simply shorten it, as many do when mounting them onto machines.

No gears between the motor and the leadscrew! Stepper motors vibrate and overshoot steps when on the move, but I think they stay steady when stationary. Servos are the best, but more expensive. They can "hunt", oscillating about the set point if the gains are set up incorrectly.

[ Much of the above is borrowed and adapted from a previous project. I made lots of mistakes before I got it right :) ]
 
Lots of industrial servo options. Have one servo motor on my desk. 50W, size of a Nema 14 motor and is fitted with a 24 bit (16 million) count Nikon encoder. Now you pay $$ for that, but there are options out there.

Gearboxes help with resolution, especially if your speed requirements are not high, but that adds a lot of cost - often more than the motor.

I will warn you that microstepping a stepper is finacky. At each increase in microstepping you lose a tremendous amount of torque. Eventually, the natural friction in the stage and even motor bearings yields no improvement in positioning resolution beyond a lower microstepping value. Servos are different and can be considered to have a flat torque response at any position or speed.

If I were designing a high performance stage, I would select a direct drive servo, a bellows or disc type servo grade coupling, and an antibacklash ACME screw (spring loaded double nut).

DMM technology offers a 50 or 100W AC servo motor with a 65536 count encoder. Paired with their low voltage drive you get a servo system meeting your needs for a bit over $300. Not cheap, but good motion control is expensive. The drive can accept serial commands if you wish to run it that way.

Teknic Clearpath offers servos with built in drives and encoders. The enhanced version only offers 6400 count encoders unfortunately. An 86W motor comparable to the DMM above that accepts RS232 program commands would run $375.

For comparison, a true industrial servo system like the Delta ASDA-B3 servo system or an Allen Bradley Kinetix 5100 would run you about $2000 for a kit of drive, motors, and cables. These are the drives offering the 24bit encoder. And these are still considered entry level.

Most servos will autotune and not hunt.
 
For the couplings, I'd personally go with a bellows coupling (sample below - choose one to fit your motor and screw)


Or a servo grade disc coupling (again only a sample)


These offer the highest torsional stiffness of all motor couplings with zero backlash. Other types (jaw, helical beam, oldham, etc.) have much worse torsional stiffness which makes a difference when going for precision.

Finally for the screw, I'd look at a small precision ACME screw and zero backlash spring loaded nut like this one offered from Roton



Ballscrews in this size of any quality will be extremely expensive and this screw should work fine for your needs.
 
Wow! Very helpful and quite a bit to think about. I may compromise at first and try a stepper with the caveat that the motor moves and stops prior to the image. That is a compromise but not a big one. Fortunately, backlash should not be a problem as the movement can be in one-direction during the shoot. The other helping factor is that it is the absolute position on the rail is not important but only that the distance between stops is consistent. Since the commercial units are stepper based, I think I will start there and improve if needed. The linear DRO is an excellent idea even if just for temporary use to make sure I'm not losing steps.

I'll have a look at the DMM Technology product too. I am aware of clearpath product and have a couple (and like them alot). Their smallest is nema 23 which is heavy for the application. I have the actuator assembly which is a THK precision screw assembly, I believe it to be of "C3" precision (accuracy?) or at worst "C5".
 
I'm a stepper rookie but an experienced photog.
I think the challenge with a stepper will be the 'buzzing' that occurs as a result of the motor holding position. My gut says that unless there is zero load on the motor the motor may vibrate as it's holding position.
Happy to be educated on this...
 
David. I have a driver where I can rig it up to back off on the current at will if needed. Because of the fine screw pitch/friction and tiny load I don't think I will need much holding power so I think I can get around that potential problem. I need enough hold to not lose steps and that is it.

Years ago I found some Parker/Compumotor OEM750X indexer/Drives. All I need to do is to send it commands via serial and it does the rest. Lots of tuning capability etc. I think I bought a dozen of them for $30 total. I've used two and gave a couple away. I guess I'm itching to put more of them to use.

Did you do any macro or extreme macro stuff?
 
So I wanted to provide info on the servo side since I love it (and work professionally in the industry). You can also get nice performance from steppers, especially nice quality ones. Probably more important than anything else is providing the motor drive with a smooth command with minimum Jerk (derivative of acceleration). The wobble of the camera will probably outweigh any vibration from the motor. Smooth "S-Curve" move profiles (if your drive supports it) will help reduce any camera shaking after the motor stops.

Give it a go and see how it works!
 
David. I have a driver where I can rig it up to back off on the current at will if needed. Because of the fine screw pitch/friction and tiny load I don't think I will need much holding power so I think I can get around that potential problem. I need enough hold to not lose steps and that is it.

Years ago I found some Parker/Compumotor OEM750X indexer/Drives. All I need to do is to send it commands via serial and it does the rest. Lots of tuning capability etc. I think I bought a dozen of them for $30 total. I've used two and gave a couple away. I guess I'm itching to put more of them to use.

Did you do any macro or extreme macro stuff?

I do a bit of macro work with flowers. I’m a real fan of the extreme macro work.
 
This might be one of the few times where a belt drive from the servo to stage would be helpful. Specifically, a round belt like those used to drive grinder spindles imparts no tooth cogging. Assuming an outer-loop linear encoder is used for positioning belt slip would not be an issue.

Id look for an ebay servo drive with dual encoder input, and a nema 17 (ish) servo with fairly high count encoder. Id also tune the inner loop (motor encoder) to be a bit tolerant so theres no chance of hunting while stopped.

You might even find a indexing drive (like the Parker mentioned earlier) that requires no external motion controller. Simple switch inputs can move the stage predetermined distances and re-home after a series of photos.

An air bearing stage woud be choice here, but unless you want to go down a very deep rabbit hole i think you'd be in pretty good shape with a small servo, linear encoder, and dual-loop indexing drive.
-Ralph
 
Back
Top