The predicted result of observing the incident spectrum of Figure 3 with this payload is shown in Figure 5.
This figure shows the pulse height spectrum predicted for a payload consisting of two EEV CCDs, which are described in detail in section 4.2.3. The simulations include a 1200Å Al/Ti blocking filter to eliminate optical and UV light. Also included in the figure is an estimate of the expected particle background rate assuming a particle rejection efficiency of 99.5%, which we expect to achieve in flight.
The detectors have
fields of view of 
per pixel for incident
energies below 1 keV, and 
per pixel
for incident energies above 3 keV, with a transition region at intermediate
energies. This is accomplished through the use of a partially transmitting
collimator, described in section 4.2.10, and results in
a bulge in the effective area above 2 keV which is visible in Figure 4.
This collimator design is critically important, since it allows
reasonably good angular resolution at low energies where galactic emission
is anisotropic and has interesting features with typical angular scales of
, but opens up to wider solid angles at high
energies where the incident flux is lower but is isotropic. This permits
the sky flux
to dominate over the expected particle background by at least an order of
magnitude over our entire instrumental bandpass.
The CUBIC spectrum shown here is a simulation of a single 50,000 s pointing.
The CUBIC experiment will provide high quality,
moderate resolution
spectra like that shown in Figure 5
over regions of 
on the sky.
The predicted sky count rate for this simulation is 
counts per
second, with an expected internal background rate (after charged particle
rejection) of 
counts per second, and the spectrum
contains 93,000 counts.
(The simulation was performed for a fairly dark portion of the 
keV
sky. In some regions of the sky, the count rate may exceed 5 cps.)
Figure 5 shows that
the major line groups below 1 keV are clearly resolved in the CUBIC \
spectrum.
If we achieve a 25% observing efficiency, this observation will require about two days of integration. Clearly considerable spectral information can be obtained on shorter timescales; however, we note that a baseline of 20,000 - 50,000 second integrations requiring a spacecraft roll no more often than once every day or two is commensurate with minimal mission planning and mission operations resources. If the target directions are chosen carefully, we could obtain such spectra for the entire sky in 15 months of operation, although it is unlikely that this will be achieved in practice. We may wish to observe individual regions for longer than 2 days in order to obtain high quality spectra for subfields, using the coarse spatial resolution of the ``pinhole'' collimator. Because of the isotropy of the extragalactic portion of the diffuse X-ray background, which dominates above 1.5 keV, the high energy data from the entire mission can be combined to provide an extremely high precision CXB spectrum.
In order to take advantage of the opportunity offered by SAC-B , we have built the CUBIC experiment in-house as inexpensively as possible. We emphasize that this is a satellite experiment built on (nearly) a sounding rocket budget. It is viewed as a relatively high-risk, low-cost opportunity to put a science payload on an engineering test flight of the Argentine SAC satellite platform. It has been built in the ``sounding rocket mode'', with every effort made to reduce costs, simplify the experiment, and finish it on schedule. In keeping with this philosophy, we minimized redundancy and the use of specially qualified components except for the most critical elements of the experiment. Our goal is to get the most science possible for the dollar, not to ensure a fixed lifetime for the experiment.
