CT Selection

Overview

This article discusses the considerations when selecting a current transformer. For most applications, the easiest approach is to:

Select a split-core CT for ease of installation.
Select the CT with the smallest opening large enough to fit around the conductor you are measuring.
Select the CT model with the lowest rated amps that are equal to or greater than the rating of the service, breaker, or fuse circuit being monitored.

For more complex applications, the following discussion provides additional details. The main issues are:

The range of currents to be measured.
The diameter of the conductor being monitored.
Is an opening CT desired to simplify installation?
Accuracy requirements.
The crest factor of the load.

Range of Currents

CCS generally specifies CT accuracy down to 10% of rated current for standard CTs and down to 1% of rated current for revenue grade CTs. The upper limit is generally 100% or 120% of rated current. This is the range over which CTs are the most accurate, so depending on your accuracy requirements and the type of CTs being used, you should choose CTs so that the nominal currents fall within this range.

Most CTs are still reasonably accurate at lower currents. For example, the ACTL-0750 models are very accurate down to 0.2% of rated current, and other conventional CTs are reasonably accurate down to 2-5% of rated current.

Many (but not all) CTs are also accurate for over-current conditions; however, the WattNode input circuitry clips at around 125% of rated current for sine wave currents and at even lower currents if the current has a high crest factor, so we always recommend selecting a CT with a rated current higher than you ever expect from the load (with the possible exception of very brief turn-on surges).

For example, suppose you are monitoring a 100A circuit with a compressor and fan. The turn-on surge reaches 85A, the normal load when the compressor is running is 65A, and when just the fan is running, 7A. The following table shows some options:

CT Model	Rated Amps	10% Point	100% Point	Comments
ACTL-0750-100	100A	10A	100A	This CT has good accuracy specifications from 1% to 120% of rated current, so will be accurate at all operating currents.
CTT-0750-100	100A	10A	100A	Because this is a solid-core CT, it will be quite accurate even below 10% of rated current, so it should work well over the entire range.

As another example, suppose you are monitoring a compressor that typically draws 125A. You could use a 150A CT, but we’d recommend a 200A CT to allow more headroom for the turn-on surge and overload. Since the compressor will rarely draw much less than 100A when it’s running, that will still be 50% of the CT rated current and the accuracy of the CT will be very good. The compressor current would have to drop below 20A before there would be accuracy concerns.

Conductor Diameter

See Conductor AWG Size vs. CT Opening Size for a table of wire gauges, current ratings, diameters and recommended CT opening sizes.

To summarize, you will get the best accuracy and generally the easiest installation if you pick the smallest CT that will fit around the conductor you are monitoring. Larger CTs will be harder to fit in the panel, harder to keep centered around the wire, and harder to keep from touching other conductors and CTs (the accuracy is best if CTs are not touching each other and are not touching other conductors).

Opening vs. Solid-Core CTs

Solid-core CTs require that the conductor be disconnected to pass it through the opening in the CT. This is not difficult when the wire gauge is small but becomes impractical with larger wire gauges and multiple parallel conductors. Installing solid-core CTs requires turning off the power to the circuit and disconnecting the conductor from the breaker or fuse, so the installation is more difficult. However, solid-core CTs are generally more accurate and less expensive then conventional split-core CTs.

Split-core CTs can be opened for installation around a conductor. For most retrofit installations, this is much faster and easier. Large electrical services with multiple parallel feed conductors traditional use large split-core CTs such as our CTBL Series CTs, installed around all the conductors or around the bus bars . Custom current ranges and sizes are are available in widths from 1.5″ to 12.0″ and leg lengths from 2.0″ to 12.0″.

Even custom sized CTs can be difficult to fit is some panels. The easiest technique to installing CTs in large panels with parallel feed conductors is to install one CT on each the individual phase conductor in a set. The CTs are installed where the conduit sets enter the panel. The black and white output leads from each CT on the same phase conductor are wired together in parallel at the meter’s CT input terminals. The meter is configured for the sum of the individual CT’s rated currents. Note that all CTs must be the same part number. Additional information can be found in the Multiple CTs – All Conductors section on the Measuring Parallel Conductors page.

Accuracy

For best accuracy, use our ACTL Series split-core current transformers or any of our other CTs available in revenue grade versions (class 0.6). For non-revenue grade CTs, use the standard ACTL models (class 1.0) or the CTT series of solid-core CTs, which are also class 1.0. When using other brands of CTs, check the phase angle specifications, because low phase angle errors (see Measurement Errors Due to CT Phase Shift) provide better accuracy for loads with low power factors.

As the measured current varies, the accuracy of the current transformer will vary. Traditional CTs are most accurate from 10% to 100% of rated current. Above 100% of rated current, the WattNode input circuitry may start to clip, resulting in rapidly degrading accuracy. Below 10% of rated current, the accuracy of traditional CTs degrades.

Crest Factor

See http://en.wikipedia.org/wiki/Crest_factor for an overview of crest factor.

The term “current crest factor” is used to describe the ratio of the peak current to the RMS current. Resistive loads like heaters and incandescent lights have nearly sinusoidal current waveforms with a crest factor near 1.4. Power factor corrected loads like PC power supplies typically have a crest factor of 1.4 to 1.5. Many common loads can have current crest factors ranging from 2.0 to 3.0, and higher values are possible.

The WattNode current transformer inputs will saturate and become inaccurate if the peak current is too high. This means you may want to be conservative in selecting the CT rated current. For example, if your load draws 10 amps RMS, but has a crest factor of 3.0, then the peak current is 30 amps. If you use a 15 amp CT, the WattNode will not be able to accurately measure the 30 amp peak current. Note: this is a limitation of the WattNode measurement circuitry, not the CT.

The following graph shows the maximum RMS current for accurate WattNode measurements as a function of the current waveform crest factor. The current is shown as a percentage of CT rated current. For example, if you have a 10 amp load with a crest factor of 2.0, the maximum CT current is approximately 85%. Eight-five percent of 15 amps is 12.75, which is higher than 10 amps, so your measurements should be accurate. On the other hand, if you have a 40 amp load with a crest factor of 4.0, the maximum CT current is 42%. Forty-two percent of a 100 amp CT is 42 amps, so you would need a 100 amp CT to accurately measure this 40 amp load.

You frequently won’t know the crest factor for your load. In this case, it’s generally safe to assume the crest factor will fall in the 1.4 to 2.5 range and select CTs with a rated current roughly 150% of the expected RMS current. So if you expect to be measuring currents up to 30 amps, select a 50 amp CT.