Servo Motor vs. Stepper Motor System and Components

[dscbob]

Hello,
I am designing a "pick and place" mechanism for a new project I have been assigned to. I am a mechanical engineer with very little servo or stepper motor experience. This motion system will be three dimensional (x-y-z) with motor (brushless servo or micro stepper) driven lead screws. I am not too concerned about meeting speed and torque requirements because it is a light duty application. The system has to fit into a 15in.(380mm) x 15in.(380mm) area. The positioning of the system should have an accuracy of +/-.001". What do I need besides the motor? I am getting lost in the info on-line....commutators, encoders, blah, blah, blah.
Any help would be appreciated.
Thanks!

~Bob

 

[MikeHalloran]

You may need a new job, if that accuracy spec can't be relaxed or redefined, and your budget is limited. Light duty robot parts will never work well enough. You'll need precise, sturdy, stiff, _steel_ slides and leadscrews, the kind sold as machine tool components, just so trajectory error and lead error doesn't consume your entire error budget. You'll need anti-backlash nuts, or maybe software backlash compensation. And you'll need encoders or precision linear transducers so you can run closed loop. Forget stepping motors; you need servos.








Mike Halloran
Pembroke Pines, FL, USA

 

[itsmoked]

+/- 0.001" = hard.

For servos look at "Gecko Drives".

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[jasongreat1]

The use of ball screws will eliminate backlash. Also +/-.001 = easy with ballscrews. I am new to this forum and have not had a chance to look over all the threads, but there is a DIY CNC group on one of those home pages(Y).I would bet that if you searched their archives for a couple of hours you could probably get all your questions answered.

 

[dscbob]

I have met with a rep from Maxon Motors to get more information on servo motors and their capabilities. Maxon informed us that the resolution of their encoders can control shaft postion to +/- .18 degrees. With this resoultion the tolerancing challenge lies in the lead screw - a fine lead (pitch) will give me better accuracy, the motor would have to increase it's velocity, but the tighter pitch should increase torque output. The downside about Maxon's product is that the servo motor motion controllers only accommodate one axis. I would need to have a motion controller for each axis (n=3). The other challenge I would be facing is that when this pick-and-place system is in its x,y position and ready to use the z-axis, to pick or place, is that the servo motor does not have a "detent" position or low "detent torque". A stepper motor has incremental positions which could require significant torque to "drift" out of place. How would I keep the servo motor from drifting in this position? I am looking into a gear head because the system would be operating at speeds below 5000 rpm. This gear head will contribute to a detent position and mightt actually increase the +/- .18 degree shaft position.

I did look into Gecko Drives, thank you, and they do have motion controllers that accommodate more thn one axis. They actually have motion controllers that acommodate up to 8 axis but the multi-axis controllers do not come cheap.

I am trying to get a "start-up" kit that I can use to create a breadboard for my testing. I want it to be a package that includes three stepper motors (w/encoders), a motion controller that accommodates multiple axis, accessories required to monitor/control the system through a PC using Lab View and most importantly - instructions for use.

Thanks for the information gentlemen!

~Bob

 

[MikeHalloran]

Ahem. The servo MOTOR does not have detent torque like a stepper, WHEN THE POWER IS OFF.

When the power is on, the servo SYSTEM ( = motor + drive + encoder ) has a LOT of detent torque; typically more than a stepper of similar size.

Go ahead and buy a stepper- based breadboard to screw around with. Plan to throw it away.



Mike Halloran
Pembroke Pines, FL, USA

 

[dscbob]

Mike,

Here is a cough drop for your throat...

So does that mean "power" never gets cut-off to any of the three axis? The servo motor receives a sinusoidal pulse width modulated signal from the driver which controls it's velocity. If this x-y-z pick and place system established it's x-y position it would have to discontinue power to the x-axis and y-axis motors in order for them to stay put. Do the drivers simply decrease the amount of current delivered to these motors or completely shut off? I am not that familiar with servo motors but the rep from Maxon seemed to think this could be a problem too. The servo motor's torque is based on how much current it receives (torque constant).

Do you represent a servo motor company? You sound like a salesperson of a servo motor company trying to steer me away from steppers. Can you get a little more technical than "a LOT of detent torque".

Thank you.

~Bob

 

[Skogsgurra]

No Bob, they do not reduce current. They reduce frequency to zero Hz. That keeps the motors locked in position with full torque.

Gunnar Englund

http://www.gke.org

 

[MikeHalloran]

No, I'm not a servo salesman. I'm a long time stepping motor user, actually. But I know a tiny, tiny bit about how servos work.

Digital PWM servos are used in model airplanes. The pulse width corresponds to the commanded output position, I think with 1.0ms representing the midpoint, and something like 0.5 and 1.5 ms representing the extremes of motion, usually +/- 90 degrees on a bellcrank. I forget what the 1/0 voltages are; I think much less than 5V.

Industrial servos use a DC voltage or a digital "word" that corresponds to the commanded position.

In any such position servo, if the pulse width or the voltage or the word doesn't change, the servo output doesn't move. You don't turn the power off to stop the output; you just stop changing the command input.

Within the servosystem, the actual position (as measured by an encoder or a resolver or a potentiometer) and the commanded position are compared, and if they are different, the output is driven in a direction that should eliminate the difference.

The servo doesn't know or care why they are different.
If you, acting as a random external force, use a stick (never put your finger in this kind of stuff) to try to push the servo away from its commanded position, it will fight you. The model airplane servo will fight pretty hard. A CNC machine's servo will feel like it's bolted in place.

You might want to rig up a PIC or a Stamp and a model airplane servo just for fun, feed it a steady train of pulses, and measure the supply current as you try to move the output.



Mike Halloran
Pembroke Pines, FL, USA

 

[mc5w]

You can also go with a direct current motor or pneumatic pistons or pneumatic motors for some of your servo applications.

If your application requires very precise positioning or a torque read current limit that you can adjust, direct current servo drives can also be used. Also, a DC drive can have less cogging problems at certain speeds. GEC Parvex can sell you a servo disk motor that uses a 6 layer printed circuit card for the rotor. This achieves 193 commutator bars in an 8 pole motor which achieves extreme smoothness. A other trick that you can do with a DC drive is that you can hook up multiple resistors and relays to the current limit so that you can get 160% of motor full load current to get fast positioning and then drop down to 80% full load current for holding torque. Just make sure that you drop from a higher value resistor to a lower value by paralleling resistors using form A contacts.

You can also program most DC computer numerical controls so that they do not get upset if it cannot reach final position because of a hard stop, such as certain kinds of press feeding operations.

If you decide to go with direct current servo motors instead of alternating current brushless servo drives bear in mind that at zero speed the motor can only develop 80% of nameplate torque. This is because at zero speed the unwanted current components from the pulse width modulation drive are at maximum.

Also, sa far as jaws that grab something, you should consider pneumatic pistons. You can get a pulse width modulated valve which controls the pressure of a small amount of air. This then goes to a volume increaser which is essentially a pressure regulator that uses the pilot air pressure instead of an adjusting screw. The advantages of compressed air is that a soft torque limit is easy to achieve such as when moving something that breaks easily or falls over easily.

Mike Cole mc5w at earthlink dot net

 

[djs]

The big difference (usually) betwen a stepper motor and a servo motor system is the in the servo motor the position is actively controlled in a feedback loop. In other words a position sensor of some type feeds a position to the servo control which then sends the correct output to the servo motor to move or keep it in the commanded position.

I would definately look into using a gear box between the servo motor and the positioning mechanism. Special gearboxes are made to be used with servo motors. They are not cheap, but most have zero or near zero backlash.

 

[IRstuff]

An off-the-shelf product with performance that exceeds the given requirements:

http://www.newport.com/store/genproduct.aspx?id=400011&lang=1033&Section=Spec

It's not cheap, but you can get 2 axes of 24-inch travel, plus z-axis of 12-inch travel, with controller for under $35K. The velocity and acceleration numbers seem to be pretty good, but the settling time might be high for your application, although it might settle faster with a 0.001 in tolerance.

If nothing else, you could buy one and play with it to get some concrete ideas for a custom design

TTFN