b. When you calibrate you compare a standard of a known accuracy

against an unknown, Under ideal conditions, the standard should be 10 times

more accurate than the unit under test. When you cross-check you only check

the probability of standards remaining within their stated tolerance. One

cannot assure that a standard is in tolerance by cross-checking; this is

done only by a proper calibration. However, you can compute the probability

error between two standards whose tolerances are basically the same. This

is determined by using the formula: Square root of the sums of the

individual accuracies squared, as shown below:

%Ep = [(%E1)2 ----(%En)2]1/2

where:

%Ep

=

Percent

of

probability error

%E1

=

Percent

of

error of standard 1

%E2

=

Percent

of

error of standard 2

%En

=

Percent

of

error of all other standards to the nth standard

Note: The probability of the standard being within tolerance is greatly

increased as more readings and standards are cross-checked on the same

parameters. In post cases %Ep is based on intercomparison of two standards.

3.

REASONS FOR CROSS-CHECKS.

a. General.

The theory of cross-checking is a functional comparison

between instruments to verify that no significant departure from assigned

tolerance has occurred.

The cross-check is a method to detect any change

(drift) in transfer standards that may establish trends and to assist in

preventing out-of-tolerance standards from being used in the performance of

calibration.

Also, the cross-check increases confidence of operator

personnel in their equipment. Secondary transfer standards are more subject

to out-of-tolerance condition because they are mobile and used in

uncontrolled

environments,

causing

premature

failure

and

aging

of

components.

b. Aging of components. As a piece of equipment is used, it is subject

to current surges as it is turned on and off, or due to input line voltage

changes. After a piece of equipment has been operating continuously for 8

hours or more, which is not uncommon for a calibration standard, the heat

will affect some components by some small amount.

When this happens day

after day for months, the component's value will change, causing the value

of the standard to change.

By making periodic cross-checks of our

standards, we may detect it drifting out of tolerance.