What is better a stepper or a servo motor?

[tcr3670]

Can anyone explain which motor would have more torque and accuracy?

Thanks

[Switcher]

What is better a stepper or a servo motor?

servo !

.

[NC Cams]

There is a place on this M/B that has answers to the most frequently asked/answered questions (FAQ). Have you checked there yet as this and many other stepper/servo questions have been asked and answered???

If not and don't have the ambition to do so, try this link:

http://www.cnczone.com/forums/showthread.php?t=17419

Answer: servo

[rexstep]

Ok, let's be fair because the question is application specific. For example, you can say that servos are usually better for high speed routers. But I would always select steppers where low speed torque is critical and high speed is for rapid traverse, i.e. plasma cutters and small to medium size milling machines.

So, which is more powerful? It depends on the speed range you need torque for. I would add that steppers don't burn up when they stall.

As for servos being more accurate it's simply not true in most cases. Remember that servos constantly HUNT for their position, always over and undershooting ever so slightly. Steppers on the other hand HOLD in rock solid position until told to move and then they stop on a dime unless software parameters are poorly set.

For an excellent overview of the two take a look at the page that deals with this on the Torchmate website. It's right on the money.

-RL

[NC Cams]

"...As for servos being more accurate it's simply not true in most cases. Remember that servos constantly HUNT for their position, always over and undershooting ever so slightly..."

I disagree that this is endemic to ALL servos and it does NOT ALWAYS HAPPEN - it is a function of tune and operation.

We have some servos tuned on our BPT mill that can hold position to withing 0.0001" and do it regularly - they don't overshoot if/when we properly program them for the desired accuracies.

We also can mill model cams for our cam grinder the hold shape to within parts per MILLION in accuracies. It takes time to do the mechanicals and more time to do the proper tuning but it can be done - that is, get them to run smooth and accuratedly WITHOUT "always" overshooting and hunting as the esteemed member contends can/will happen.

"...Steppers on the other hand HOLD in rock solid position until told to move and then they stop on a dime unless software parameters are poorly set...."

We know of folks who've tried to apply steppers to CNC cam grinding and OD grinding machines. Since they "step", they can't help but dupicate their "steps" into the surface finish of the part - even when holding a constant diameter when they should be "rock solid".

Of all of the CNC cam grinders that I've seen, the only ones that create mirror like, step/chatter free finishes are those operated by servos - which ALL high end grinders are fitted with BTW.

Yes, the maladies noted ARE found in servo and stepper systems but these are definitely "application specific" - as a friend once said about the help he hires, "... you need to find the right guy for the right job..."

Steppers and servos respond NO differently.

[rexstep]

I don't think we're in disagreement necessarily. Remember that my comments are in responce to the other esteemed member's blanket statement that servos are better.

0.0001" is a very tight responce but I'm addressing the "dither" characteristic intrinsic to all servos which is that even in so tight a response, it had to pass the exact correct line on the encoder before backing up onto it. Further, unless I'm out of line here, there's do device in the closed loop system to LOCK the encoder on that line so that even at .0001" axis travel (and that is axis travel) the hunting must still be there even though it is so small that it's not an issue. If there's a servo system that has absolute "zero" dither, I will readily recant from this doctrine which I learned from a very good source at Geckodrive Inc.

Now, what you say about the stepper's unsuitablity to grinders makes perfect sense to me and I'd like to reiterate my original that the choice between the two is application specific.

For the record, in pointing to some stepper advantages and some servo disadvantages I did not make the claim that steppers are superior. No offense meant.

Kind Regards,

-RL

[Al_The_Man]

Originally Posted by rexstep If there's a servo system that has absolute "zero" dither, I will readily recant from this doctrine which I learned from a very good source at Geckodrive Inc.

I am positive I get this result with the servo systems I have been using, namely Galil and Acroloop motion cards.
The output of the control is ±10v analogue, this feeds a transconductance type PWM amplifier (torque/current control), the motor current is directly proportional to the analogue signal voltage.
If the encoder reads that it is in the desired/command position, then the command output is zero, therefore motion is zero.
If I check the in-position error and it is zero then the command is zero.
Al.

[imserv]

Dithering is a problem with resolver based systems much more than optical encoder based. A modern servo system will have an optical encoder feedback with much finer resolution than the machine requires. The control system has a deadband within which no control motion is needed as the system is considered in position. The deadband might be somewhere between .001 and .0001 inch, depending on the machine. As long as position is maintained within this accuracy, no controlling power is needed or applied and no dither is possible. Freewheeling Z axes may present a problem with dithering, but safety regulations would also require that a free wheeling Z axis have a deadman brake to prevent crushing an operator if power was disconnected. It becomes less expensive to provide a balanced counter weight. A counter weight won't provide a constant load that pushes the encoder outside the deadband. A safe machine shouldn't dither either.

To properly answer the question, the post from tcr3670 should ask about a specific stepper and servo motor. It may be best to show some examples of what can be expected from steppers and servos, which might help others.


Typical inexpensive stepper is 200 steps per rev. Can be accurately positioned in CNC applications down to 1/2 step, which means the available resolution is 200 or 400 steps per motor rev. Most systems use direct drive to take advantage of the low end torque available.

Servo resolution is a function of the encoder slots. Inexpensive encoder wheels are available with 32-2500 slots. The quadrature resolution is 4 times those numbers, ie typically 128-10000 counts per encoder revolution. Servo motors are usually belt reduced by 2:1 or 3:1 ratios, increasing the resolutions to 256-30000 counts per leadscrew revolution.

We use 500 line encoders and 2.8:1 belt drives to provide 112,000 closed loop steps per inch on a CNC ready Sherline or Taig mill. Typical stepper systems provide 4000 or 8000 open loop steps per inch. Maximum usable speeds of 25-30 ipm with steppers and 45-50 ipm with servos. On our routers we drop this down to 28,000 steps per inch and run at 175 ipm, using the same servo motors, drivers, and control system.

It should be clear that stepper resolution is fixed with respect to motor rotation, but that servo resolution is a function of the designer's encoder wheel selection. At no additional cost ( the slotted wheels are all the same price whether 32 slots or 2000, 2500 slots does add an extra $10 to encoder cost) , it can be an order of magnitude better than a stepper.

Both steppers and servos are available in a wide variety of sizes and toques, up to about 1100 oz-in. There are few steppers available over 1100 oz-in. Since just about any non-cogging, PM DC motor can be used for a servo, there is a wide selection of both new and surplus motors available for use as servos. An example of this is the ubiquitous treadmill motor with 1/2 inch shaft, which shows up in nearly all the surplus motor shops for around $20 (or in garage sales for $5, but attached to a treadmill).

Fred Smith - IMService
http://www.cadcamcadcam.com/hobby