[femm] Re: DC MOTOR DESIGN BOOKS

[Dcm3c@xxxxxxx]

In a message dated 11/1/99 3:54:05 PM Eastern Standard Time, 
tsiporak@xxxxxxxxxxx writes:

> I am looking for "old" books for DC Motor design calculation, such as
> number of poles, number or turns per coil, number and sizes for iron
> sheets, etc. I know that I can FEA analysis but to better get the
> "feeling" for the basics of motor physics. I am looking for books and
> or articles published upto the 1970's before DC motor designers had the
> FEA tools. Any help will be appreciated/

Check out the on-line book, Vector Control of Induction Machines, at:

ftp://murray.newcastle.edu.au/pub/reb/elec414_material/Notes/Elec414Notes.pdf

Although its title says "induction machines," there is a very good section at 
the beginning that talks about modeling of DC machines, and especially the 
assumptions that you make in reducing them to a form that is easy to model.

As far as physical books, try:

G. Selmon, "Electric Machines and Drives," Addison-Wesley, 1992

which is a pretty comprehensive book.

Another nice tool that is useful in practice is the program Wiretron, freely 
available at:

http://www.wiretron.com/free.html

This program contains a lot of practical information about wire sizing, 
current capacity, different types of insulation, etc, etc.

As far as stuff like laminations and magnets, the only good source of 
information is manufacturer's spec sheets. Unfortunately, these guys don't 
seem to be very webified yet, or maybe I'm just not aware of where this 
information resides. 

This is sort of a peeve of mine: finite element are a useful tool to motor 
designers, but it's generally a very bad idea to use _just_ finite elements. 
It's generally the most productive to do a design using an analytical 
approach, and then use finite elements to look at the design and investigate 
the effects that you might neglect when making the analytical models (like 
saturation, higher order harmonics, and so on). Another good use of FEA is 
to use finite element models to obtain parameters that are used in analytical 
models, but which are hard to calculate by hand. I use this trick a lot, too. 
The problem with using only finite elements is that the design ends up being 
sort of a numerical version of cut-and-try. Since you don't have any insight 
into _why_ the model is doing what it is doing, your only recourse is many 
iterations of making changes to the design and re-running the model until you 
get what you want. Typically, there are too many parameters that describe 
any given machine to do this effectively.

Dave.