Chapter 7-PRINCIPLES OF MEASUREMENT 



An acceptance for limited use indicates that 

 an instrument which has failed certain tests or 

 which has not been tested against all acceptance 

 criteria is nevertheless suitable for certain 

 specified (limited) usage. In such a situation, 

 a limited use label (rather than a CALIBRATED 

 or QUALIFIED label) is placed on the instru- 

 ment to draw attention to the conditional ac- 

 ceptance. In addition to the label, a limited use 

 tag is attached to the instrument. This tag is 

 filled in by the servicing activity; it includes 

 a full description of the reservations or pre- 

 cautions which should be observed in using the 

 instrument. The label and the tag indicating that 

 the instrument is suitable only for limited use 

 must remain on the instrument until the next 

 calibration or qualification. 



The term rejected is used when an instrument 

 fails to meet the acceptance criteria during 

 calibration or qualification and when it cannot 

 be made to meet these criteria by incidental 

 repair. Under these conditions, a REJECTED 

 label is placed on the instrument and all other 

 servicing labels are removed. In addition to the 

 label, a REJECTED tag is attached to the in- 

 strument. The tag, which is filled in by the 

 servicing activity, gives the reason for rejec- 

 tion and such other information as may be re- 

 quired. 



Repair is defined as the repair and/or 

 replacement of malfunctioning parts of a meas- 

 uring instrument or standard to the degree 

 required to restore the instrument or standard 

 to an operating condition. 



Traceability is when the accuracy of a meas- 

 urement made by the fleet can be directly trace- 

 able through the echelons of calibration to 

 Reference Standards maintained by the National 

 Bureau of Standards (unbroken chain of properly 

 conducted calibrations.) 



INSTRUMENT ACCURACY 



As we have seen, all measurement is subject 

 to a certain amount of error. The international 

 and national standards have some error, even 

 though it is almost unbelievably small. The 

 error in secondary standards, while still ex- 

 tremely small, is somewhat greater than the 

 error in the primary standards. When we get 

 down to the actual measuring devices used even 

 for precision measurement, the error is larger 

 still. 



Although it may sound backwards, the accu- 

 racy of an instrument is expressed by giving the 

 amount of error of the instrument. For example. 



an instrument with an accuracy of 1 percent is 

 said to have an error of ± 1 percent. 



The error of an instrument is the difference 

 between the reading shown on the instrument and 

 the true value of the variable being measured. 

 Error may be expressed in scale units, in per- 

 cent of scale span, in percent of range, or in 

 percent of indicated value (iv). By agreement 

 among instrument manufacturers, error in in- 

 struments with uniform scales is most com- 

 monly expressed as a percentage of the full 

 scale length, regardless of where the measure- 

 ment is made on the scale. The exception to 

 this general rule is that the measurement is not 

 made at the extreme top or the extreme bottom 

 of the scale, since an instrument is almost sure 

 to be less accurate in these areas than in the 

 working range of the scale. 



Using the full scale length as a basis for 

 determining instrument error can lead to some 

 confusion. For example, consider several pres- 

 sure gages, each one of which has a guaranteed 

 accuracy of 1 percent. K the scale reads to 30 

 psi, the allowable error is± 0.3 psi. If the scale 

 reads to 100 psi, the allowable error is + 1 psi. 

 If the scale reads to 500 psi, the allowable 

 error is + 5 psi. If the scale reads to 1000, 

 the allowable error is ±10 psi. 



As far as accuracy ratios are concerned, 

 there are recommended low and high accuracy 

 ratios that should exist between the test and 

 measuring equipment, and the measuring system 

 or Standard, also between echelons of Standards. 



The lower limit ratio should be at least 4 to 

 1; a ratio below this limit is impracticable for 

 technical reasons. The upper limit ratio should 

 not be more than 10 to 1; if the ratio is higher, 

 equipment costs will become excessive. 



Calibration error is taken care of by line- 

 arity and range errors . Linearity error is when 

 the lowest and the highest indications are cor- 

 rect, and there is an error in between these 

 indications. Range error occurs when the lowest 

 indication is on and the highest indication is off, 

 above or below the true value. 



If an instrument does not give the same read- 

 ing when it comes from the top of the scale down 

 to the point of measurement as it does when it 

 goes from the bottom of the scale up to the point 

 of measurement, the error is called hysteresis. 

 Hysteresis occurs from a variety of factors that 

 cause loss of energy within the instrument; it 

 might occur because of friction or binding of 

 parts, fatigue of a spring, excessive play in 

 gears, or other mechanical difficulties. 



155 



