What prevents heat generation in magnets?

[sdk_imported]

It appears that my motor is generating heat in the magnet or the rotor yoke. Under what conditions does rotor heating occur? What advice do you have for reducing it?

Thanks.

 

[jbartos]

Suggestion: There are several causes:
1. Voltage harmonic distortion causing voltages of higher frequency that causes higher Eddy Currents and Magnetization losses.
2. Lower voltage at the motor terminals. This causes higher currents flowing through the windings which heats up motor iron.
3. Higher voltage at the motor terminals which causes the higher magnetization current thus higher magnetization losses.

 

[sdk_imported]

Are eddy currents enduced more by high frequency variation or low frequency variation? Do you have any recommendations for troubleshooting or eliminating this problem?

 

[electricpete]

in general eddy currents are roughly proportional to frequency^2 divided by area^2.

What type of motor is it?

 

[sdk_imported]

I'm confused about the area term in your eddy current equation. I would think that eddy currents would be a function of material volume. What area are you referring to? Also, do you think switching losses are as significant as eddy currents? Are switching losses a function of frequency square also?

This is a four pole high speed (30 krpm) brushless dc motor. The motor uses a high strength extruded neo magnet. The rotor is 17-4 stainless.

Thanks.

 

[jbartos]

Suggestion: Reference:
1. Gordon R. Slemon "Magnetoelectric Devices Transducers, Transponders, Transformers, and Machines," John Wiley and Sons, Inc., 1966,
Section 2.2 Eddy Currents.
Equation 2.37 on page 124 indicates:
p=[(c x w x B)**2]/(24 x ro)
p is average Eddy Current Power Loss in Watt/(cubic meter)
c is lamination thickness
w is angular frequency = 2 x pi x f
B is maximum flux density
ro is the resistivity of material (notice that the high resistivity of material is reducing Eddy Current Losses)

 

[electricpete]

I think jbartos' reference gives a good description of what I was trying to say.

Assuming constant flux density somewhere near saturation of core steel, the voltage induced around a loop within the core (not interrupted by lamination boundary) will be proportional to the perpendicular area through which the flux flows. V = d/dt(B*A)

jbartos' expression I believe will estimate the heating associated with a section of core formed in a long thin lamination of area length * c, c<< length. I think I can derive that result P ~ c^2 as follows:

Rloop ~ 2*length+2*c ~ 2*length.

I ~ V / R ~ A / length ~ (length*c)/length ~ C.

P ~ I^2*R ~ c^2

 

[sdk_imported]

I'm comfortable with determining eddy current and switching losses in the stator laminations. I'm not familiar with dealing with these losses in the magnet or rotor yoke. I'm concluding that there are several things that I can do to reduce eddy currents in the rotor.

1. reduce the slot opening size
2. adjust the pwm frequency
3. adjust the pwm harmonic response
3. laminate the magnet
4. laminate the rotor yoke
5. reduce the pole count

Can you think of other things I should consider? Do you think that skewing the lamination stack would help?

 

[UKpete]

I don't think skewing will help.
One more to add to your list - neodymium magnets have higher resistivity than samarium cobalt.