Temperature Coefficient 1

[Clyde38]

When calculating permanent magnet electric motor performance, the variation of the magnet (temperature coefficient of B_r) and the winding (temperature coefficient of Resistance) over a temperature range is factored in using the value of thermal resistance (at some ambient temperature) published for the motor assembly. The temperature coefficients published for magnet materials and winding materials is assumed to be linear over a certain temperature range. Does the thermal resistance vary with respect to the change in temperature? If so, how is this characterized?
See reference at this site:

http://www.specamotor.com/en/case_brushlessmotor.php

I feel that the thermal resistance should vary since the materials that make up this resistance (air, copper, steel, etc.) change properties with temperature.


[link

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

]

[/url]

 

[IRstuff]

But, your thermal model, as described in your citation, is dominated by the convective resistance, but there's nothing on that page that indicates what the 10.3 K/W really refers to. It could be already determined for a high temperature, or, you need to manually iterate the calculation and supply a new convective resistance for a couple of iterations.

TTFN

FAQ731-376

 

[Clyde38]

The 10.3K/W refers to the thermal resistance of the windings to the housing (Rth 1). The thermal resistance of the housing to surrounding ambient is Rth 2. Data sheets refer to an ambient temperature for information such as resistance, torque constant etc., so it seems that the thermal resistance falls under this criteria (it is not differentiated in the data sheet). I'm using an iterative method to determine the temperature rise of the motor

http://www.specamotor.com/pdf/overheating_white_paper.pdf

(see attached spreadsheet), however I have to use the thermal resistance as reported by the data sheet. Am I looking for something like a Nusselt number (

http://en.wikipedia.org/wiki/Nusselt_number)

that takes into account the convective heat transfer coefficient and the conductive heat transfer coefficient? If so, can this be calculated from any published material properties? Does the thermal resistance change with temperature, and is it an issue to properly calculate the temperature rise of the motor?

[link

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

]

[/url]

 

[IRstuff]

According to the page you linked originally 10.3 K/W is Rth2.

And yes, the convective coefficient can be calculated:

http://www.efunda.com/formulae/heat_transfer/convection_forced/calc_lamflow_isothermalplate.cfm#calc

Note that this is a subscription site, so copy all that you can from the get-go. You only need to go the Prandtl Number link and the theory link. The Reynolds Number is on the same page as the Prandtl Number.

This issue will be that you don't know the actual parameters used, but you should be able to scale the equations to follow the temperature behavior of the equations used.

TTFN

FAQ731-376

 

[Clyde38]

IRstuff,
Thanks.
Rth2, my mistake.

Would it be correct to say that you agree that the thermal resistance will change with temperature? And if so, do you feel (gut reaction) that it will influence the calculations enough to pursue? Typically the temperature rise that I'm reffering to is about 100C.

[link

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

]

[/url]

 

[IRstuff]

The simple solution is to use a convection coefficient compatible with the expected temperature. The resultant numbers will be "close enough" since many assumptions that go into the calculations are not always valid anyways. A 50°C change in the surface temperature should result in about a 6% change in convection coefficient

TTFN

FAQ731-376