Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
ACCUMCY OF MTA As stated earlier, methods and equipment with proven r e l i a b i l i t y in the laboratory were used in the f i e l d test. The equipment was calibrated both before and after the test. Experience in using this equipment was depended upon to detect possible sources of error during the test. Representative deflection data are shown in Fig. 15. This figure shows deflections measured in Test I at Locations 2h and 26.(â¢^) Deflections were measured by a different level for Scale 2k than for ScaJ.e 26. Both levels had a sight distance of about 70 f t , one of the longer sight distances. I t i s seen that there i s greater scatter about the curve drawn for data measured at Scale 26 than for data measured at Scale 2k. Such scatter i s attributed to less than ideal lighting and lack of experience of personnel operating the levels. I t i s f e l t that in general accuracy of readings were within tO.Ol in. Strains measured with the manually operated strain indicator were re- corded to the nearest 5 millionths. G^ypical strain data from Test I are shown in Fig. 16. Individual strain readings are within a narrow band of about l3 millionths along the load-strain curves. Difficvilty was encountered in Tests I and I I in measuring load using load c e l l s . The soirrce of the problem was later traced to a faulty strain indicator. I t was not possible to obtain a replacement for the defective indicator unt i l Test I I I , indicating the obvious desirability of a stand-by instrument. Average loads recorded in Test I I I using load ce31s were generally within 1 2 percent of loads determined by the pro- duct of average ram area and hydraulic pressure. A comparison of loads determined by these two methods i s shown in Fig. 17. 2- 11
Hydraulic pressure was measured using Bourdon type d i a l gages^ Bourdon type strip chart recorders, and a pressure c e l l . A l l units were c a l i - brated before and after the tests. One strip chart recorder was dam- aged i n shipment after the tests and i t s performance was different from that before the tests. Pressure c e l l readout was made using the same strain indicator used for load c e l l s . Consequently, pressures recorded for Tests 1 and I I were in error because of the faulty strain indicator. In Test I I I , the pressure c e l l indicated about 3 percent less pressure than that shown by the d i a l gage as seen in Fig. l8. In Test I , both strip chart recorders had a zero shift not discernible from the pressure record. Pressures by the recorders were 15O psi and 50 psi less than pressure measured by the dial gages. Pressvires from the strip charts compared satisfactorily with those recorded from d i a l gages in Tests I I and I I I . Recorder versus di a l gage pressures for Test I I I are shown in Fig. 19. Recorder pressures in Tests I I and I I I were within about t 3 percent of dial gage pressiire with few exceptions. In the f i n a l analysis of test data,^-^^ loads were taken as the product of average effective ram area and pressure measiired by d i a l gages. 2-12
