Because of its size, DUMAND II was subdivided into two phases. Phase I consists of the construction, deployment and installation of three strings of 24 optical detector modules, each with the associated auxiliary and support systems (figure 2) and of the junction box at the DUMAND site. Phase I was to include the laying of the 36 km-long twelve fiber electro-optical cable from the junction box to shore (one optical fiber for each string, one for the video cameras and to control the sonar responders, and two spares), connecting the three strings to the junction box with a deep sea robot or a manned submarine, and taking this configuration, the Triad, into operation. Phase II comprises the completion of DUMAND II.
Figure 2: Details of a sensor string, divided into three sections, shown side by side, with all the auxiliary and support systems. Full length 415 m.
By fall of 1993 the hardware for Phase I approached completion, and deployment of the Triad was fixed for late November. However, production and testing of strings was behind schedule, and further some ship resources including the deep sea robot and the deep dive submarine were diverted by U.S. government order due to an emergency. Consequently, the unavailability of this equipment made it impossible to connect the strings to the junction box at the sea floor.
In order to deploy at least part of the Triad for test purposes, one single string was attached and connected directly to the junction box prior to deployment and, as shown in figure 3, deployed together with the junction box and shore cable in a single complex oceanographic operation in December 1993.
Figure 3: Deployment operation of December 1993. The upper end of the sensor string is deployed first, followed by the junction box (JB) and the shore cable. After completion of deployment and cable laying the sacrificial anchor (SA) at the end of the string is released and the string rises into vertical position.
The deployment was very difficult but successful. The complete subsystem was brought into operation and its functionality fully confirmed. Unfortunately, after about ten hours of operation problems developed that suggested that a micro leak had developed in the controller housing. Never during the numerous tests prior to deployment did such a problem occur. Nevertheless the few hours of operation of the single string yielded extremely valuable scientific data. These delivered proof that the integrated subsystem is conceptually best suited to fulfill its task. Parts of the results are already published [5].
Six weeks later, in January 1994, we recovered the string, which had been shut-off in the mean time. Recovery was conducted from a ship by sending a sonar command which activated the string release mechanism, disconnecting the string from the junction box. One hour later the string floated to the surface and was taken on board the ship. A detailed inspection later on in the laboratory revealed that an electrical penetrator at the string controller housing developed a tiny leak that led eventually to a short circuit in the power handling section of the controller. The auxiliary and support systems, such as sonar and video, were not affected by the penetrator failure since they have their autonomous control and data transmission systems. The damaged parts of string one are presently being repaired. In addition, string one, together with three additional strings, are subject to minor upgrading based upon the very useful experience gained in these operations.