Some developing ideas on Seafloor cabling and copper/fibreoptic split points where this talks builds and borrows from the work of VENUS, MARS, NEPTUNE and NEMO, particularly that of F. Ameli, M. Musumeci, M. Sedita G. Hallewell Centre de Physique des Particules de Marseille KM3NeT Work Packages Meeting, Paris, Feb 23-24, 2009

Background:  Electro –optic wet mateable connectors for intermediate (~500V) voltages (Ocean Design Hybrid Mark 2 type) have been the most unreliable component in ANTARES  Ocean Design is sole source of electro-optic WM connectors: try to replace this technology with something cheaper, with multiple sources (competition)  Alcatel is having trouble with a 10KV electric-only WM connector for its deep sea nodes (NEPTUNE, NEMO) Try to develop a strategy to use copper backbone from OMs, down vertical drops and across the seafloor to secondary junction boxes

An early NEMO (2003?) geometry idea for redundant ring and star cabling

OTHER IDEAS - VENUS: Generalized Observatory Configuration: hook ring & spurs to science packages Platform R Backbone cable Junction Box Node BU Shore Terminal R R Repeaters Science Platform R Shore Terminal R Science Platform Node BU R Junction Box R R Junction Box BU Here is a generalized configuration for an underwater observatory consisting of a single backbone cable loop between two shore terminals. The terminals provide power to energize the cable, terminate the communications links to/from the underwater nodes, and provide a connection to land based networks. The terminals may be located at a single site or some distance apart to provide geographic diversity. Network nodes are connected to the backbone cable by means of underwater branching units (BUs) and spur cables, which may be from 5 to 25 km in length. The BUs provide power switching capability and routing of optical data channels to the nodes. The nodes distribute communications and power to a series of science platforms and secondary nodes via extension cables. All cables into a node have connectors to allow nodes to be removed for maintenance. Depending on the network design, optical amplifiers in underwater housings (repeaters) may be used to boost the communications signals along the backbone cable. Both repeatered and “unrepeatered” designs will be considered here. The dotted blue line indicates the boundary of the system being considered. Because repair times for shore station equipment are short, the contribution to unavailability is negligible. Junction Box R BU Spur cable Branching Units Node Node Junction Box Junction Box Science Platform Junction Box

OTHER IDEAS - NEPTUNE: Very long deep water network

A Junction Box approach could be based on VENUS/NEPTUNE node technology Penetrator MEOC cable entry into node (see UJ bend limiter)

Example of a completed UNIVERSAL JOINT cable repair box with internal cable anchorage and external bend limiters

UVic VENUS October 2007

Node internal cylinder structure Fibre splices CTA MVC Interconnections within Alcatel Node: HV and optics separated: 10KV W.M. connector still a problem HV conn CTA MVC Cable termination assembly uses UJ (Universal Joint) penetrator etc.

Node internal cylinder structure Pre-Qualified (i.e. warrantied) MVC (Medium voltage converter: 10kV  400V uses C6F14 fluorinert as cooling medium (electronics completely immersed) Fibre splices CTA MVC HV conn CTA MVC It would make sense to consider adding an MVC cylinder ‘ante-chamber’ in which the HV connector could be dry-mate, the cylinder deck-filled with C6F14 –an excellent dielectric

UNDER DISCUSSION ONLY Node internal cylinder structure Pre-Qualified (i.e. warrantied) MVC (Medium voltage converter: 10kV  400V uses C6F14 fluorinert as cooling medium (electronics completely immersed) Fibre splices CTA MVC HV conn CTA MVC Threaded rod extensions might allow fitting of an Antechamber (cylinder+ penetrator flange) UNDER DISCUSSION ONLY

Node insertion into TRF, MARS

Example: standard single conductor and multi-fibre branching unit Universal Joint Technology (penetrators) might allow us to build a scalable sea floor JB-linked infrastructure without intermediate WM HV connectors Example: standard single conductor and multi-fibre branching unit

MORE ON BRANCHING UNITS BU’s can be very sophisticated with F/O commanded power switching A B.C, AB+C, BA.C, BA+B, CB.A, CB+A Also colour dependent fibre drops All this can get very expensive ~ 1M€ per B.U. But simpler passive configurations are possible ~ 300k€ / B.U.

Possible (Cu) star & (fibre) hook ring geometry Fibres + high voltage circulate between JBs (sea electrodes (HV) near each JB) One or two MEOCS Ring of Neptune/NEMO-like JBs with internal fibreoptic drops, DWDM and multi-Cu to fibre conversion. 400V DC multiple generation from 7-10kV DC. Circulating electro-optic MEOC between JBs linked by penetrators

MEOC Penetrator connection between between JBs Ring of external identical Neptune-like JBs with internal fibreoptic drops, DWDM & 370V DC multiple generation to power lines in star config. MEOC Penetrator connection between between JBs (Voltage converter cylinders with two antechambers) MEOCS to shore

Maybe simpler to deploy with multiple BU’s MEOC to shore All spurs have a length ~ 2x water depth for individual JB retrieval MEOC to shore

Better optical budget + one less BJ but more difficult to deploy… MEOCs to shore

Consideration of copper radial (star) connections Consideration of copper radial (star) connections to circumferential junction boxes  (Ocean Design Hybrid Mark 2 type) have been the most unreliable component in ANTARES  Ocean Design is sole source of electro-optic WM connectors: Tests with copper backbone cable are quite encouraging…

From Fabrizio Ameli’s talk at VLVnT08…

One of several electric only WM connectors: GISMA Series 80

Tests with copper backbone cable Tests with copper backbone cable are quite encouraging…  Can several hundred Mb/s be transmitted over ~ 500m across seafloor (DU divided into several descending copper backbones, each for a ‘sector’?  Ocean Design is sole source of electro-optic WM connectors: but there are many mfrs (~10?) of WM electric-only connectors – even subconn connectors (a la ANATRES OMs) could be a possibility Electrical pulse reforming at DU anchor should be possible, for retransmission a further ~ 500m: Questions: how many conductors to transmit data to JBs from each DU anchor?

Summary. Some ideas on independently-liftable ring of identical Summary Some ideas on independently-liftable ring of identical junction boxes linked to each other in a redundant fibre + power ring using standard deep sea electro-optic cable with penetrator technology such as used in Universal Joint ing Possible solution to Alcatel HV intranode HV cable connection problem using MVC cylinder extension ‘ante-chamber’ & fluorinert ® dielectric fluid filling Possibilities for copper transmission across sea floor with base-of-line reformed electrical pulses. Multi 100 Mb/s probably within reach of TP cables and connectors