[en] After considering numerous focusing and monochromatizing schemes, we have settled on an optical geometry which we consider near ideal. The main optical element is a conical crystal which focuses 15 mrad of horizontal divergence from 2 to 20 keV, and which focuses 12 mrad of horizontal divergence up to 40 keV. A model of the key conical crystal bender has been designed and is out for bids. We plan to test the model optically and then with synchrotron radiation from Cornell. Should the conical crystal geometry prove too difficult, we are prepared to fall back on the BNL paraboloidal mirror concept