JasonLion said:
All it takes is a little resistance between the hot wire and the case, 5 or 6 ohms would probably do it, and a regular breaker won't trip. You then get a loop between hot and ground across the pump casing with the pump casing ending up at some intermediate voltage due to resistance in the various wires involved...
...and say that the bonding wire to the pump is missing and there aren't any other connections between electrical ground and the bonding system... ...Then there is a circuit from the pump casing through the water to anything metal, say a ladder, and back through the earth to the local electrical ground.
I don't want to belabor the point unnecessarily, but to continue with your example, there is electrical current somehow flowing from the faulty pump into the water in the pump housing, through the pool plumbing, into the water in the pool. And the path of the current flow is from there to the earth through a ladder, and from there through the earth to the electrical ground. The maximum potential at the hot wire is 110 volts nominal. (It is difficult to imagine a scenario where anyone in a pool might be exposed to 220 volts - it would require both hot wires to be connected to different parts of the pool. But if one wire did become connected to the water in the piping, and the other to the bonding system of the pool, and the swimmer was the only connection between the two, then the GFCI would not be of any help since it cannot prevent a shock between the two hot wires, only between a hot wire and ground. But I digress.) There are all kinds of resistances in the proposed scenario, your 5 or 6 ohm "short", the water itself which, even loaded with salt, is far from a good conductor, and the not inconsiderable resistance through the earth itself to the electrical service ground rod(s). This resistance limits the current flowing in the system, and the voltage drop across each resistance reduces the voltage. The metal ladder is a better conductor than the water or the earth, and connects the two, meaining there is virtually no local difference of potential between a swimmer in the water and the ladder or the deck it is attached to.
The bonding system of the pool is by far the most important electrical safety consideration. If concerns about a defective bonding system are the motivation to add a GFCI to the pool pump, then what about the many other more likely sources of potentially lethal voltages that can affect our pools? Ignoring lightning, overhead power lines, the shorted well pump leaking current into the garden hose, and the shorted sewage transfer pump up the street leaking current back through the waste line, just consider stray voltage in the ground near the pool. The ground itself is often "energized" and the sources of the voltage can be difficult to find. Maybe the neutral circuit in your next door neighbor's house is leaking current into the ground through their ground rod (maybe their ground rod is closer to your pool than your ground rod). Maybe your neighbor's pool pump is shorted and current is flowing from it through his pool into the ground near your pool. If your bonding system is broken you are at risk regardless of how many GFCI breakers you have.
I guess for a very far-fetched, yet definitive, answer for whether purchasers of 220-volt GFCI breakers are getting "pool stored", we would need some calculations. Assume a 110-volt (or let's say 130-volt, for worst case conditions) hot wire is connected inside one end of a water-filled pipe. Assume the conductivity of the water is as high as a swimming pool gets, with high salt and other mineral levels. Assume the pipe is as short as is found in typical swimming pool installations. Assume the other end of the pipe is connected to a container filled with the same water, and a swimmer is standing chest deep in the water. The swimmer holds in his hand, outside the container a bare copper conductor connected directly to electrical ground. Is the potential voltage difference across the swimmer's body high enough to make him regret volunteering for the experiment?