
Average versus percentile: two different questions
The CFE bill gives you one number per billing period. It is an average: it adds up the whole month and divides. It is useful for calculating your penalty or your credit, and nothing else.
The Grid Code (Código de Red) does not ask what your average was. It asks whether the value held, and for that it is measured in short intervals and assessed on what proportion of those intervals comply. That is a percentile criterion, and it is a completely different question.
The difference matters because the two answers can point in opposite directions. A plant with a 0.98 average and motor starts that drag it down for twenty minutes every shift has a bill that looks fine and a profile that does not comply. The rest of the month swallowed it in the average.
Why the average hides exactly what matters
Think about what a monthly average does: it spreads the twenty bad minutes across the forty-three thousand two hundred minutes of the month. They vanish. But those twenty minutes happened, they were measured, and under a percentile criterion they count as non-compliance.
And they are precisely the minutes that cause everything else: the simultaneous start that sets your billable demand, the transient that trips the protection, the moment the capacitor bank cannot keep up. The average erases exactly the information your plant needs.
That is why a serious compliance report is not made from twelve bills. Bills say how much you paid; they do not say what happened.
What actually answers the question
A measurement campaign. A Class A analyzer installed at the service entrance and at the suspect switchboards, for at least seven calendar days, per IEC 61000-4-30.
Seven days, not two, because you need to capture the full operating cycle: the shifts, the weekend, the Monday startup and the maintenance shutdown. Any given Tuesday does not represent your month.
Class A, not just any instrument, because the standard defines how each parameter is aggregated, over what window and with what accuracy. That is what makes two different analyzers, at the same point and at the same time, give the same result — and what makes your compliance report defensible before a third party.
That campaign yields the data the bill will never give you: the full distribution, not the average. And with it, whether you comply or not, and at exactly which moments you stop complying.
A warning about the figures going around
Percentages, thresholds and dates circulate on this topic that almost no material cites against its instrument published in the Official Gazette (Diario Oficial). We checked several and could not find the reference: repeating them would be easy and would do you no good.
What is verifiable and sufficient to decide is the above: the current framework is RES/550/2021, published in the DOF on December 31, 2021, and the power factor requirement for load centers is set out in the Load Center Connection Regulatory Manual.
When we present a number to you, it will come with its reference. And if someone presents one without it, ask for it before making an investment decision on that basis.
Your bill answers how much you paid. The standard asks how your plant behaved. Only a measurement campaign answers the second, and that is the one that decides whether you have a problem.











