Quote:
"Is that why you try to correct as close to the load as possible, if installation costs permit such a design?"

Yes. The motor still has reactive power (Var) requirements, which the capacitor is now providing. The wire between the two still carries this reactive power.
By the same token, if you don't connect the capacitor where it comes on and off with the load, you'll have the opposite problem. When the motor goes down, if the cap stays on, you now have just as bad a PF, except now it's leading instead of lagging.
In terms of your bill, this probably isn't that much of a concern since you usually get penalized for your PF at peak demand.
However, even at no load the caps would be hot and you would have some issues.

But remember that there is a law of diminishing returns to apply here. There is a point where the cost of additional capacitors is not justified.
In other words, correcting to 95% PF might save you tons of money, but correcting to 98% may be twice as expensive and only return slightly better savings.
If billed on a PF penalty, you would need to know what the POCO's limits are.

[This message has been edited by WFO (edited 02-25-2006).]