Since the start of operations in mid-1983, JET has undertakne two major machine upgradings. The first in 1986-87, when the plasma current capability was increased from 4.8MA to 7MA and X-point magnetic configurations were made possible up to 5MA. This upgrading was accompanied by progressively covering the vessel walls with graphite and/or beryllium tiles supported by beryllium deposition onto the whole suface of the vessel. The synergy between these interventions led JET to achieve maximum performance (Q~1), but only for a fraction of a second.

A second major upgrading was carried out in 1992-93 with the installation of a pumped asymmetric poloidal divertor, with plasma current capability up to 6MA. This allowed active impurity control and plasma exhaust and three configurations have been foreseen with a progressively more-closed structure. JET is now operating with the second of these configurations (Mark II).

Extensive experimental work with three main plasma scenarios (hot-ion mode, optimised shear and Elmy H-mode) in D-D have allowed definition in detail of the forthcoming D-T experimental campaigne DTE1 (June-August). Afterwards, the Mark II structure should be replaced with Mark IIGB structure, using remote handling techniques, already fully tested by installing ~20% of Mark II tile structure remotely. The 'gap control' of the plasma boundary and shape has been further improved by increasing the number of the plasma-vessel wall gaps to be controlled. Machine instrumentation has also been enhanced to better monitor and assess the effects of the interaction between plasma and machine structures during disruptions. On some occassions, significant radial displacements (< 7mm) of the whole vacuum vessel, were observed.

Extensive studies have been carried out in 1996-97 to prove that the toroidal magnetic field could be increased from 3.45T to 4T. The assessment includes detailed computer modelling of the coils and of the vacuum vessel to evaluate stresses and tests on slices cut from a faulty coil, replaced in 1989 after six years of operation due to an interturn short circuit. The tests show no deterioration of the brazed joint and of the fibre-epoxy bonding. The coils have already been commissioned up to 3.8T and up to a plasma current of 4.2MA. This is the set of parameters to be used during DTE1, aiming at increasing the neutron yield by 20% to 30%.