Why Your Sub-Meter Will Never Match TNB's Meter

Your meter and TNB's meter will never agree, and you can't win that argument

Install energy monitoring, run it for a month, then line your numbers up against the TNB bill. They won't match. Your total kWh will be a little off, your maximum demand will land on a different number, and sooner or later someone in a meeting will ask the obvious question: which meter is right? The uncomfortable, honest answer is that the question has no useful answer. Your meter isn't wrong and TNB's isn't wrong. They are two different instruments measuring slightly different things in slightly different ways, and there is no neutral referee who will declare a winner. Understanding why saves a lot of wasted arguing, and it changes what you should expect your own metering to do for you.

They were never measuring the same thing the same way

Two meters on the same supply disagree for reasons that are baked into the physics and the standards, not because one is faulty.

Different accuracy classes. TNB's meter is a revenue-grade instrument, built and sealed to a tight accuracy class (typically around 0.2% to 0.5% under reference conditions) with revenue-grade current transformers to match. A typical sub-meter or monitoring meter is held to a looser class, and the headline figure on its datasheet is only part of the story.

The CTs are the real source of error. Most of the inaccuracy in a metering installation lives not in the meter but in the current transformers feeding it. The combined system accuracy of a meter plus its CTs is always worse than the meter's own rating, and it degrades badly at low load. The relevant standard only specifies CT accuracy down to 10% of rated current, and allows the error to roughly double there, so a CT that's tight at full load can drift at a lightly loaded one. Oversize the CT for the actual current and it gets worse fast. A well-documented example of a 600:5 CT used on circuits averaging only 80 amps produced an 8% billing discrepancy on its own. Two correctly working meters with different CTs will simply read differently.

Different measurement points. TNB measures at the point of supply. Your meter usually sits somewhere downstream, which means small distribution losses, and sometimes whole sub-circuits, live in the gap between the two. You are not metering the same electrons at the same place.

Different demand windows. Maximum demand is the worst offender. MD is an average over an interval, and the number depends entirely on where that interval starts and stops. TNB's metrology samples and integrates demand on its own clock and its own block, and your meter does it on yours. Feed two meters the identical load with their 30-minute windows misaligned by even a few minutes and they will report different peaks. We get into how the peak is defined in how to calculate maximum demand.

Stack these together and a stable difference of a couple of percent on energy, and a more noticeable gap on demand, is the normal, healthy state of affairs. It is not a defect to be chased to zero.

The part nobody can argue with: TNB's meter is the legal one

Here is the bit that ends the meeting. TNB's meter is the revenue meter. It is the contractual, legally recognised instrument of record, sealed, type-approved and periodically tested. Your monitoring meter, however good, is not a legal revenue instrument, so it carries no standing to dispute the bill. You cannot walk into TNB with your own sub-meter readings and demand a correction, because the basis of the contract is their meter, not yours.

The only real recourse, if you genuinely believe the utility meter is faulty, is to request that TNB test or verify their own meter through their formal process. That checks the revenue meter against a reference standard. It does not adjudicate your meter against theirs. So the dispute you might imagine having, "my data says my MD was lower, so my bill is wrong," is not a dispute you can actually have. There is no forum for it.

This sounds discouraging only if you bought monitoring for the wrong reason.

So what is your own meter actually for?

Not for arguing the bill. For understanding it. The value of your own metering was never absolute agreement with TNB. It is everything the revenue meter can't tell you:

- Attribution. TNB gives you one number for the whole site. Your meters tell you which building, which floor, which chiller, which tenant, which process is behind it. That's the entire point of sub-metering, and the revenue meter is blind to all of it.
- Trends and anomalies. Your meter watches consumption continuously and flags when something drifts, the value the monthly bill can never deliver because it arrives too late and too coarse.
- Verifying savings. When you make a change, your own metering is how you prove it worked, the basis of measurement and verification.
- Managing the peak before it's set. You use your demand reading to see the peak forming and act inside the window, which is how you actually lower the maximum demand charge. The fact that your MD figure won't exactly equal TNB's billed MD is irrelevant to that job.

For every one of these, perfect agreement with the utility meter is unnecessary. What matters is that your meter is consistent, and that it moves with the bill.

The rule that makes the discrepancy useful instead of annoying

Stop trying to make the two numbers equal. Watch the offset between them instead. A healthy installation shows a small, stable gap month after month: your meter always reads, say, two percent under the bill, and that's fine. The offset is your friend once it's stable, because then your meter is a faithful proxy for the billed total and you can manage against it with confidence.

What you actually care about is the offset changing. If your meter tracked the bill within a steady couple of percent for a year and then suddenly diverges, that's a real signal: a CT problem, a wiring fault, a metering error, or a genuine change in what's being measured. The discrepancy you were tempted to eliminate becomes a diagnostic the moment you treat it as a baseline rather than a flaw.

This is the honest position we take with clients, and it's how CobiNeural is set up to work: monitoring calibrated to be consistent and reconciled against the TNB bill, not pretending to replace the revenue meter. Its billing and tariff modelling expects a gap between measured and billed and works with it, rather than promising an exact match it can't deliver.

Your meter will never agree with TNB's to the last kWh, and you will never win a dispute by waving its readings. Once you accept that, your monitoring stops being a second opinion on the bill and becomes the thing it was always meant to be: the only view that tells you why the bill is what it is. If you'd like metering set up to manage your energy rather than to argue about it, book a CobiNeural walkthrough.