The grazing incidence (GI) telescope will be built up from about 7,000 pairs of primary (P) and secondary (S) reflector elements.These pairs are azimuthal segments of a circular shell. Reflector pairs from many shells are nested within modules (right) that provide mechanical support, alignment, and thermal and electrical services. The telescope concept has an 8.3 m diameter inner section filled with 125 shells, and four 60 deg fold-out sections (one split) giving a partially filled 16 m diameter mirror (below), with 250 shells. Each section contains several modules.

To achieve 0.1" half power diameter (HPD) resolution at 1 keV, our preliminary error budget simply divides in quadrature and allocates 0.05" to each of Figure, Alignment within a module, Module-to-Module alignment, and Margin.

Our critical technical innovation is to develop adjustable mirrors using thin optics with thin film electro-active actuators deposited directly on the back surface. The surface figure can be locally corrected via surface-parallel strains without the need for a reaction structure. Such optics would be adjusted very infrequently (during an extensive initial on-orbit calibration,and then on ~year time-scales) to remove figure errors that could not be measured and controlled accurately enough on the ground and maintained through the launch environment. The actuators (red cross-hatch in the figure to the right) may be either piezo-electric material or electrostrictive. When the voltage across the actuators changes, they expand or contract imparting a strain on the mirror (shown in blue). The deposition of electrodes on the back surface would create a pattern of cells (figure below) which would individually control the local slope errors.

On-orbit we would use a high throughput X-ray imager which could be positioned at specific locations in several different planes forward of the telescope focal plane. The positions are chosen so that the converging rays from each discrete shell are resolved as portions of separate rings. An ideal mirror would produce exactly uniformly filled rings (right, below). By measuring the profile across each ring in narrow azimuth bins, we can diagnose what figure corrections are needed along that particular axial strip. For example, in the figure to the left below, we simulate the profile for an axial element which is perturbed by a sum of Legendre polynomials with amplitudes up to 0.1 micrometer. Neglecting scattering, such a perturbation would produce 0.11" HPD in the focal plane.