Ieee std 1159-1995, ieee recommended Practice for Monitoring Electric Power Quality

A.2 Calibration issues
A.2.1 Drift rate 
The time span of a speciÞcation indicates the length of time an instrument can be expected to remain within
the speciÞed limits. One to two years are common time spans for portable instruments. Uncertainties will
increase for longer time spans due to a drift rate. If a drift rate speciÞcation is included, a buyer can calculate
the uncertainty for the time span required.
A.2.2 Temperature coefÞcient
This represents the amount that uncertainty increases with temperature variation from a speciÞed tempera-
ture spread. A typical temperature spread is 23 ¡C (73¡F) ±5 ¡C. A wide temperature spread and a small tem-
perature coefÞcient permit on-site calibration since the instruments are outside the protected environment of
the calibration laboratory.
A.2.3 Where to calibrate
Calibrating a working instrument in its real environment is inherently more accurate because local affects
are taken in account, and on-site calibration is safer since the instrument is not subject to damage in transit.
Also, the instrument is out of commission a shorter time. Contrary argument says, Òshould Ôlocal effectsÕ be
included or purposely deleted from calibration?Ó If it affects and creates a more accurate measurement at
one site, does not that naturally mean it may produce less accurate data at another site?
A.2.4 Calibration intervals
A broad guideline is provided by military speciÞcation MIL-L-45662B [B28]. Test equipment and standards
should be calibrated at intervals established on the basis of stability, purpose, degree of usage, precision,
accuracy and skills of personnel utilizing the equipment.
Figure A.1ÑInstrument accuracy is the combination of (a) random errors and
 (b) systematic component errors

---

IEEE
Std 1159-1995
IEEE RECOMMENDED PRACTICE FOR
62
Some instruments offer built-in calibration. It is important to know exactly what a built-in calibrator is test-
ing. Internal self-test can be limited by its own processing procedure, so it is important to determine what
type it uses. Internal testing can be limited by providing only a couple of points on a spectrum to try to vali-
date performance whereas a laboratory calibrator would generate a whole spectrum of points; 80 points
might be an average. Internal test may only test one section at a time and at low signal levels. The user must
determine if that method is adequate. Built-in self-calibration and self-test may not be able to generate high
voltages necessary to test the front end circuitry of the instrument. A calibration laboratory would normally
use a standard that is four times more accurate than the instrument being calibrated to validate accuracy. This
standard is difÞcult to reproduce with a built-in device.