How much loss occurs between inverter output and the revenue meter?

Typically a few per cent of gross generation, split between transformer no-load losses (constant), transformer load losses and cable losses (both rising with the square of current), and auxiliary consumption. The exact figure depends on the collection-system design, cable runs and transformer specification. The number matters less than its behaviour: a stable MV loss is a fixed cost to design around, while a growing one is an early sign of degradation. Because the loss sits between two good measurements — inverter output and the meter — it can be quantified from data you already collect.

Can transformer degradation be detected without MV instrumentation?

To a useful degree, yes. A degrading transformer shows up as a rising load loss at equivalent throughput and temperature, which appears as a drift in the inverter-to-meter ratio over time. Normalising for ambient temperature separates ordinary resistive heating from genuine degradation. This inference will flag that something in the MV path is worsening and characterise how; confirming the specific component and its severity is where direct instrumentation — transformer metering, temperature or dissolved-gas monitoring — earns its place.

What causes a growing inverter-to-meter divergence?

A widening gap between summed inverter output and metered export, at equivalent conditions, points to something degrading in the MV path: a loosening connection adding resistance, a transformer whose windings or insulation are deteriorating, a developing cable fault, or rising auxiliary consumption. Because the divergence grows rather than staying fixed, it is distinguishable from the plant's baseline MV losses — and catching it as a trend, before it becomes a failure, is the entire value of watching the ratio.

The MV Blind Spot: Losses Between the Combiner and the Revenue Meter

Ellume Engineering

Most solar analytics stop at the inverter or the combiner. But energy still has to travel through medium-voltage transformers and cabling to reach the revenue meter — and that path loses a few per cent, some of it fixed and some of it growing as transformers age. Because nothing measures inside the gap, the loss gets attributed to the array or written off as noise. You can infer it from inverter and meter data you already have.

Key Takeaways

→The medium-voltage path sits between two trusted measurements — summed inverter output and the revenue meter — yet almost nothing monitors inside it.
→MV losses are typically a few per cent of gross generation: large enough to matter, small enough to get rounded into 'other'.
→The losses decompose into fixed no-load losses, load-dependent transformer and cable losses, auxiliary consumption, and developing faults — and each behaves differently.
→Because no-load losses are constant while load losses rise with the square of throughput, the inverter-to-meter ratio across the production range separates them without new instrumentation.
→A drifting inverter-to-meter ratio at equivalent conditions is the earliest warning of a degrading transformer or cable — a trend caught long before a failure.

There is a stretch of your plant that almost no monitoring system watches, and it sits between two numbers you trust completely. On one side is inverter output — measured, granular, dashboarded to death. On the other is the revenue meter — the number you actually get paid on, measured to settlement-grade accuracy. Between them lies the medium-voltage collection system: the transformers that step the inverter output up to collection voltage, the buried cabling that carries it across the site, and the substation that hands it to the grid. Energy is lost all along that path, and almost nothing measures inside it. The result is a blind spot with a specific, expensive property: it is the last thing between your generation and your revenue, and it is the thing you understand least.

You measure what the inverters made and you measure what the meter sold. The difference is real money, and most plants call it noise.

The measurement gap between combiner and revenue meter

Follow a megawatt-hour from where it is generated to where it is sold. It leaves the modules as DC, becomes AC at the inverter — the last point most analytics measure in detail — and then begins a journey the monitoring system largely stops watching. It passes through a step-up transformer, incurring core losses (constant, present whenever the transformer is energised) and winding losses (rising with load). It travels along MV cable, losing energy resistively in proportion to current squared. It may pass through a second transformation at the substation. Finally it reaches the revenue meter. Each step is a small percentage. Together they are typically a few per cent of gross generation — a number large enough to matter and small enough to ignore, which is the worst possible combination, because it is exactly the size of loss that gets rounded into 'other'.

What actually lives in that gap

The losses in the MV path are not a single thing, and distinguishing them is what turns 'we lose a few per cent somewhere' into an actionable picture.

•Transformer no-load (core) losses. Present whenever the transformer is energised, regardless of production. At night, a solar plant's transformers can quietly consume energy. Over a year, no-load losses are a fixed tax that many energy models omit entirely.
•Transformer load losses. Rising with the square of load, these peak when the plant is generating hardest. They are the price of moving power, and they grow as a transformer's windings age and its insulation degrades.
•MV cable resistive losses. Also proportional to current squared, and dependent on cable length, cross-section and temperature. A hot day increases cable resistance exactly when current is highest, compounding the loss.
•Auxiliary and parasitic loads. Tracker motors, inverter cooling, monitoring, security, lighting — the plant's own consumption, which sits between generation and export and is easy to overlook.
•Connection and degradation faults. A loosening connection, a partially failing transformer, a developing cable fault — these show up as a growing gap between inverter output and metered export, and they are the ones worth catching early.

Inferring MV losses from inverter and meter data

The elegant thing about this blind spot is that it is bracketed by two good measurements. You know what went in (summed inverter output) and what came out (the revenue meter). The difference is the total MV-path loss, and while a single number, it can be decomposed with the data you already have.

1.Separate the fixed from the load-dependent. No-load losses are constant; load losses rise with the square of throughput. By looking at the inverter-to-meter ratio across the full range of production levels — from near-zero at dawn to peak at midday — you can separate the fixed component (visible at low load, and even at night as consumption) from the load-dependent component (which grows with output). The shape of that relationship is a fingerprint of the collection system's health.
2.Track the ratio over time. A healthy MV system has a stable inverter-to-meter relationship. When that relationship drifts — the same inverter output yielding progressively less metered export at the same conditions — something in the path is degrading. This is the single most valuable signal the blind spot offers, and it requires no new instrumentation, only the discipline to compute the ratio and watch it.
3.Normalise for temperature. Because both transformer and cable losses depend on temperature, a fair comparison over time has to account for ambient conditions. A gap that grows only on hot days is consistent with resistive heating; a gap that grows regardless of temperature points to something else, like a developing connection fault.

Transformer degradation signatures

Transformers fail slowly and then suddenly, and the slow part is visible if you know where to look. Rising load losses at equivalent throughput and temperature suggest winding or insulation degradation. A growing no-load loss can indicate core problems. Where transformer temperature or dissolved-gas data is available, it corroborates the electrical signature. The point is that a transformer's decline shows up as a trend in the inverter-to-meter relationship long before it becomes a failure — if anyone is tracking the relationship.

When the gap is worth closing with hardware

Inference from inverter and meter data will localise a growing loss to the MV path and characterise its behaviour, but it will not always pinpoint the exact component. Direct MV instrumentation — metering at the transformer, temperature and dissolved-gas monitoring, cable-fault sensing — is warranted when the inferred loss is large or growing, when the plant is valuable enough that a transformer failure would be catastrophic, or when a specific degradation signature needs confirming before a costly intervention. For most operators the sequence is: infer first from data you have, instrument second where the inference points.

A reconciliation method

The disciplined way to handle the MV blind spot is to fold it into a full energy reconciliation: start from expected generation given the weather, deduct every named loss in sequence — array, inverter, MV path, auxiliary — down to metered export, and refuse to leave a residual. When the MV path is an explicit line in that reconciliation rather than part of an unexamined 'other', two things happen. The fixed cost of moving power becomes visible and can be designed around. And any growth in that line becomes an early-warning signal for a degrading transformer or cable, caught as a trend rather than as a failure.

Frequently Asked Questions

How much loss occurs between inverter output and the revenue meter?: Typically a few per cent of gross generation, split between transformer no-load losses (constant), transformer load losses and cable losses (both rising with the square of current), and auxiliary consumption. The exact figure depends on the collection-system design, cable runs and transformer specification. The number matters less than its behaviour: a stable MV loss is a fixed cost to design around, while a growing one is an early sign of degradation. Because the loss sits between two good measurements — inverter output and the meter — it can be quantified from data you already collect.
Can transformer degradation be detected without MV instrumentation?: To a useful degree, yes. A degrading transformer shows up as a rising load loss at equivalent throughput and temperature, which appears as a drift in the inverter-to-meter ratio over time. Normalising for ambient temperature separates ordinary resistive heating from genuine degradation. This inference will flag that something in the MV path is worsening and characterise how; confirming the specific component and its severity is where direct instrumentation — transformer metering, temperature or dissolved-gas monitoring — earns its place.
What causes a growing inverter-to-meter divergence?: A widening gap between summed inverter output and metered export, at equivalent conditions, points to something degrading in the MV path: a loosening connection adding resistance, a transformer whose windings or insulation are deteriorating, a developing cable fault, or rising auxiliary consumption. Because the divergence grows rather than staying fixed, it is distinguishable from the plant's baseline MV losses — and catching it as a trend, before it becomes a failure, is the entire value of watching the ratio.

Where this leaves you

The most expensive stretch of your plant to lose energy in is the one your analytics stop short of — the path from the inverter to the meter. But it is bracketed by two measurements you already trust, and the difference between them tells you both the fixed cost of moving your power and the early warning of a transformer or cable beginning to fail. Make the MV path an explicit line in your energy reconciliation, and the blind spot stops being blind.