Accurate CT Calibration for the Model 1133A

Overview test, ratios of (1000*M/N):5, (800*M/N):5, and so on

can be provided. This allows a wide range of test CT
The Model 1133A Power Sentinel™ offers accuracy ratios to be calibrated with a single accurate refer-
in power measurement of 0.025%. Existing CTs ence artifact.
provide accuracy of a few tenths of one percent.
How then can the Model 1133A actually provide
accuracy of 0.025% in the field? Excitation Current Source

The answer is calibration of the CTs. By comparing To provide the high currents required to perform the
each CT to an accurate reference CT, traceable to calibration, we also plan to provide an excitation
national standards (for example, NIST in the USA), current source. This will work on the principle of a
the errors of the test CT can be determined at “CT in reverse,” that is, a ‘donut’ having a multi-turn
several currents and the correction factors entered primary excited at a reasonable current level (5 to
into a table in the Model 1133A. The Model 1133A 10 amps maximum). This device will allow a heavy-
uses these correction factors, interpolating between gauge winding to be placed through its center,
them based on the measured current, to correct exciting it at the high currents required (up to 2000
measured power (and current) to the actual values, ampere-turns).
in real time, within the instrument’s stated accuracy.
While it is possible, and perhaps even desirable on
some accounts, to provide a regulated, solid-state
Calibration Process source to drive the excitation coil, this would be very
expensive and heavy considering the powers in-
To calibrate the CTs requires three things: first, a volved (up to 1000 VA). Due to the accurate
calibrated reference CT; second, a source of high ratiometric measurement technique to be proposed
current; and third, some means of comparison. Each in the next section, it is not necessary to have the
of these will be addressed below. excitation current be particularly accurate or even
stable. Therefore, we can use a much simpler,
‘brute-force’ design based on a multi-tap autotrans-
Calibrated Reference CT former and Variac to set the current. The design
contemplated will offer settings for input voltage
range (100, 120, 200, 220 and 240 Vrms) and output
Arbiter Systems plans to develop an accurate, multi-
range (100, 50, 20, 10 and 5% of full-scale), and
ratio reference CT suitable for these calibrations,
continuous adjustment from zero to the selected
and suitable for calibration by national standards
output range. The device would be powered from the
laboratories at an accuracy which exceeds 0.01%
customer’s ac main power.
(the present limit of traceability at NIST). This CT
will have ratios of 1000:5, 800:5, 750:5, and 600:5. It
will operate accurately at currents from 0 to 200% of
Comparison Technique
rating. Using a two-stage, amplifier-aided design,
this CT should be fundamentally capable of accuracy
better than 10 ppm (0.001%). The most direct method of comparison involves
making a measurement using the reference CT, then
By using a multi-turn primary through the center of using the unknown, and then calculating the ratio
the reference CT, current ratios of (1000/N):5, (800/ correction factor and phase angle from the data
N):5, (750/N):5, and (600/N):5 can be provided as (figure 1). This has the drawbacks that the current
well (with N equal to the number of turns). By using source and measurement device must be substan-
a multi-turn primary on the CT under test, ratios of tially more accurate and stable than the desired
(1000*M):5, (800*M):5, and so on can also be measurement result, typically by a factor of 4 or 5. If
provided. And finally, by using both N turns through this were possible, it would greatly increase the cost
the reference CT and M turns through the CT under and size of the equipment.

Arbiter Systems, Inc. · 1324 Vendels Circle, Ste 121 · Paso Robles, CA 93446
Tel +1 805 237 3831 · Fax +1 805 238 5717 · E-mail sales@arbiter.com · http://www.arbiter.com 1
Accurate CT Calibration for the Model 1133A

REFERENCE CT

Step 1 A

Ia = rIef
Ib = (Itest - rIef)

TEST CT

Figure 2

Step 2

Figure 1

A better technique is called nulling (figure 2). In this standards-lab performance can readily be achieved
method, the outputs of the two CTs, reference and in the field using rugged, portable equipment de-
test, are subtracted electrically by using superposi- signed for field use, operated by service technicians
tion (Kirchhoff’s law), and the difference (error) with little or no training in metrology techniques.
measured directly. This subtraction is exact, and
introduces no error. Using a second channel of the
same measuring instrument, the actual secondary Accessories
current can also be measured, and the ratio correc-
tion factor and phase angle are then determined This setup will be provided with the necessary
mathematically. Best of all, if the measuring instru- accessories to perform the required calibrations. The
ment can make the two measurements simulta- cable carrying the high primary current will be AWG
neously, then effects of source variations cancel 4/0 welding cable, capable of up to 500 amps
out. Furthermore, since the quantity being measured current. For applications where the cable can be run
is a small error, equal to at most about 1% of the through the center of the test CT, a single long
secondary current, the accuracy required of the length (about 15m or 50 ft.) of cable will be provided,
measuring instrument can be relaxed as well. allowing for multi-turn setups. To perform a calibra-
tion of, for example, a 1000:5 CT at currents up to
The ideal instrument to make this comparison is the 2000 amperes, four turns are required through each
Arbiter Systems Model 931A Power System Ana- of the excitation CT, the test CT, and the reference
lyzer™. This instrument samples both of its selected CT.
input channels simultaneously, meeting the require-
ment stated above. Best of all, measuring the error For applications where the current must be provided
current with 0.05% accuracy, the Model 931A is to a bar running through the test CT, four shorter
capable of making this comparison at a level of 5 (3m or 10 ft.) cables will also be provided. The
ppm of the secondary current (0.05% of 1%). This lugged ends of all of the cables are connected to the

Arbiter Systems, Inc. · 1324 Vendels Circle, Ste 121 · Paso Robles, CA 93446
2 Tel +1 805 237 3831 · Fax +1 805 238 5717 · E-mail sales@arbiter.com · http://www.arbiter.com
 Accurate CT Calibration for the Model 1133A

primary of the test CT. This would form a single-turn

loop, with each cable carrying 500 of the 2000
amperes of current.

Software will be provided to automate the data

gathering and reduction process, using a serial
connection to the Model 931A and prompting the
operator to perform the proper hookup and settings
of the other equipment. The data could later be
printed out, and it can be stored in a data file suit-
able for use when configuring the Model 1133A,
eliminating a manual step where errors could enter
into the process.

Hardware required to connect the lugged ends of the

cables together, completing the high-current loop,
and tools to tighten the hardware to specification,
will also be a part of the kit.

Conclusion

Taking advantage of the full accuracy of the Model

1133A Power Sentinel™ requires calibration of the
user’s CTs. This paper presents a method to per-
form these calibrations. Using two new products to
be developed by Arbiter Systems, a high-accuracy
reference CT and an excitation current source, and
an Arbiter Model 931A Power System Analyzer™,
calibrations will be done at primary currents up to
2000 amperes (higher under some conditions) with a
transfer accuracy of 5 ppm and an overall traceable
accuracy of 0.01% or better.

Arbiter Systems, Inc. · 1324 Vendels Circle, Ste 121 · Paso Robles, CA 93446
Tel +1 805 237 3831 · Fax +1 805 238 5717 · E-mail sales@arbiter.com · http://www.arbiter.com 3