The problem of controlled thermonuclear synthesis of rarefied plasma, confining by magnetic field, exists about 40 years. At the present time the problem solution follows in the traditional way of transition to larger and more expensive facilities. The way is near to saturation and its perspectives are vague. For the reasons takes place sharp decreasing of financing. Taking into account for the considerations it is expedient to analyse a nontraditional way, when between the plasma filament and first wall is formed a vacuum gap. Physical aspects of the way are considered in the paper. 28 refs.; 8 figs.