Grading margins of around 0.3 to 0.4 seconds are often quoted as a starting point for overcurrent coordination. They are useful rules of thumb, but they are the result of a set of assumptions — not a fixed rule. Knowing where the number comes from tells you when it can reasonably be reduced, and when it should not be.
What the margin has to cover
When a fault occurs downstream, both the downstream and the upstream relay start timing. The upstream relay must not reach its trip point before the downstream breaker has interrupted the current and the upstream relay has stopped timing. The margin therefore has to absorb:
- The interrupting time of the downstream circuit breaker — take this from the breaker's data rather than a typical figure.
- Overshoot of the upstream relay: the time it continues towards operation after the current has gone. Small for numerical relays, more significant for induction-disc relays.
- Timing errors of the relays, as declared by the manufacturer for the characteristic in use.
- CT errors, which shift the current the relays actually measure.
- A safety allowance for what has not been modelled.
Building the number
One commonly published formulation adds the percentage timing errors of the two relays and the CT error, applies them to the downstream relay's operating time t, and then adds the fixed terms:
t′ = ((2·E_R + E_CT) / 100) · t + t_CB + t_o + t_s
Here E_R is the relay timing error (%), E_CT the CT error (%), t_CB the breaker interrupting time, t_o the relay overshoot and t_s a safety allowance. As an illustration only, assume E_R = 5 %, E_CT = 10 %, t = 0.5 s, t_CB = 0.06 s, t_o = 0.02 s and t_s = 0.033 s. The result is 0.10 + 0.06 + 0.02 + 0.033 ≈ 0.21 s. With assumptions more typical of electromechanical relays — E_R = 7.5 %, t_o = 0.05 s and t_s = 0.1 s — the same arithmetic gives about 0.34 s.
These figures illustrate the method. They are not recommended values: the inputs should come from the actual relays, CTs and breakers, and the criteria agreed for the project.
The margin is a budget. Every millisecond not needed for uncertainty is time the upstream stages do not have to wait.
One margin is not a coordination study
Inverse-time curves converge and diverge across the current range, so a margin that is adequate at one fault current may be too small at another. Coordination is normally checked across the range of currents each pair of devices may see — including minimum and maximum fault conditions, and any definite-time stages — and repeated when the system changes.
Why it matters during modernization
When older relays are replaced with numerical devices, margins that were sized for electromechanical relays may be larger than necessary. Recalculating them with the new devices' data may allow faster upstream clearing times, depending on how many grading stages there are and on the other constraints in the system.
Our coordination tool lets you change curves, multipliers and the fault current and see the time margin at the selected current. It is an educational illustration, not a substitute for a study.