Grounding system testing and soil resistivity field guide

Ground resistance testing, and what the number tells you

Grounding system testing measures the resistance from an electrode to remote earth and confirms the path to ground is low enough to do its work. A ground that reads high looks fine on the print and fails when it is needed. The number is in ohms, and the lower it reads, the better the electrode dumps surge, lightning, and the system's voltage reference into the earth.

This guide is the testing side. The install side, choosing and bonding the electrodes, sizing the grounding electrode conductor, the Ufer, the two-rod rule, lives in the grounding electrode system and bonding field guide. The difference between grounding and bonding, and why the dirt never clears a fault, lives in the grounding versus bonding guide. Read those for what gets built. This one is about proving what got built actually performs.

One thing to settle up front, because it is the source of most field arguments. The earth connection does not clear a ground fault. That job belongs to the bonded metal path back to the source, and you do not test it with a ground resistance meter. Ground resistance testing answers a narrower question: how well does this electrode tie the system to the planet for surge, lightning, and reference. A high reading there is an ineffective ground, not a non-tripping breaker.

What is ground resistance?

Ground resistance is the resistance the current sees traveling from the electrode, through the soil around it, out to remote earth far enough away that it no longer matters. It is not the resistance of the rod or the copper. The metal is a fraction of an ohm. Almost all of the reading is the soil packed against and around the electrode.

Picture shells of dirt around the rod. The shell touching the rod is small in area, so it carries the current through a tight cross-section and contributes the most resistance. Each shell farther out is larger, carries the current through more area, and adds less. Past a certain distance the shells are so large they add nothing, and that distance is what remote earth means. This is why the first few feet of soil around an electrode set the reading, and why a rod in a small pocket of good dirt surrounded by rock still reads high.

Lower is better, full stop. A low reading means the electrode is well coupled to a large volume of earth and can move surge and lightning energy without the voltage at the electrode climbing. A high reading means the opposite, and the targets in the next section are how you judge whether the number you measured is good enough for the job.

What is a good ground resistance value?

There is no single number the way people want one. The most quoted figure, 25 ohms, comes from NEC 250.53, and it is widely misread. It is the trigger that decides whether you must add a second rod to a single rod electrode, not a performance target the finished system has to meet. Drive one rod, test it, and if it reads 25 ohms or less you may leave it as a single electrode. Read over 25 and you add a supplemental electrode, and once the second one is in, the NEC stops asking for a number.

So where do the low targets come from? The project specification and the equipment, not the code. Sensitive and critical installations are held far tighter than 25 ohms, and those numbers come from IEEE guidance and the owner's spec. Substations are commonly designed to 1 ohm or less. Data center and telecom specs often call for 5 ohms, with the more critical tiers pushed toward 1 ohm. These are recommendations and contract values, not NEC mandates, and the right target is whatever the spec, the equipment listing, and the adopted code edition call for. When a contract names a number, that number wins.

What is soil resistivity?

Soil resistivity is the property of the dirt itself that decides how hard it is to push current through it, measured in ohm-meters or, in older references, ohm-centimeters. It is the input. The ground resistance you measure at an electrode is the output, and resistivity is the largest thing driving it. Same rod, same depth, two sites: the one in conductive clay reads a fraction of the one in dry sand or rock.

The number ranges over orders of magnitude, not percentages. Wet clay and loam can run a few tens of ohm-meters. Sand and gravel run into the hundreds. Rock, dry sand, and frozen ground run into the thousands and higher. That spread is why you cannot guess a ground from the rod alone, and why a design that works on one site drives twice the steel on the next.

Because resistivity drives everything downstream, it is the first thing measured on a job big enough to design the grounding rather than default to two rods. You measure resistivity with the Wenner four-pin test, covered below, and feed that profile into the electrode design so the finished system hits its target on the first try instead of by trial and error with a sledge.

What changes soil resistivity: moisture, temperature, salt

Soil resistivity is not a fixed property of a site. It swings with moisture, temperature, the soil chemistry, and how packed the soil is, and the swing is large enough to turn a passing ground into a failing one between seasons.

Moisture is the biggest lever. Dry soil is a poor conductor; add water and the dissolved salts in the soil carry current, and resistivity drops fast. This is why a rod tested after a week of rain reads low and the same rod in a drought reads high. Temperature works the same way at the cold end. As soil approaches and passes freezing, resistivity climbs steeply, and frozen ground is close to an insulator, which is why a reading taken in January over frost can be several times the summer number. Salt content and soil chemistry matter because it is the dissolved electrolytes, not the water alone, that conduct. A little natural salt in the soil reads low; clean washed sand reads high even when wet.

The field lesson is to test at the worst case, not the best. A ground that passes in wet spring and fails in a dry, frozen winter is a ground that fails when a winter storm sends the surge. If you test in the good season, note the condition, and treat a marginal pass as a fail until you have seen it in the dry or frozen worst case.

How do you test soil resistivity?

You measure soil resistivity with the Wenner four-pin method: four small pins driven in a straight line, equally spaced, wired to a four-terminal resistance meter. The two outer pins inject test current into the soil and the two inner pins read the voltage the current develops, and from that the meter gives a resistance reading you turn into resistivity.

The arithmetic is straightforward. Resistivity equals 2 times pi times the pin spacing times the measured resistance, with the spacing and the depth in matching units. Drive the pins shallow, a fraction of the spacing, so the result reflects the soil and not the pins. The single most useful fact about the method is that the reading represents the average resistivity down to a depth roughly equal to the pin spacing. Set the pins 5 ft apart and you read the average soil to about 5 ft. Spread them to 20 ft and you read the average to about 20 ft, picking up the deeper layers.

So you do not take one reading. You take a series, widening the spacing each time, and the set of readings builds a profile of resistivity versus depth. That profile is what tells the designer whether the conductive soil is near the surface, where a ground ring or shallow grid will work, or deep, where rods have to be driven long to reach it. Run the traverse in two directions across the site to catch soil that varies sideways as well as with depth.

ρ (rho)

Soil resistivity in ohm-meters, or ohm-centimeters in older tables

a

Pin spacing, equal between all four pins; the reading reflects soil to about this depth

R

The resistance the four-pin meter reads between the inner potential pins

How do you test ground resistance?

The accurate way to measure an electrode's resistance to earth is the fall-of-potential test, also called the three-point or three-terminal test. It is the reference method, the one acceptance specs call for, and the one IEEE 81 describes for verifying a ground.

You drive two auxiliary pins in a straight line out from the electrode under test. The far pin, the current probe, goes a long way out, well past the resistance area of the electrode, often a hundred feet or more for a single rod and much farther for a large grid. The meter pushes a known current from the electrode to that far current pin. The middle pin, the potential probe, reads the voltage at points along the line between them. Resistance is voltage over current, read straight off the instrument.

The catch that makes or breaks the test is distance. The current pin has to be far enough out that the electrode's resistance area and the current pin's resistance area do not overlap. If they do, there is no valid reading anywhere on the line, and the test has to be re-run with the current pin moved farther. How you confirm the spacing was adequate is the 62 percent rule and the plateau, which is the next section, and it is the part most field failures come down to.

The 62 percent rule and the plateau

The potential pin does not go just anywhere. The valid reading sits at about 62 percent of the distance from the electrode under test to the current pin, and that figure is not a rule of thumb. It falls out of the math for the way the two resistance areas overlap, and at 62 percent the two effects cancel and the meter reads the electrode's true resistance.

In practice you do not trust a single reading at 62 percent. You take readings as you walk the potential pin out, at several points along the line, and you plot resistance against distance. A valid test shows the curve rising, then flattening into a plateau through the middle of the run, then rising again as you near the current pin. The flat part is the true resistance, and the 62 percent point should land on it. A common field shortcut reads three points, near 52, 62, and 72 percent of the distance, and if all three agree closely, the current pin was far enough out and the middle value is your answer.

Here is the failure that voids a test. If the curve never flattens and just climbs steadily from the electrode to the current pin, there is no plateau, the resistance areas are overlapping, and no point on that line is a valid reading. The fix is not to pick a number off the slope. It is to move the current pin farther out and run it again. A reading without a plateau is not a low number or a high number. It is no number.

Why you disconnect the electrode for fall-of-potential

The fall-of-potential test measures the electrode you connect it to, and if that electrode is still tied into the rest of the system, you are not measuring the electrode. You are measuring it in parallel with every other ground path in the building: the other electrodes, the water pipe, the bonded steel, the utility ground coming in on the neutral. The reading comes back low and wrong, because all those parallel paths pull it down.

So the rule for a true single-electrode fall-of-potential reading is to disconnect the electrode from the system first. Open the connection at the electrode, test it in isolation, then put it back. That gives you the resistance of that electrode by itself, which is what acceptance of a specific electrode usually wants.

This is also where the danger lives, and it is worth being blunt. The grounding electrode conductor can be carrying current, and on an energized system, disconnecting it can put you in series with fault or neutral current and can leave equipment unreferenced while it is open. You do not casually open a ground on a live system. If the electrode cannot be safely isolated, that is exactly the case the clamp-on tester was built for, because it reads without disconnecting anything.

What is a clamp-on ground tester?

A clamp-on ground tester, sometimes called a stakeless tester, measures the resistance of a ground without driving any pins and without disconnecting anything. You clamp it around the grounding electrode conductor or the rod, and it works by inducing a voltage into the loop with one coil and reading the current that flows with another. From voltage and current it gives a resistance.

What it actually measures is the loop: the resistance of the electrode you clamped plus the parallel resistance of every other ground path in the system, in series. That is the catch and the strength at the same time. On a multi-grounded system with many electrodes tied together, all the other paths in parallel add up to nearly nothing, so the loop reading is close to the resistance of the one electrode you clamped. Fast, no pins, no disconnect, done in seconds, and you can read every rod on a tower line in the time it takes to set up one fall-of-potential test.

The limit is hard and it traps people. The clamp-on needs a parallel return path to work. Clamp it on a single isolated rod, like a typical house with one electrode and no other ground path, and there is no loop to read, so the number is meaningless. It also cannot test an electrode that is disconnected, since disconnecting it opens the loop. Single electrode, or one you have isolated to test? That is a fall-of-potential job, not a clamp-on job.

Fall-of-potential vs clamp-on: which to use

Both methods read ground resistance, and they are not interchangeable. Pick by whether the electrode has a parallel path and whether you can take it out of service.

Fall-of-potential is the accurate, reference method, and it is what an acceptance spec almost always means by a ground test. It measures the true resistance of the electrode in isolation, it works on a single rod, and it produces the plateau plot that proves the reading is valid. The price is time, room to run the pins out a long way, and disconnecting the electrode. On a tight urban site with no open ground to drive pins, it can be hard to run at all.

The clamp-on is the fast field check. No pins, no disconnect, seconds per electrode, good for the periodic survey of a system with many interconnected grounds: tower lines, pole grounds, a yard full of rods tied together. It cannot test a single isolated electrode and it cannot test a disconnected one, and because it reads a loop, it includes the rest of the system rather than the one electrode alone.

The practical split most techs use: fall-of-potential for commissioning and acceptance, where you need the true number on record and the electrode can be isolated, and clamp-on for routine maintenance checks on multi-grounded systems, where speed matters and the parallel paths are real. When the two disagree, the fall-of-potential is the one that goes in the report.

The two-point (dead-earth) method

The two-point or dead-earth method is the quick option, and it is worth knowing mostly so you know its limits. You measure the resistance between the electrode under test and a second, known low-resistance ground already in place, commonly a metal water pipe system in earth contact, with no auxiliary pins.

The reading you get is the two grounds in series: your electrode plus the reference ground plus the wire between them. It works only when the reference ground is so much lower than the electrode you are testing that you can ignore it, and you have to trust that it is, because the method gives you no way to separate the two. There is no plateau, no proof, just a single series number.

Use it for a rough go or no-go where a known good reference ground exists and you only need to confirm an electrode is in the same ballpark, not put a defensible number on record. For anything that has to hold up, acceptance or a real value, it is fall-of-potential. The two-point method tells you a ground is roughly there. It does not tell you what it is.

The ground resistance tester and its leads

The instrument is a dedicated ground resistance tester, the type long known by the Megger brand name, and it is built differently from a multimeter. It sources its own test current, usually at a frequency offset from 60 Hz so it can reject the stray power-frequency current already in the earth, and it reads the small resulting voltage cleanly.

The terminal count tells you what it can do. A four-terminal instrument runs both the Wenner soil resistivity traverse and the fall-of-potential test. A three-terminal instrument runs fall-of-potential but not the four-pin soil test. The kit includes the auxiliary stakes for the current and potential pins and long reels of lead wire, because the current pin on a fall-of-potential test can sit a hundred feet or more from the electrode. The clamp-on stakeless tester is a separate instrument, and many techs carry both: the clamp for fast loop checks, the staked tester for the acceptance number.

Two field habits keep the readings honest. Get good contact at the auxiliary pins, because dry or loose pins raise the auxiliary resistance and can throw the reading or stall the test; wetting the pin holes helps in dry soil. And keep the leads from running alongside energized cable where induction can couple in, which is one more reason the test current sits off the power frequency.

Using soil resistivity to design the electrode system

The reason to measure soil resistivity before driving anything is that the resistivity profile tells you what electrode design will actually hit the target, instead of finding out after the steel is in the ground. The Wenner traverse gives resistivity versus depth, and the design follows the conductive soil.

If the low-resistivity soil is near the surface and the deeper layers are worse, a horizontal design wins: a ground ring around the building, a buried grid, or radial conductors that stay in the good soil near the top. If the surface is dry or rocky and the conductive layer is deep, you drive rods long enough to reach it, because a 10 ft rod that never leaves the bad layer does little. The profile decides rod length, rod count and spacing, and whether a ring or grid is the better spend.

This is the handoff to the install. How those electrodes get chosen, bonded into one system, and sized, the Ufer, the rods, the grounding electrode conductor, lives in the grounding electrode system and bonding field guide. The testing job here is to supply the resistivity number the design is built on, and then to verify with fall-of-potential that the finished electrode hit the target the profile predicted.

Lowering a ground that reads too high

When an electrode tests over its target, you have a short list of fixes, and they work in proportion to how much earth you can reach and how conductive you can make it.

Depth beats width almost every time. Driving a rod deeper reaches soil that stays moist and unfrozen year round, so a longer rod usually does more than a second short one beside it. If you do add rods, space them at least a rod-length apart, because rods crowded together share the same shell of soil and the second one barely helps. Beyond rods, a ground ring or a buried grid spreads the contact over a large area and pulls the number down, which is why large facilities use them.

When the soil itself is the problem and you cannot drive deeper, ground enhancement material is the lever. Bentonite clay and engineered ground enhancement compounds packed around the electrode hold moisture and lower the resistivity right where it matters, in the first shells of soil against the conductor. They take time to work. The material has to absorb water and cure, often a week or more, before a retest reads true, so do not test the morning after you backfill and trust a falsely high number. A concrete-encased electrode, the Ufer, is the same idea built into the footing, and on new work it is the cheapest low ground you will ever get.

The install detail for all of these, the rods, the ring, the Ufer, lives in the grounding electrode system and bonding field guide. The testing job is to confirm the fix actually moved the number, after the enhancement material has had time to cure.

Testing a large ground grid

A substation or large data center does not sit on two rods. It sits on a ground grid: a buried mesh of conductor bonding the whole site into one low-resistance, low-difference plane. Testing one is a bigger job than testing a rod, and it asks two different questions.

The first is the overall grid resistance to remote earth, measured by fall-of-potential, but on a grid the resistance area is huge, so the current pin has to go very far out, often several times the diagonal of the grid, to clear it. Run the pins short on a grid and you get the same no-plateau, invalid reading as on a rod, only it is easier to fool yourself because the number looks plausible.

The second question, and the one that actually protects people on a high-current site, is step and touch potential. During a fault a grid does not stay at one voltage; the surface potential rises, and the difference a person bridges between their feet, the step potential, or between hand and feet, the touch potential, is what can hurt them. IEEE 81 covers measuring earth surface potentials for exactly this, and on a substation the step-and-touch survey matters more than the single grid-resistance number. The data center and substation grounding work, the bonded plane and the grid itself, is in the grounding electrode system and bonding field guide; the verification of it, grid resistance plus step and touch, is the testing job.

When to test, and how often

Test at commissioning, and test again on a cycle. The acceptance test, run before the system is energized and put in service, is the one that proves the ground was built to the spec and gives you the baseline every later test is compared against. Skip it and you have nothing to compare to when a number drifts.

Grounds do not hold still. Connections corrode, especially the buried ones in wet or chemically active soil, electrodes get damaged by excavation, and soil conditions change as a site is developed around it. So a maintenance retest on a regular interval catches the slow failures before they matter. NETA, the body that publishes acceptance and maintenance testing standards for electrical power equipment, gives the framework for both the acceptance test and the periodic maintenance test of grounding systems, and the interval is set by the criticality of the site and the spec. A substation or data center is tested more often than a warehouse.

Time the retest to the worst case where you can. A ground that is going marginal will show it in the dry or frozen season, not after a week of rain, so a test scheduled for the hard part of the year is more honest than one taken when the soil is at its best. Note the season and soil condition on every test, because a number without its conditions cannot be compared to the next one.

Safety during the test

Ground testing puts you out in the yard with long leads, auxiliary pins, and sometimes a disconnected ground on a live system, and each of those is a way to get hurt.

The disconnect is the sharp one. Opening a grounding electrode conductor on an energized system can break the reference for equipment and, if there is fault or neutral current present, can put that current across the gap you just opened, which means across you or your tools. Treat an in-service ground as potentially current-carrying. Where the system can be de-energized for the test, do that; where it cannot, the clamp-on method that reads without disconnecting is the safer tool.

Out in the field, the test current and a faulted or energized grid make their own hazard. On a high-current site a fault during the test can drive the grid and the surrounding earth to a dangerous potential, so you keep clear of the auxiliary pins and leads while the instrument is injecting, and you do not handle a pin and the electrode at the same time, because that is step-and-touch potential with your own body as the meter. Keep the leads out from under foot and away from traffic, and treat every pin as live while the test is running.

What to record

A ground test that lives only in someone's memory is a test that has to be re-run the next time anyone asks. The record is what makes the next reading mean something, because a ground resistance number is only useful against its baseline and its conditions.

Capture the test method, the instrument and its terminal count, the electrode or grid tested and whether it was isolated or read in place, the measured resistance, the target it was held against, and the spacing geometry for a fall-of-potential test so a reviewer can confirm the current pin was far enough out. For soil resistivity, record the pin spacings and the reading at each, so the profile can be rebuilt. And record the conditions every time: the date, the season, recent rain, and whether the ground was frozen, because the same electrode reads differently across the year and a number without its conditions cannot be compared.

A field tool that timestamps the reading, holds the photo of the plateau plot or the meter face, and ties it to the site and the electrode keeps the baseline and every retest in one place instead of on a clipboard that walks off the job. The point of the record is the comparison: this electrode, this number, these conditions, against what it read last time and what the spec demanded.

Common mistakes

• Reading an electrode in fall-of-potential without disconnecting it, so parallel paths pull the number low and wrong.
• Running the current pin too short, so no plateau forms and the reading taken off the slope is meaningless.
• Putting the potential pin somewhere other than about 62 percent, or trusting one point instead of plotting the curve.
• Using a clamp-on on a single isolated rod with no parallel return, where the loop reading means nothing.
• Treating the 25-ohm figure as a system performance target instead of the NEC supplemental-electrode trigger.
• Testing only in wet weather and never seeing the dry or frozen worst case.
• Designing the electrode system with no soil resistivity data, then driving steel by trial and error.
• Retesting ground enhancement material before it has absorbed water and cured, and trusting the falsely high reading.
• Accepting a high resistance because the breaker still trips, confusing the earth reference with the fault-clearing path.
• Recording a resistance with no note of method, geometry, or soil conditions, so it cannot be compared later.

Field checklist

Standards and references

IEEE 81, the guide for measuring earth resistivity, ground impedance, and earth surface potentials of a grounding system, is the document the testing methods come from: the Wenner four-pin soil resistivity traverse, the fall-of-potential method with its 62 percent rule and plateau, and the step-and-touch surface potential survey for grids. When an acceptance spec says test the ground, it usually means an IEEE 81 fall-of-potential test.

The NEC, NFPA 70, is where the 25-ohm figure lives, at 250.53, and it is the single most misread number in grounding. It is the trigger for adding a supplemental electrode to a single rod, not a resistance the finished system must achieve, and once a second electrode is in, the code asks for no number. The low targets people quote, the 5 ohms and the 1 ohm, are recommendations from IEEE guidance and from owner and industry specs such as BICSI 002 for data centers, not NEC mandates. NETA publishes the acceptance and maintenance testing standards that set what gets tested and how often on commissioned and in-service gear, including the grounding system.

Two honest hedges belong on every grounding submittal. Section numbers and target ohms shift between code cycles and between specs, so confirm them against the adopted code edition, the project specification, and the equipment listing before you cite them. And follow the ground tester manufacturer's instructions for the instrument you are holding, because the pin spacing, the lead handling, and the valid measurement range are specific to the meter.

Units, terms, and abbreviations

Ground testing carries a small vocabulary that gets used loosely, and a couple of pairs that look alike but mean different things.

Resistance and resistivity are the pair to keep straight, because they are not the same unit or the same thing. Ground resistance is in ohms and belongs to a specific electrode in specific soil. Soil resistivity is in ohm-meters, sometimes ohm-centimeters in older tables, where 1 ohm-meter equals 100 ohm-centimeters, and it belongs to the dirt regardless of what electrode sits in it. Spacing and depth on the Wenner test share units, feet or meters, and the reading reflects soil to about the spacing depth.

Ground resistance

The resistance from an electrode through the soil to remote earth, in ohms; the output you measure

Soil resistivity

A property of the soil itself, in ohm-meters or ohm-centimeters; the input that drives ground resistance

Fall-of-potential

The accurate three-point electrode resistance test, using a current pin and a potential pin

62 percent rule

The potential pin position, about 62 percent of the way to the current pin, where the reading is valid

Plateau

The flat part of the resistance-versus-distance plot that confirms a valid fall-of-potential reading

Stakeless / clamp-on

A loop ground test that needs no pins or disconnect but requires a parallel return path

Wenner method

The four-pin soil resistivity test, where the reading reflects soil to about the pin spacing

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What is the difference between grounding and bonding?
• What is the difference between grounding and bonding?
• What is the 62 percent rule in fall-of-potential testing?
• What is the difference between conventional and addressable fire alarm systems?
• What is a separately derived system?