Most days, an equipment grounding conductor does nothing at all. It carries no load current. It does not light anything up. It just waits.
Then a fault happens, and that quiet wire suddenly has the most important job in the circuit.
What it actually does in a fault
Say a hot conductor touches a metal enclosure. Without a low-impedance path back to the source, that enclosure can sit at a dangerous voltage while nothing trips. Someone touches it, and they become the path.
The equipment grounding conductor gives fault current a deliberate way home. Current rushes back through it, the overcurrent device sees a large fault current, and it opens the circuit. Fast.
That is the whole point. It is not there to carry normal current. It is there to make a fault loud enough, electrically speaking, for protection to act — and to keep exposed metal from staying energized while it does.
Two things matter in those few moments: the path has to be low enough impedance for enough current to flow, and the conductor itself has to survive the heating until the device clears. A wire that melts open before the breaker trips has failed at the one moment it was needed.
The table starts from the breaker, not the load
In NEC work, Table 250.122 sizes the equipment grounding conductor from the rating or setting of the overcurrent device ahead of it, not from the load current alone. A bigger breaker ahead of the circuit means a bigger minimum grounding conductor, because the fault current that device will let through before it clears can be larger.
That logic trips people up. They size the phase conductors for the load, then pick a grounding conductor to match the load as well. The table works the other way round. Look upstream first. What is protecting this circuit? That rating sets the floor.
And a floor is all it is. Meeting the table does not automatically mean the conductor will handle the heating in a real fault, or that the rest of the grounding system is adequate. It is a minimum, not a design verdict.
When the phase conductors grow, so does the ground
Here is the part that gets missed on long runs.
If you upsize the ungrounded conductors for voltage drop, the equipment grounding conductor has to grow proportionally, in circular mil area. The reason is simple once you see it: you have lowered the impedance of the phase conductors, so more fault current can flow, and the grounding conductor needs to keep pace. Leaving it at its original size while the phase conductors get bigger leaves the fault path weaker than the circuit it is supposed to protect.
It is an easy mistake, because voltage drop feels like a separate problem from grounding. On a long feeder, it is not. The two decisions are tied together.
A worked example: the thermal check
The table gives a minimum size. A separate question is whether a given conductor can absorb the heat of a short fault without getting too hot. That is a thermal check, and it uses different inputs: fault current, clearing time, and a material factor called k.
The formula is:
Amin = I x sqrt(t) / k
where Amin is the minimum thermal area in square millimetres, I is the RMS fault current in amps through that particular conductor, t is the clearing time in seconds, and k depends on the conductor material and the permitted starting and final temperatures. The companion checks are I squared t <= k squared A squared, and, for tape, minimum thickness = Amin / width.
Take the worked example from the GridCircuits calculator. It uses an illustrative fault current I = 10,000 A, a clearing time t = 1 s, and an assumed k = 143 A sqrt(s)/mm². That k value is an illustrative assumption, not a universal factor for bare copper — it has to be replaced with the value for the actual material and temperatures on a real project.
Run the numbers:
- Amin = 10,000 x sqrt(1) / 143 = 69.93 mm²
- Fault energy I²t = 100,000,000 A²s
- For a tape 25 mm wide, the required thickness is 69.93 / 25 = 2.80 mm
- A 25 x 3 mm tape has an area of 75 mm², so it passes this single assumed thermal check
Notice what that last line does and does not say. It passes this one thermal check, with those assumed inputs. It does not prove the grounding system is safe. It does not check touch and step voltage, ground-potential rise, current division, corrosion, joints, or mechanical requirements. The calculator itself is explicit about that, and it is an adiabatic check for short durations of up to 5 seconds — not a complete grounding design.
That distinction matters. A conductor can be thermally adequate and still be part of an unsafe system. The thermal check answers one question: will this piece of metal survive this much heating for this long? Other checks answer the rest.
If you want to try the same kind of thermal check with your own fault current and clearing time, you can run the numbers in the GridCircuits grounding conductor sizing calculator.
The habit to keep
Size from the protective device, not just the load. Grow the grounding conductor when you grow the phase conductors. Then check the heating, with real fault current, real clearing time, and a k value that matches your material and temperatures.
None of those steps is hard on its own. The trouble starts when one of them is skipped, usually because the table was treated as the whole answer instead of the floor it is.
These results are for preliminary design, estimating, and education. Check them against the code edition and standards adopted for your project, and have final designs checked by a qualified engineer.
Top comments (0)