Germanium detectors are extensively used in astronomical instruments for far infrared observations. To meet the science objectives of future space projects, large-format far IR detectors are needed. As a first step toward this goal, we have fabricated a 2x16 Ge:Sb array with the 1x32, SB-190 CTIA cryogenic readout. The detector design as well as the preliminary results of our parametric tests are presented here. The array exhibits very good noise performance with an NEP as low as 4.0E-18 W/√Hz, and confirms the viability of the design for large format arrays.Far infrared and sub millimeter detectors and detector arrays have drawn increasing interest from NASA and the astronomy community is general. In the most recent workshops on far IR and submillimeter detector technology sponsored by NASA in 2002 [1,2], the pivotal role of far IR arrays was reiterated and the need for further advancement in this area was reemphasized. The experts' consensus expressed in these workshops was also emphatically echoed in the report issued by NASA's Infrared, Submillimeter, and Millimeter Detector Working Group (ISMDWG) [3] as well as the Origins Roadmap [4].Far IR arrays constitute the key technology for SOFIA (Stratospheric Observatory For Infrared Astronomy) and the future far IR space telescopes that are currently on the drawing board. NASA's Single Aperture Far-Infrared Observatory (SAFIR) is conceptualized as a 5K, 10m space telescope operating in the 40µm-1mm wavelength range [5,6]. Its target launch-date is 2020. SPICA (SPace Infrared Telescope for Cosmology and Astrophysics), a.k.a. HII/L2, is a Japanese 3.5m space telescope targeted to be launched in 2010 and will cover the 5-200µm wavelength range [7]. Both of these telescopes require large format (larger than 64x64) far IR detector arrays with cryogenic readout multiplexers, the technology that is currently immature.In order to take full advantage of the scientifically rich far IR and submm spectral range, large-format arrays similar to those currently operating in near and mid IR regions are needed. The fundamental advantage of large-format arrays, of course, is that the observing time scales as (Sensitivity) 2 /(Number of pixels). The desire for large formats can best be fulfilled by a suitable direct hybrid architecture employing planar, indium bump-bond technology. Although this technology has been used successfully with short wavelength detectors for many years, its extension to long wavelength (>50µm) is neither automatic nor trivial. This is due to the fact that the operating condition as well as the material for far IR and submm detectors are substantially different than those for the near and mid IR detectors. In particular, the requirement for low bias levels, deep cryogenic operation, the glow from the readout electronics, detector heating, and thermal mismatch between the readout and the detector array impose different restrictions that are incompatible with the present direct hybrid design. A different approach is, therefore, needed to address all the issue...