Can I use an insulation tester to do the same test?
No. This is a common error. Field operators are often issued a MEGGER ® instrument from stores, without its being checked to determine whether it's an insulation or ground tester. Insulation testers are designed to measure at the opposite end of the resistance spectrum from a ground tester. No one wants grounds that measure in megohms Insulation testers use high test voltages in the kilovolt range. Ground testers are limited, for operator safety, to low voltages. Insulation testers do commonly have low-voltage, low-resistance continuity functions and these are frequently misused to make jiffy ground tests. However, a continuity test can only make an arbitrary measurement between an installed electrode and a reference ground, which is assumed to have negligible resistance. This does not afford a reliable measurement of the resistance the earth offers to a ground fault current. Even this arbitrary measurement may not be reliable, since a dc continuity test can be influenced by soil transients, the electrical noise that is generated by utility ground currents trying to get back to the transformer, as well as other sources.

The required measurement is of resistance; why can't I use a multimeter?
For the same reasons that a continuity range on an insulation tester should not be used. Measurements made with a dc multimeter are subject to distortion by electrical noise in the soil. A multimeter offers no means of verifying that the resistance displayed represents anything other than an arbitrary measurement between two convenient points. With a multimeter, one can measure the resistance of the soil between a ground electrode and some reference point, such as the water pipe system, but a fault current may encounter a higher resistance. Genuine ground testers are amenable to field-developed standard procedures that have built-in cross-checks that expose insufficient test conditions.

What is the difference between a two-point, three-point, and four-point test?
Literally, the number of points of contact with the soil. More specifically, these commonly used terms refer to dead earth, fall of potential, and Wenner method tests, respectively. In the dead earth method, contact is made at just two points: the ground electrode under test and a convenient reference ground, such as the water pipe system or a metal fence post. In the fall of potential method, a genuine ground tester makes contact via the test electrode, plus the current and potential probes. With the Wenner method, no ground electrode is involved, but rather the independent electrical properties of the soil itself can be measured using a four-probe setup and a recognized standard procedure.

Can't I simply make a measurement to a reference ground?
This common method often uses an instrument other than a dedicated ground tester. It is referred to as the dead earth method because the reference ground is only being used for the test and is not normally part of an electrical system. It can be the water pipes, a metal fence post, or even a rod driven just for the test. The method is popular because of its ease and generality, but is not recommended. Since the reference ground happens to be located by a combination of convenience and chance, it is only a matter of luck if the soil resistance to it actually represents the true electrical ground resistance. Furthermore, the measurement has to be accepted on faith, because there is no way to validate it, as there is with accepted standard methods. The method has no independent recognition from standards authorities, and so is of no use in establishing the reliability of results or in liability protection.

How do I know if my ground is good?
The most widely used specification is that of the NEC, which mandates for residential grounds a resistance of 25 or less. This is not a particularly difficult specification to meet. Others are more demanding, and may be specified by the engineer designing an electrical system, or by a client, or may come as part of the warranty requirements for advanced equipment. The most commonly encountered specification for industrial grounds in general is 5 or less. Computers and process control equipment may demand as little as 1 or 2 .

How often should I test my grounds?
Odd intervals of 5, 7, or 9 months are recommended so that the various seasons will all be encountered in succession. This is because the quality and effectiveness of a ground are profoundly affected by weather and seasons. If quarterly or semiannual testing schedules were used, certain months would consistently be missed, and these could be the ones in which the grounding is most stressed by the weather. Adopting irregular intervals, on the other hand, ensures that worst case seasons will be revealed. Since a ground fault, potential fire or accident, can happen at any time, your protection is only as good as the ground condition in the worst time of year.

How do I go about designing a good ground?
Traditionally, the trial-and-error method was used and could be enhanced by an individual's experience. This consists of designing the ground during the process of installation, and repeatedly testing it in progress, until the desired spec has been met. For example, a rod can be driven then tested. A second rod can be coupled to the first, driven deeper, and tested again. This procedure is repeated until spec is met. Similarly, a second rod can be added in parallel (driven into the soil separately and connected by a conductor rather than coupled end-to-end) and tested. Additional rods can then be added to form a ground bed until a sufficiently low resistance is obtained. The trial-and error method is still frequently used and often works well. But its limitation is that it is subject to the Law of Diminishing Returns: more and more work for less and less reward. In optimal soil conditions, a satisfactory ground can be achieved with only a few retests. But in more difficult environments, one can end up wasting the day without realizing the goal.

A more ordered approach is available and begins with the manufacturers of grounding materials. They will commonly provide, at little or no cost, engineering programs to calculate the best theoretical design in advance. These are usually available on computer disc and require the input of some data, which include resistivity measurements of the available area, and the ground resistance specification that must be met. From this data, the program calculates a design which can be installed as a unit rather than the piecemeal approach of trial-and-error. A single resistance test should then verify that the spec has been met and, if not, only a simple addition or adjustment should be sufficient.

How do I hook up my leads?
With a three-terminal model, the common is connected to the ground being tested. The shortest available lead is used for this connection because the test ground should be close at hand and the lead resistance will be part of the measurement. If the tester is a four-terminal model, the operator has some discretion. The C1 and P1 terminals can be jumpered or not. If jumpered electronically by means of the appropriate test button (newer models have separate test buttons for three- or four-terminal operation), the lead should be taken from the C1, not the P1. The internal connection made by means of the test button is resistive, not a dead short. If the test lead is run from the P1, this internal resistance will show up in the measurement. If both terminals are to be used, two leads are run to the test ground. In this genuine four-wire bridge configuration, no lead resistance enters the measurement. So the choice is resolved between time and accuracy. Three-wire is quicker because it requires less hookup. If the test spec isn't unusually low, the little bit of lead resistance doesn't hurt. Four-wire requires a bit more work, but if you must meet one or two W , it may be helpful to make the accuracy as sharp as possible.

The longest lead is connected to the C2 terminal, and normally, or in the absence of any other specifics, stretched out to its full length. The P2 receives the lead of intermediate length and this will probably be moved several times between the tested ground and the current probe, according to the dictates of the specific procedure being used.

How do I know how far to extend the leads?
Unfortunately, there's no failsafe method of determining this in advance without the possibility of some trial-and-error. That's because soil conditions are infinitely variable. In highly conductive soils, the resistive volume surrounding the test ground is comparatively small and acceptable measurements may be made from as little as 25 ft away. As soil type and conditions worsen, and/or the resistance spec goes down, the resistive field enlarges and can become quite extensive. Then, extremely long test leads may be required in order to get out of the resistive field of the electrode being tested. How can this very important consideration be approached?

First, industry rules-of-thumb exist and are often seen reproduced in tables in the literature. These vary somewhat depending on the disposition and experience of the issuing agent, but in general it can be said that the current lead is extended four to five times the length of the maximum dimension of the ground being tested (some more rigorous authorities recommend ten times). For example, with a driven rod, the current probe would be placed five times (or whatever multiplier is being used) the depth; with a ground bed, five times the diagonal; for counterpoise, five times the diameter, and so on. Just remember, these are only rules of thumb. They are not strict requirements. The underlying concept is that the practice will provide a reasonably good likelihood of getting an acceptable test on the first try. But this is neither guaranteed nor mandated. If the calculated distance is impractical or impossible, there is nothing wrong with testing at whatever distances are available. The only thing being risked is the increased possibility that the result won't qualify, and the test will have to be repeated; i.e., more trial-and-error.

Secondly, you may just use whatever is available. If it is not inconvenient to run the leads out to the maximum (100 ft or more), there is a good chance that most tests will be adequate. As long as an acceptable test procedure is followed, bad results caused by insufficient probe spacing will be recognized and thrown out. The more you can live with trial-and-error, the less concern there need be about getting it right the first time. And, operator experience is invaluable. Familiarity with local conditions and previous tests are often all that is needed for the experienced operator to make a practical decision on the first try.

And thirdly, the test procedure being used has something to do with the amount of lead length that will be needed. Some methods (e.g., Slope Method) tend to require less distance than others (e.g., fully developed Fall of Potential) for a given test ground. Just remember, as far as lead length is concerned the ends justify the means. If an acceptable test result (that is, coherent and repeatable) is obtained, the leads were long enough.

What is the significance of the provided lead lengths?
As a generalization, they are long enough to yield an acceptable test on the first try in most common environments. The one false conclusion that must not be inferred is that they are the exact lengths needed in all situations. That is to say, do not merely stretch the leads out to their full lengths, conduct a test, and leave. You may, and most likely will, have a good test but there is no certainty. There are many atypical environments in which the provided lead lengths are not enough. Any attempt at a universal lead set would be highly impractical, inordinately long, and a nuisance to work with in most routine test situations. The provided leads will be adequate in most instances,\ but a margin must be allowed for situations in which longer lengths will have to be used. No lead set design can serve as a preordained substitute for good fundamental practice.

Do I just go out 62 feet and take my measurement?
This practice does have limited applicability, but should not be relied upon. It frequently appears in the literature as the "62% Rule" (more specifically, 61.8). It is based on calculations that have shown that, under ideal conditions, a Fall of Potential graph will encounter the value that represents the theoretical true ground resistance at a position that is 61.8% of the distance to the current probe. The theoretical true resistance is the value for that ground electrode in that location if the resistance of the entire planet could be measured. Of course, it cannot but fortunately for practical considerations all but a miniscule amount of that resistance is determined by the conditions in the immediate area which can be measured. Because resistance rises well above theoretical true in the vicinity of the current probe, the total graph therefore will include the value equivalent to theoretical true.

The limitation of the "62% Rule" is that it assumes ideal conditions. These include adequate probe spacing and homogeneous soil. Soil is rarely entirely homogeneous (of a thoroughly consistent nature) and around graded construction sites, it will be particularly disturbed. Therefore, by pure misfortune, the potential probe at 62% of the total distance may be driven into a localized hot spot where the soil in a small area is not representative of the prevailing condition. This could cause a reading to be too high, resulting in unnecessary improvement of the ground. Conversely, it could make the reading low, resulting in the facility being left without the intended protection. And unlike other standard test methods, the result of the 62% test has to be taken on faith; i.e., there is no built-in safety check to toss out readings taken with inadequate spacing. Again, unless the probes happen to be separated by sufficient spacing, a reading could be taken on the rise of the resistance vs. spacing curve, and falsely indicate that the ground has met spec.

So, where can this method be successfully used? It may be that in some non-critical applications, the speed and ease of this test may make it attractive to risk the occasional bad test slipping by. Unless thoroughly justified, however, that is not a recommended idea. Utilities may sometimes scour an area with these abbreviated tests for a line of pole grounds, for instance. But there is always a back-up team that performs a second test by a more rigorous means on any locations that aren't in agreement with expectations from previous records. The most advantageous use of this method, then, is where the assumptions have been met; that is to say, where previous thorough testing has established the distances, points of probe placement, and so on, and they are recorded in a maintenance history. In such cases, subsequent maintenance testing can be performed without as great an expenditure of time and effort by using the "62% Rule".