Re: how deep do the coilers prefer doing it? : )

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: Jim Lux <jimlux@xxxxxxxxxxxxx>

At 09:39 PM 2/11/2006, Tesla list wrote:
> Original poster: "Dmitry (father dest)" <dest@xxxxxxxxxxx>
>
> Hallo.
>
> yesterday i got a 1/2 hp, 2780 rpm capacitor run motor - you can see
> some pics of it here - http://himplast.ru/father_dest/temp/370w/
> this puppy has no "dead poles" at all, and was fully rewound by hands
> of some drunk electrician % )))
> i see that nobody on the list really know, how deep should one shave
> that poor rotor down to perform so called "salient pole" conversion.
> i mean nobody know any formulas, nobody can explain how wide exactly
> should one cut that flats, nobody know how one could verify that
> performed conversion was optimal in terms of lost torque and etc.
> am i correct on this?


I suspect that in the motor mfr business, it's done empirically.  You 
build a bunch of prototype motors to get the dimensions, then go 
manufacture 10,000 more of the best.

One CAN model these things, but it's quite complex.  I am under the 
impression that there aren't any simple analytical expressions that 
are useful (that is, the simple expressions can give you a 
qualitative feel, but can't tell you the last decimal place) and that 
they use fairly sophisticated FEM models to do detailed 
analysis.  Lots of interacting effects (geometry of the poles, the 
windings, etc.)


> coz all that i can find in archives is something blurry like this:
>
> "The info I supplied is more of a general guideline than anything --
> you will inevitably see variations on the number of poles, etc."
>
> "Maybe the flats should still be 3/4" wide like the 1750rpm motor, I
> don't know.  There are various opinions."
>
> "In other 3600 rpm cases I made the flats narrower, but the motors
> seemed weaker.  But these were all different motors so it's hard to
> judge."
>
> does anybody know any real physics that stands behind all of this? what
> should one monitor in process - input currents, any phase shifts,
> something else?
> or coilers prefer just to cut 1/3 of od and don`t want to know nothing
> more? : )


If you have a motor dyno (something that most motor manufacturers 
have) you can run speed/torque, current & voltage curves as you cut a 
bit more each time.  Once you find the "optimum" for your application 
you now have a recipe to make more identical ones.  The problems are this:
1) The recipe is peculiar to the specific motor model
2) The recipe is peculiar to the particular load (torque vs speed) 
characteristics. (A RSG is probably somewhere between torque 
proportional to speed and torque proportional to square of speed)
3) In the case of a synchronous motor with a load, the inertia of the 
motor rotor and the load being driven, as well as the elasticity of 
the shaft, will all interact.

So, you just hack away to make it good enough, and unless you're in 
the business of making RSGs, you're probably only making one, and 
once it works, you call it done.