Torque Pulsations

[SPIGUY1]

I have a 24VDC BLDC servo motor that we are using for a force feedback application. One of the critical requirements is that there be as little cogging/torque ripple as possible. This is a 3 phase sinusoidal controlled motor.

We are testing a 12 slot/4 pole design now but would like to know:
1. What's better at reducing cogging - skewed laminations or skewed magnets?
2. Is it possible to wind a straight stack and then skew it? What particular design features should be evaluated - slot design, wire tension, etc.?
3. Would a 15 slot/4 pole motor design be better without the need to skew either the magnets or the stack? What degree of improvement could be expected between the two designs? Would a 18/4 be better yet provided that the slots could accommodate the wire (AWG #24 or #26)and the turns?
4. How accurate are 2D and 3D FEAs in predicting the possible cogging/ripple reduction?
5. Any good articles on reducing torque pulsations?

 

[electricpete]

I would think skewing the magnets would be effective.

I don't see what benefit at all comes from skewing the laminations.


=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Clyde38]

Skewed laminations and skewed magnets will work equally as well. Both methods are aimed at smoothing the variation in reluctance in the air gap. Magnets with a tapered raidus will also reduce reluctance torque (cogging).
It is possible to wind straight and skew later, but you probably don't want to do it. You might want to consider the length of the stack to the diameter and number of slots (is it safe to assume that your interest is in an outside stator?). Slot pole configuration will also reduce reluctance torque. There is however more to it than just the number of slots and the number of poles. You must also consider the pole arc length with respect to the number of slots. The idea is to make the change in airgap reluctance as smooth a transition as possible. FEA (in the right hands) can be quite accurate in predicting the airgap reluctance. Have you considered a coreless motor?
If I can recall any liturature about the subject before you have solved your problem, I will post.

[link

http://www.linkedin.com/in/clydehancock

]

[/url]

 

[electricpete]

maybe I don't understand what is meant by skewed laminations.

If the slots are skewed, I could see that could minimize torque pulsations. Is that what you're referring to?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Clyde38]

Here is one article. I'm sure there are many more. By the way, what is the size of this motor? One other thing to mention. Any thing you do to reduce reluctance torque will also reduce the desired torque. This can be compensated for.

http://www.docstoc.com/docs/2380520/Torque-Ripple-Analysis-of-a-PM-Brushless-DC-Motor-Using-Finite

[link

http://www.linkedin.com/in/clydehancock

]

[/url]

 

[Clyde38]

I'm not sure what SPIGUY1 meant. I did make the assumption that SPIGUY1 was referring to skewing the "lamination slots". I don't know what else to skew with the lamination.

[link

http://www.linkedin.com/in/clydehancock

]

[/url]

 

[Clyde38]

One more and I'll stop!

http://www.industrialcontroldesignl...ionid=TLC0ZX5QNF3PSQSNDLPSKH0CJUNN2JVN?pgno=4

[link

http://www.linkedin.com/in/clydehancock

]

[/url]

 

[electricpete]

I don't know what else to skew with the lamination.

The laminations are perpendicular to the shaft. If they were not, that would be skewed. But it has no logical purpose as you say. So from your comments I understand the intent is skewed slots. Thanks.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[SPIGUY1]

Yes, skewed laminations means skewed slots. Sorry if I didn't make that clear. The amount of skew is .86 to 1.0 stator tooth.
To answer another question - this motor is a NEMA 23 frame size and has a peak torque of 2.3Nm (through a gearbox) with an averge torque of 0.23Nm.

 

[Clyde38]

Sorry about the mis-communication on my part. I believe that it is more accurate to describe the skew as the "lamination stack". I have indeed seen the lamination slot/tooth skewed (not perpendicular to the shaft), and still have the stack of laminations not be skewed. This is not for reduction of the reluctance torque. It is common to skew the lamination stack on the armature of PM DC motors.
The size and torques required are well within the capabilities of a coreless (slotless) motor.

[link

http://www.linkedin.com/in/clydehancock

]

[/url]

 

[SPIGUY1]

Thanks for all the input. I would think that skewing the laminations would be easier to develop instead of skewing the magnets that would require more design work. Are there existing skewed magnets sized for the rotor body OD and length?

 

[cswilson]

While I agree with the others that you should do as much as possible to reduce cogging torque in the design of the motor (and a coreless motor has no cogging torque), I do want to point out that it is not that difficult to compensate significantly for the remaining cogging torque.

I know multiple applications where a simple table-based correction (as a function of motor angle) reduced the velocity ripple resulting from cogging torque by about two-thirds. To find the table values, they just noted how much torque was needed to hold position of the (unloaded) motor at each point in the table. In operation, this torque amount was simply added to the computed servo command.

Curt Wilson
Delta Tau Data Systems

 

[onemilimeter]

You may consider to use fractional-slot permanent motor with similar slot and pole number (e.g. 12-slot/10-pole, 9-slot/8-pole) to minimize cogging torque. But the unbalance magnetic force might be higher.

To cswilson:
Would you mind to share how to measure "cogging torque"?

Thanks.

 

[cswilson]

1mm:

Measuring cogging torque turns out to be very simple if you've got a digital positioning servo controller. I'll expand on my brief explanation above.

To hold an unloaded motor at a certain position with no error requires that the controller apply a torque that is equal and opposite to the cogging torque at that position. So you have the controller command that position and wait for it to settle at that position with zero error, then read the controller's servo-loop output value (which is the torque command if you are not using a velocity-mode amplifier).

You've got the value in the internal units of the controller now. If you want to convert that to physical units (e.g. N-m), you have about 3 system constants you need to multiply this value by. But typically you want the value in the controller's internal units, because now you can automatically add this value in to the active servo-loop command.

You will repeat this process -- of move, settle, read the torque command number, and put this number in a correction table -- for every few degrees of a motor commutation cycle or revolution. I have automated the process to be able to do it in less than a minute.

Curt Wilson
Delta Tau Data Systems

 

[onemilimeter]

Thanks cswilson. I never thought of that before. Your suggested method is very efficient. By the way, if we do not have a 'digital positioning servo controller', what other options we have to measure cogging torque?

Thanks.

 

[sreid]

Take a Look at Maxon Motors for Brushless Slotless Motors.

http://www.maxonmotor.com/product_overview_details_ec_motor.html

 

[electricpete]

Here is an article that talks about skew of the slot opening to reduce cogging torque in bldc (not the entire slot, just the opening from slot to airgap in semi-closed slot)

http://www.aece.ro/bfiles/files/1_2...shless_dc_motors_using_slot-opening_shift.pdf

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

Here's one that talks about "adaptive control" of torque ripple in bldc. May be similar to what cwilson was saying, I'm not sure:

http://www.advanced-energy-conv.com/pdfs/apec98.pdf

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

Google shows a lot of discussion of this subject:

http://www.osti.gov/bridge/servlets/purl/100261-AWMg7z/webviewable/100261.pdf

http://pemclab.cn.nctu.edu.tw/W3lib/confers/IMCSD/Hartman.imcsd00.pdf

Here on page 13/15 talks about cogging torque reduction for axial flux bldc machines:

http://journal.esrgroups.org/jes/papers/5_1_2.pdf

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[sreid]

Beyond the reluctance cogging torque from the stator teeth there will be torque variations from rotor position. A brushless motor will only have a constant torque constant if 1) the drive currents are pure sine waves 2) the magnetic field from the magnets is a pure sine and 3) the rotor winding distribution is a pure sine.

These requirements are more easilly approached in slottless motor designs.