The stress level at which failure occurs at temperature, T, 

 and after a time, t, is reported to be a function of K. Thus by- 

 testing at high temperature, we should be able to observe fail- 

 ures in reasonable testing periods. The data so obtained should 

 then permit determination of time-to-failure at normal temp- 

 erature by using the factor, K, It is also approximately true 

 that failure stress is a linear function of K. Hence, the logarithm 

 of life may be a linear function of stress for the time and stress 

 levels of interest. Life data have therefore also been plotted 

 vs. stress on a senai-log graph. Life has necessarily been 

 estimated by extrapolation to operating pressures from test 

 data so plotted. These plots are given in Figs. 4-7 through 

 4-10. There is much scatter in the data due to the lack of 

 uniformity within any batch. Note that we estimate the following 

 life expectancies at 11 C: 



XX 1000 PSI, 10^<life<10^ hrs., avg. life-^io'^ hrs. 



Marola, 1500 PSI, 10-^< life <10^ hrs. , avg. life'-'lO^ hrs. 



? 10 7 



Marola, 1000 PSI, 10 < life < 10 hrs., avg. life/*/10 hrs. 



Figs. 4-11 and 4-12 contain photographs of collapsed floats. 



4. 2 Sensor Calibration and Environment Testing 



It is the intention of this section to describe simply the 

 tests performed prior to installation, it is not felt that 

 detailed information is of general interest. 



4. 2. 1 Thermistors 



a. Standard R-T curve: property of semiconductor 

 material checked for each lot at factory. 



b. Individual thermistor bias from standard curve 

 determined at MIT/IL in ice water bath, August 

 and April 1968. No change of bias resistance was 

 detected. 



c. Time constant as installed in sensor housing: 

 continuous resistance record after immersion in 

 ice water bath at MIT/IL, September 1967. 



d. Gross check in pressure environment: compar- 

 ison with crude indicator during SRI pressure 

 tank test up to 1300 PSI, 6, 7 October 1967; 

 reasonableness check during SRI pressure 

 tank test up to 1200 PSI, 29, 30 July 1968; 

 proper operation during deployment to 600 

 meters depth, 24 November 1967. 



223 



