On 3 Jan 2006 16:04:40 -0800, jim rozen
wrote:

In article , Robert Swinney says...

IMO, you need to lose the thinking of a RPC as being a form of generator.
It isn't. Think more of the RPC as a network in which parts of it rotate in
order to supply current throughout. Part of the RPC is the load motor. The
idler generates nothing without the load as part of a network. IMO, a
flywheel on the idler cannot act as anything more than additional dynamic
load on the network. It would be aprox. the same to put the flywheel on the
load motor instead. Forget flywheels and spend the money on enhancing the
idler-load network with proper capacitance. Complex current flows in all
parts of the RPC. In simplistic terms, the idler-load current paths can be
viewed as series resonant circuits. Such circuits are "tuned" via
capacitance placed in series.

Granted this kind of tuning is the very *first* thing one would do
before considering flywheels.

I specifically recall Gary Coffman claiming they would reduce transient
response, and yet there's a considerable group of well-informed
individuals on the practical machinist board who say they improve
matters.

I have to say I find *both* sides to be persuasive, at least at the
'hand-waving' level.

My suspicion is that flywheels probably help up to a point, if one
models the rotor as having zero mass to start with. And that the
optimum flywheel size will wind up being about one rotor unit in
size! This is what a former boss of mine calls 'the schwarz law
of the initial maximum.'

Ie, if it works the first time you set it up, anything you do to it
after that makes it work worse.

:^)

We just call that syndrome "fix it 'til it's broke".

Snarl