Grounding issues and ground tester

A standard three-prong outlet tester will have three lamps (usually neon bulbs): hot to ground, hot to neutral, and neutral to ground. The first two should be lit, and the third should not be lit. There's not a lot to them, and neon bulbs are pretty bullet-proof, so they should be reliable.

An "open ground" reading is when the "hot to neutral" is lit, but the "hot to ground" and "neutral to ground" aren't. The only bad reading is the "hot to ground" bulb, which means that there's not enough ooomph between the hot and ground terminals to light the neon bulb. It's quite possible that your multi-tester (a DVM?) is far more sensitive than the tester, and so is reading a voltage that is only a phantom, i.e. isn't supported by any current capacity. In any case, a bad ground shouldn't cause shocks to your horses. And, the presence or absence of ground fault receptacles wouldn't change an outlet tester's reading.

You might try dropping one probe from your presumed DVM into the horse tank (or just touch it to the tank if it's metal) and pushing the other probe into the (damp) ground, and seeing what voltage you come up with.

Those plug in testers can be total crap. I have about a dozen of them and 2/3rd of them show good circuits while the other 1/3rd show problems, and not all the same problem.

You might be surprised to find out with a high impedance meters ( the type most meters are) you can plug the meter lead in the hot terminal (the smaller one) you may read a voltage very close to your supply.

As for the neutrals and grounds being in the same bar this correct at the main panel but every place else it is not allowed. Since you just replaced the water heater and are having problems I would add a ground bond to the outlet pipe, animals are more sensitive to leakage currents than we are and this might solve the issue. I would also suggest to drive a 2nd ground rod.

Standard rods are only 8' long in the past when I have used pipe I use a full stick that is usually 10' long. Current code requires 2 if this is your only grounding system 6' or further apart. Unless the single rod has a 25 ohm or less resistance (takes a special meter to measure so most just drive 2 ).

Also, I was initially thinking we were discussing a water heater, now with the edit I see it is a "cal rod heater". Put your meter probe in the water and then to ground. If you measure a voltage the heater has leakage and is faulty, take it back. All my tanks are metal and sit on the ground outside and we only have a few weeks a year where we have to break ice. I would encourage you to put this heater on a GFCI outlet! But I believe you have a faulty heater.

Even with no ground the hot conductor of the cal rod is in the center protected by a metal jacket. It is basically a big resistor that gets hot so either the jacket is cracked or the connection points are not properly sealed - this is the reason the horses are getting shocked. Not a grounding issue.

3 lamp tester, into the trash it goes.

Grounding systems are simple but critical. My philosophy is "nuke it from space, it's the only way to be sure". The 2014 Electrical Code agrees with me, it greatly liberalized the rules for retrofitting grounds.

I break grounding systems into 3 sections, from the, um, ground up.

The Grounding Electrode System

This is the ground rod or water pipe tie-in (or Ufer if your concrete pouring guy was a smart cookie). Modern standard for ground rods is two of them some distance apart, connected by copper wire to your service panel's chassis and its grounding bar. I leave nothing to chance, bonding to both the chassis and the grounding bar.

The Neutral-Ground Bond

In your case, the neutral bar and the ground bar are the same bar, which is the sincerest neutral-ground bond possble. These often tie to the panel chassis with a green "ground bonding screw", I have seen cases where this screw was burned up by a prior overload.

The Equipment Grounding Conductor (EGC)

This is the "ground wire" that should be present in every circuit from the ground bar to the point-of-use. Because people understand the importance of it, this isn't usually where the problem lies.

I assume you checked this on the wiring to your horse trough, so there is a real possibility that the problem isn't this ground exactly, but all grounds in your house. The horses just discovered it first.

Ground fault detectors

GFCI "receptacles" can go right next to 3-lamp testers, as far as I'm concerned. Too many people buy a 10-pack at Home Depot and replace every receptacle with one, without a moment of thought, this creates a giant mess and a "Yo Dawg" joke. It also wastes money better spent on other safety apparatus. GFCI protection is a great concept and is important, but it is better installed intelligently, and only needs to exist at 1 point in a circuit. From there, it can protect everything downline of that point.

Since you have at least 1 circuit that runs around outside, yeah, GFCI protection is required for that, and in theory will stop horses from getting shocked. Put the GFCI device somewhere early in the circuit, so the unit is still indoors. Putting a GFCI device outside is just wasteful.

You can and should fit GFCI on this circuit, I recommend as a breaker or receptacle inside. However will simply convert it from a "horses getting shocked" problem to a "GFCI trips and refuses to reset" problem. The core problem will remain the same: your tank heater has a ground fault.

The only upside of opening up every receptacle to change it to GFCI is that it forces you to open up every receptacle and inspect it. There you may find all sorts of interesting problems left for you by history, and one or more might explain the problem you're having today. But just do that, then; no need to spend $20 per socket.

Or ground is different at the horse trough

A longshot possibility is that the tank heater is perfectly grounded, to the house, but the ground the horses are standing on is at a different potential than the house's ground rod. That could be caused by a significant amount of leakage from some other apparatus in the area. The earth has basically become a giant rheostat (open resistor) between two different points having differing voltage in the same system.

However for that to happen, somebody's electric meter would have to be spinning pretty quickly. And that would also be the case even if the horse trough heater's breaker was off. So powe off the heater and see if the horses still get shocked.

Here are the manufacturers 15N instructions. You might want to pay special attention to the Troubleshooting part called The livestock are being shocked and also Read and follow all instructions in this sheet.

Note: If all else fails there is an 800 help and tech line at the bottom.

Good luck and keep those horses safe.