STS Prepaid Meters and the TID Rollover Explained

A prepaid meter has no idea what year it is. It counts minutes from a date fixed in 1993, in a field that was only ever 24 bits wide — and a counter that narrow eventually runs out of room.

That single design decision has quietly forced the biggest coordinated field operation in prepayment metering history. For anyone buying STS meters for an estate, a utility or a housing scheme, it is also a quick way to tell a current product from an obsolete one.

We supply metering and utility management systems into East and West African markets, and this is the question we ask a supplier before any other. Simfy Exim sources these systems; we do not operate the vending platform or hold the keys, so what we can do is make sure the hardware you buy is on the right side of the rollover.

What is STS and who controls it?

What makes STS unusual is that it is genuinely open at the point of sale and tightly closed at the point of cryptography. Any vendor in the network can sell a token for any compliant meter, which is what makes prepayment work in markets with thousands of small retail outlets. The security sits underneath: CLOU describes a Key Management Centre operated by Eskom in South Africa, which handles Supply Group Codes and security module initialisation, and notes that licensing through the Association ensures appropriate encryption key management practices. A meter belongs to a supply group; tokens are generated against that group’s key; nothing else will unlock it.

What is the TID and why did it run out?

The TID exists to stop a token being used twice. A meter remembers the identifiers it has already accepted and refuses repeats, which is what makes a token safe to print on a slip of paper and hand across a counter. That protection depends on the identifier always increasing. When the counter wraps back to zero, newly issued tokens carry identifiers the meter has already seen and treated as spent, so the meter does the correct thing and rejects them. The failure is not a bug; it is the anti-replay protection working exactly as designed, against a clock that has run out of digits.

What was the deadline, and what happens to a meter that missed it?

That is the whole commercial consequence in one sentence. A meter that was not updated does not fail visibly, does not alarm and does not stop supplying whatever credit it still holds. It simply refuses the next token the customer buys, and it refuses every token after that. For a utility, the failure arrives as a wall of support calls from customers who have paid and cannot get power. For an estate or a landlord running submetering, it arrives as a building full of residents with no way to top up. This is why the rollover status of a meter is a purchasing question rather than a technical footnote.

How big was the affected population?

Those numbers explain why the transition took years rather than months, and why it shaped the market for new meters. A utility with a million meters in the field could not treat this as a firmware push. Every one of those meters represented a physical visit, a customer interaction and a support risk, and the work had to be sequenced against ordinary operations. For buyers today the practical inheritance is that the installed base and the supply chain both moved, and a meter that predates the change is not a bargain — it is a liability someone else has already decided not to carry.

How is a meter actually updated?

The two-part nature of the operation is what makes it slow. Clearing the stored identifiers frees the meter to accept the new numbering; changing the cryptographic key moves the meter onto the new base date so that future tokens are generated correctly for it. Both steps are entered at the meter itself, which is why the customer-performed route exists at all — issuing the sequence to a household and asking them to key it in is far cheaper than sending a technician to every wall. It is also why the operation generated so much support load: a step keyed in wrongly leaves a meter in a state the customer cannot diagnose.

What should a buyer check before ordering STS meters?

The Supply Group Code is the one that catches new buyers. Meters are initialised against a supply group, and a batch initialised for the wrong group is not a configuration problem you fix on site — it is stock that will not accept tokens from your vending system. That decision has to be made before the meters are keyed at the factory, which means it belongs in the enquiry rather than in the delivery discussion. We ask for it up front for that reason, and where a buyer does not yet have a vending arrangement we will say so rather than shipping meters that cannot be sold against. The same discipline runs through the rest of what we source; see supplier verification for what a factory audit can and cannot establish.

Prepaid or postpaid: which suits an estate?

The trade is between capital and receivables. A prepaid installation costs more per point and requires a vending arrangement to exist and keep working; in exchange, nobody consumes electricity that has not been paid for, and the awkward business of disconnection largely disappears. Postpaid is cheaper to install and needs a meter reading process, a billing run and a collections function behind it. Where an estate already has strong collection and stable tenants, postpaid with remote reading is often the calmer answer. Where arrears are the operator’s main problem, prepayment addresses the problem directly rather than managing it.

Where does the water side fit?

The figures that matter on a water point are Q1, the minimum flow at which the meter still meets its error limits, Q3, the permanent flow rate under normal conditions, and the R-ratio that divides them. YOUNIO lists common market R-values running from R40 up to R1000, with fixed relationships of Q2/Q1 = 1.6 and Q4/Q3 = 1.25. A high R-ratio matters in residential submetering because the meter has to register the slow trickle of a dripping tap as honestly as it registers a filling bath — and unregistered trickle across a few hundred apartments is where an estate’s water losses actually live.

How accurate does a water meter have to be?

The same source gives Class 1 as plus or minus 1 per cent in the upper zone and plus or minus 3 per cent in the lower zone, with both classes permitted a wider margin above 30 °C — 3 per cent for Class 2 and 2 per cent for Class 1. Those lower-zone allowances are the ones an estate operator should read carefully. A five per cent tolerance on slow flows sounds small until it is applied to the hours each day when a building is only leaking rather than using, which is exactly the flow band a submetering scheme is trying to capture.

What sits behind the meter?

An estate metering package is really three purchases stacked together — the meters, the communications, and the software that turns readings into money. Buying only the first and assuming the rest will follow is the most common way these projects stall. It is worth deciding early who operates the vending relationship, how residents top up, and what happens on a public holiday when nobody is in the office, because those answers change the hardware specification. Where water and gas points sit alongside electricity, the same question applies across all three, and a single management layer is usually cheaper than three parallel ones.

Does the origin of the meter matter?

We source country-agnostically. India is one important origin among the network, not the default, and we do not operate local offices or hold local stock unless the quotation says so. What we fix first is the technical and key arrangement — compliance, base date, supply group, vending compatibility, communications — then source against it so offers can be compared like for like. If the quotations in front of you use different delivery terms, FOB, CIF and DDP are not three prices for one thing.