I really don't understand the objection to grounding probes used in conjunction with GFCI.
As I understand this article, the author mentions three potential problems with using a grounding probe:
1) If there is a short in a single device -- within itself -- this would not be a "problem" (really?), but a problem would be created by addition of a grounding probe because it would take the short out of the device "itself" and send the current to ground. Clearly this would be a problem with no GFCI. But with a GFCI, this would cause a trip and you'd investigate and find the problem A good result. Better than having a device shorting "within itself" without you knowing it.
2) If you have a ground probe, it will keep you from finding a problem. Better to know about it and diagnose it. Ok. But, again, if you have a properly working GFCI you will know about it because you'll trip your GFCI.
3) You're creating a path to ground that is in an invitation to allow current to travel through your tank, which it otherwise wouldn't if you didn't provide the invitation. Clearly true. And, clearly, a reason why a ground probe without a GFCI is a bad idea. But, since a fish tank without GFCI is a bad idea, that's redundant.
If you click on the article linked within the article, the author does mention one scenario -- albeit unlikely -- in which you definitely want a grounding probe IN ADDITION TO a GFCI. Here it is:
xxxFor example, if one power head's hot lead (black wire) is exposed to the water and another power head's neutral lead (white wire) is exposed to the water, then there will be a current flow between them as well as any sea life that gets in the way. Lets say that both power heads were plugged into the same GFI. Since all the current coming in the black wire is going out the white wire (some through the power heads, some through the water and fish)... there is a balance, but an unhealthy one for the tank. The GFI would not trip. Adding a ground probe would cause the GFI to trip in this example. xxx
What he doesn't say is that not having the grounding probe means that when you stick your hand in the tank in the path of the current, you become the grounding probe.
His final objection to grounding probes is that there is some current that is added to water by induction from the devices you have plugged into your GFCI. (These will not trip the GFCI because they are currents caused by the devices, and so the hot and neutral voltages measured by the GFCI remain constant.) These induced currents, he says, will not be harmful to the fish when there is no grounding probe, but may be harmful if there is a grounding probe, because the potential in the tank may be different from the potential in the ground in your home and thus these induced voltages and currents will find a path in grounding probe. I'm not entirely convinced.
But it seems to me that, if you're reading this article and trying to decide whether or not to use a grounding probe in conjunction with your GFCI, it is a mistake to read this article (as many here seem to be reading it) to say that a grounding probe is a negative. It's a choice you need to make as an aquarist, and it seems relatively narrow: which am I more concerned about: (1) a relatively unique scenario where current is flowing through the tank from a bare hot wire of one device to the bare neutral of another v. (2) small amounts of induced currents finding a path through my grounding wire that may present unpleasant zones in my tank for my fish. I know my answer. But each is different.
The lesson seems to be USE GFCI. Whether or not you use a grounding probe is a more nuanced question, with no "right" answer.
Disclaimer: I am not an electrician. I am a hobbyist trying to make sense of complicated matters of physics. All of this could be wrong. Do your own due dilegence before sticking your hand in salt water near devices drawing hudreds or thousands of watts.
