At a Glance. A current transformer is the least glamorous component in an energy monitoring scheme and the one that most often destroys its accuracy. The meter is usually blamed, but the CT decides the measurement: the wrong primary ratio, the wrong accuracy class, or a secondary cable that exceeds the CT's rated burden will corrupt the reading long before the meter sees it. This guide covers the five decisions that define a CT — ratio, secondary current, accuracy class, VA burden and physical format — and then explains how Socomec turns CT output into usable data through DIRIS meters and the Digiware measurement bus.
1. Why the CT, Not the Meter, Decides Accuracy
Current transformers work by magnetic coupling. The primary current creates a flux in the core, and the secondary winding delivers a proportional current into the measuring circuit. Two error mechanisms live inside that magnetic path and neither can be corrected by the meter.
The first is ratio error — the secondary current is not exactly primary divided by the nominal ratio, because excitation current is diverted to magnetise the core. The second is phase displacement, a small angular shift between primary and secondary current which becomes a significant power and energy error at low power factors.
Both errors are small at rated current and grow as the load falls towards the bottom of the range. That is the single most important thing to understand about CT selection: a CT is accurate in the upper part of its range, not across all of it. A 1200/5 CT carrying 60 A is being used at 5 % of rated current, where no accuracy class promise applies. The right answer in that case is a smaller primary, not a bigger budget.
On top of the magnetic errors sits the secondary circuit itself. The CT is specified to drive a certain apparent power into the secondary, expressed in VA, and if the connected wiring and meter demand more than that, both ratio and phase error degrade further and the class guarantee is void.
2. Step 1 — Choose the Primary Ratio
The primary rating must cover the maximum expected load current, including a reasonable margin for growth, while staying close enough to normal load that the CT operates in the accurate part of its curve.
| Feeder | Primary rating to choose | Socomec example |
|---|---|---|
| Small distribution or final circuit | Close above maximum load | TCA21 200/5, TCA21 300/5 |
| Motor control centre outgoing | Next standard step above full-load current | TCB44-63 200/5 cl.0,5 |
| Large feeder, low current expected | Prefer a wound primary over an oversized through-feed | T2RB115 125/5 cl.0,2s |
| Main incoming, high current | Match the busbar rating | TBA103 1200/5 cl.0,5, TCB26-40 750/5 cl.1 |
Two practical notes. First, if the calculated full-load current is low but the feeder rating is high, an oversized through-feed CT is the wrong tool: the stored data for the T2RB115 125/5 cl.0,2s 7,5VA records it as a wound primary converter, the design used exactly where a through-feed CT would have to be paired with an impractically small primary rating. Second, do not confuse the CT primary with the protection setting. The CT is a measurement device; the circuit breaker or fuse downstream is what protects the cable.
3. Step 2 — Choose the Secondary: 5 A or 1 A
The standard secondary ratings are 5 A and 1 A. Socomec's wound and through-feed CTs in this range are predominantly 5 A secondary devices, which is what most European panel instruments expect.
The choice matters almost entirely because of the burden calculation in Section 5. The apparent power dissipated in the secondary circuit scales with the square of the current, so at the same cable resistance a 5 A secondary consumes twenty-five times the VA of a 1 A secondary. For long runs from a remote switchroom, a 1 A secondary CT — or a low-level sensor with a digital output — is often the only way to stay inside the rated burden.
Never connect a 5 A secondary CT to a meter whose current input is designed for 1 A, or the reverse. The measurement will be wrong by the ratio of the two and the meter input may be damaged.
4. Step 3 — Accuracy Class: 0.2S, 0.5, 1 and What They Cost
IEC 61869-2 classes ascend in refinement, and the price ascends with them. The class is a defined maximum error at defined load points, not a vague quality ranking.
| Class | Where it is used | Socomec example |
|---|---|---|
| 0.2S | Billing and revenue metering, energy performance contracts, tariff boundaries. The “S” extends the specified range down to about 1 % of rated current instead of 5 %. | T2CB32-40 500/5 cl.0,2s 5VA, T2RB115 125/5 cl.0,2s 7,5VA |
| 0.5 | General sub-metering, energy management, cost allocation inside a plant. | TCA21 200/5 cl.0,5, TCB85-100 750/5 cl.0,5 |
| 1 | Indication, trend monitoring, non-billing operational data. | TCB318-20 200/5 cl.1, TCB26-40 750/5 cl.1 |
Matching matters in both directions. Fitting a class 0.2S CT to a feeder that only drives an operator display is money wasted. Fitting a class 1 CT to a tenant billing point is worse than wasted — the billing will not survive an audit, and the meter's own 0.5S or 0.2S class cannot rescue an error that happened upstream of it. The practical rule is that the overall error of a measurement chain is roughly the vector sum of the CT error and the meter error, so the weakest link sets the standard.
One further marking to read: the FS value, the instrument security factor or overcurrent limiting factor. It states the multiple of rated primary current at which the composite error is still bounded — the T2CB32-40 records FS 5, meaning the composite error stays limited up to five times rated primary, which protects the meter from a fault current. A lower FS is protective; a high FS implies the CT will faithfully reproduce a very large fault current into the meter input.
5. Step 4 — Calculate the VA Burden
This is the step that gets skipped, and it is the reason many installations measure well at full load and badly at part load. The CT's rated VA must be at least the total burden of everything connected to its secondary.
That total has three parts: the meter's current-input burden, the contact and terminal resistance, and the connecting cable. For short runs inside a cubicle the meter dominates, but for any cable longer than a few metres the cable is usually the largest term. Copper cable burden is calculated as:
S = I² × R where R = 2 × L × ρ / A, with ρ = 0.0175 Ω·mm²/m for copper at 20 °C, L the one-way cable length in metres and A the conductor cross-section in mm².
Worked example — the same 10 m run, 2.5 mm² copper, two cases:
- 5 A secondary: R = 2 × 10 × 0.0175 / 2.5 = 0.14 Ω so S = 5² × 0.14 = 3.5 VA. That alone exceeds the 2.5 VA rated burden of a TCA21 200/5 cl.0,5 2,5VA — before the meter's own input burden is added. The class guarantee is already lost.
- 1 A secondary: the same cable gives S = 1² × 0.14 = 0.14 VA, a negligible fraction of any standard rating.
Three ways out of the problem: shorten the run, increase the conductor cross-section, or move to a CT with a higher rated burden such as the TBA103 1200/5 cl.0,5 10VA. A fourth option removes the arithmetic entirely — see the low-level sensors in Section 6.
6. Step 5 — Pick the Physical Format
Format follows the installation: what has to pass through the CT, and whether the feeder can be disconnected for the work.
| Format | How it works | Best for | Socomec example |
|---|---|---|---|
| Through-feed (cable type) | Primary conductor passes through a window; the hole diameter sets the maximum cable or bar | New panels, cables up to the window size | TCA21 200/5 (21 mm window), T2CB32-40 500/5 (40×10 mm) |
| Busbar / large window | Wide opening sized for a busbar rather than a cable | Main incoming feeders | TBA103 1200/5 cl.0,5 |
| Wound primary | The primary winding is part of the CT itself | Low primary currents where a through-feed CT would be poorly used | T2RB115 125/5 cl.0,2s |
| Three-phase block | Three CTs combined in one housing for a three-phase feeder | Compact three-phase metering, faster wiring | TCB318-20 200/5 cl.1 |
| Low-level RJ12 sensor | Solid-core sensor with a plug-in digital output feeding a measurement bus | Digital monitoring architectures; no burden calculation | TE-35 63-250A with CABLE RJ12 4x2M |
The through-feed window dimension is the specification people forget to check. Note the difference between a 21 mm circular opening on the TCA21 and the 40 × 10 mm rectangular opening on the T2CB32-40: one takes a cable, the other takes a small busbar, and neither will accept the other's primary. The T2CB32-40 data also records a copper rail and snap mounting, which reflects how it is intended to be installed.
The RJ12 family — TE, TR, iTR and TF sensors — is architecturally different again. The TE-35 63-250A is a solid-core through-feed sensor with a 21 × 21 mm opening and a plug-in secondary, connected with a CABLE RJ12 4x2M whose data sheet lists it specifically as the connection cable for TE, TR/iTR and TF current sensors. Because the output is low-level rather than 5 A, the burden calculation collapses: the sensor drives a defined input, not an unknown length of 5 A loop. This is the family to use when the monitoring architecture is digital rather than analogue.
7. From CT to Data: DIRIS Meters and the Digiware Bus
A CT produces a current; a meter turns it into information. Socomec splits that job across three levels.
Standalone analysis meters
The DIRIS A-41 is a panel meter for analysis duty. Its parameter set includes current, voltage (ammeter and voltmeter functions), frequency measurement, active power and reactive power, and pulse counting. Communications are modular: optional RS485 or Ethernet ports carrying Modbus RTU, Modbus TCP or Profibus, with a supply spanning 110–400 VAC or 120–350 VDC so it can be fed from the same control supply as the rest of the panel.
Power quality meters with low-level inputs
The DIRIS A-100 333MV takes a different approach. Rather than a 5 A input, it accepts a 333 mV signal, which means solid-core, split-core or Rogowski coils can all be connected. That matters for retrofit work, where the feeder cannot be disconnected to slide a solid CT over the bar. Its active energy accuracy is recorded as class 0.1S to IEC 62053-22, its maximum measurable current is 10,000 A, and it provides four voltage and four current measurement inputs with three digital inputs and one digital output. Communication is Modbus RTU over RS485 — a single protocol rather than a menu of options, which simplifies specification.
Display, gateway and multi-point systems
DIRIS Digiware D-70 is not a CT and not a meter in the ordinary sense. It is a multi-point display and communication gateway for the Digiware system: it collects the measurements from the distributed measurement modules on the Digiware bus and presents them, while simultaneously acting as the gateway to the outside world. Its communication list is unusually broad — Modbus RTU and Modbus TCP, SNMP v1/v2/v3 with traps, BACnet IP, HTTP(S), FTP(S) and SMTP(S) — which is what allows energy data to be pushed into a BMS, an SNMP manager or a reporting server without additional hardware. It is a 96 × 96 mm panel instrument with a 24 VDC supply, IP65 front protection and a −10 to 70 °C operating range.
Where a project needs hundreds of measurement points, this architecture changes the CT conversation entirely: instead of running a 5 A secondary loop from every feeder back to a central panel, the measurement moves into the busbar compartment and only the bus and a low-voltage supply travel to the display. The full Socomec measurement and monitoring range is catalogued under Current Measurement & Monitoring, and the wider portfolio under Socomec.
8. Six CT Mistakes That Destroy Accuracy
- Oversized primary. A 1200/5 CT on a 60 A feeder operates at 5 % of range, outside the specified accuracy. Choose the smallest standard primary that covers the maximum load.
- Burden exceeded. Long, thin 5 A secondary loops routinely consume more VA than the CT is rated to drive. Calculate before you order the cable.
- Class mismatch. A class 1 CT behind a class 0.2S billing meter. The installation is only as accurate as its weakest element.
- Window too small for the primary. Check the through-feed opening against the actual cable or busbar, not the cable's nominal cross-section.
- Reversed polarity. Swapping P1/P2 or S1/S2 inverts the measured power direction, which looks like a meter fault but is a wiring fault. Keep the star point consistent across all three phases.
- Open secondary. An energised CT with its secondary open-circuited produces a dangerous voltage across the terminals. Always short-circuit the secondary — at the CT, with a dedicated shorting link — before you disconnect the meter.
Frequently Asked Questions
What does accuracy class 0.5 mean on a current transformer?
Class 0.5 defines a maximum composite error of 0.5 % at rated current, with defined limits at lower currents and at rated burden. It is the usual choice for sub-metering and internal energy management. Billing applications normally require class 0.2S, which extends the specified range down to about 1 % of rated current rather than 5 %.
How do I calculate the VA burden of a CT circuit?
Add the meter's current-input burden, the contact resistance and the cable burden. The cable term is S = I² × R with R = 2 × L × ρ / A, using ρ = 0.0175 Ω·mm²/m for copper. For a 10 m run of 2.5 mm² cable the result is 3.5 VA at 5 A secondary but only 0.14 VA at 1 A secondary.
When do I need a class 0.2S rather than a class 0.5 CT?
Whenever the measurement has a commercial or contractual consequence: tenant billing, energy performance contracts, tariff boundaries, or any figure that will be audited. For purely operational data — a display an operator glances at — class 1 is usually sufficient and costs less.
What is FS 5 on a Socomec current transformer?
FS is the instrument security factor, also called the overcurrent limiting factor. FS 5 means the composite error remains bounded up to five times the rated primary current, which limits what a fault current can present to the connected meter. A lower FS value is the more protective choice.
Can I connect a 5 A CT to a meter with a 1 A input?
No. The meter will read five times the true current and the input circuit may be damaged. Secondary rating is not an adjustment; it is a property of the CT and of the instrument that has to be matched. Low-level sensors with digital outputs avoid the issue altogether because their output is defined by the protocol, not by a 5 A or 1 A convention.
Related Products
Current transformers, low-level sensors and meters referenced in this guide:
- Through-feed CTs: TCA21 200/5 cl.0,5 · TCA21 300/5 cl.0,5 · TCB44-63 200/5 cl.0,5 · TCB85-100 750/5 cl.0,5 · TCB26-40 750/5 cl.1
- Billing class: T2CB32-40 500/5 cl.0,2s 5VA · T2RB115 125/5 cl.0,2s 7,5VA (wound primary)
- Three-phase block: TCB318-20 200/5 cl.1
- Busbar / high current: TBA103 1200/5 cl.0,5 10VA
- Low-level sensors and cable: TE-35 63-250A · CABLE RJ12 4x2M
- Meters and gateways: DIRIS A-41 · DIRIS A-100 333MV · DIRIS Digiware D-70
- Browse the category: Current Measurement & Monitoring · all Socomec products
Buying CTs and Meters From Glorient
Glorient supplies Socomec current transformers, low-level sensors and DIRIS metering systems with competitive pricing and worldwide delivery. Send us the maximum load current, the cable or busbar dimensions, the accuracy class you need and the meter model you are pairing with, and our engineers will confirm the correct reference within 24 hours. Start from the quotation form, or see the companion guide to Socomec fuse and circuit protection selection if your panel also needs protection devices.
Measurement accuracy is decided in the specification, not in the commissioning. Choose the ratio, class, burden and format correctly and the CT becomes invisible — which is exactly what a good measurement device should be. Our companion Socomec power solutions guide covers the rest of the range, from transfer switches to UPS and DC switchgear.