Bowin 



prograxfis. The manifestations of this failure were excessive 

 stepping of the automatic spring tension controller for the gravity 

 meter, water depths were changed to zero, and the digital display 

 showed unbelievable numbers. Diagnostic programs failed to 

 reveal any difficulties. Although adjustments of a bias voltage in 

 the analog to digital converter helped relieve the trouble, the 

 failing printed circuit card was not identified until the ship had 

 returned to Woods Hole and an IBM engineer from San Jose, 

 California worked on the system. The problem appears to have 

 been caused by components that failed only when loaded by the 

 computer doing multiple operations. The diagnostic tests did not 

 reveal the trouble because they carry out only one test at a time. 



We now believe that most serious difficulties during the first 

 three and a half years of operation were the result of radio 

 frequency interference. Now, of course, we can only surmise 

 that many of the earlier difficulties were caused by RFI. However, 

 the capriciousness and m.ysteriousness of the earlier troubles have 

 a strong similarity to the later troubles which were positively 

 identified to be caused by RFI and leads us to the strong supposition 

 that this was a major source of trouble earlier. The manifestation 

 of the interference was the alteration of values in the computer 

 program. For example, one of the first indications of such 

 trouble was erroneous values of ship's speed and water depth. In 

 both these cases the trouble was identified as occurring either 

 during the analog to digital conversion or immediately following. 

 Somehow the interference was coming into the computer system. 

 Many attempts were made to determine the source of these troubles; 

 including the removal of antenna lead-in lines coming directly into 

 the main laboratory and near the computer facility, installation of 

 radio frequency suppression filters in the ship's radio power lines 

 and in the 1710 system, and improvement of the system, ground. 

 These attempts usually temporarily alleviated the problem, but 

 were not a permanent solution. In spite of these efforts, RFI 

 problems continued which eventually developed into a mutual 

 interference between the ship's radio and the computer system in 

 which transmission would disturb values within the computer 

 system. Also the units of the computer, in particular the type- 

 writers and digital plotter, could be heard on many frequencies of 

 the ship's radio receiver. When this interference became strong 

 on 500 Khz (international calling and distress frequency) the 

 problem demanded immediate correction. We tried the drastic 

 action of removing all the personal radio receiver lines installed 

 haphazardly throughout the ship by the scientists, crew, and 

 officers, including the Captain's and the cook's. Older abandoned 



258 



