An improved timing scheme has been conceived for operation of a scanning satellite-borne rain-measuring radar system.The scheme allows a real- time-generated solution,which is required for auto targeting. The current timing scheme used in radar satellites involves pre-computing a solution that allows the instrument to catch all transmitted pulses without transmitting and receiving at the same time. Satellite altitude requires many pulses in flight at any time, and the timing solution to prevent transmit and receive operations from colliding is usually found iteratively. The proposed satellite has a large number of scanning beams each with a different range to target and few pulses per beam. Furthermore, the satellite will be self-targeting, so the selection of which beams are used will change from sweep to sweep. The proposed timing solution guarantees no echo collisions, can be generated using simple FPGA-based hardware in real time, and can be mathematically shown to deliver the maximum number of pulses per second, given the timing constraints.

The timing solution is computed every sweep, and consists of three phases: (1) a build-up phase, (2) a feed-back phase, and (3) a build-down phase. Before the build-up phase can begin, the beams to be transmitted are sorted in numerical order.The numerical order of the beams is also the order from shortest range to longest range. Sorting the list guarantees no pulse collisions.

The build-up phase begins by transmitting the first pulse from the first beam on the list. Transmission of this pulse starts a delay counter, which stores the beam number and the time delay to the beginning of the receive window for that beam. The timing generator waits just long enough to complete the transmit pulse plus one receive window, then sends out the second pulse. The second pulse starts a second delay counter, which stores its beam number and time delay. This process continues until an output from the first timer indicates there is less than one transmit pulse width until the start of the next receive event. This blocks future transmit pulses in the build-up phase.

The feedback phase begins with the first timer paying off and starting the first receive window. When the first receive window is complete, the timing generator transmits the next beam from the list.When the second timer pays off, the second receive event is started. Following the second receive event, the timing generator will transmit the next beam on the list and start an additional timer. The timers work in a circular buffer fashion so there only need to be enough to cover the maximum number of echoes in flight.

When there are no more beams to transmit on the list, the build-down phase begins. In this phase,receive events begin when their respective timers pay off.When the timers have all paid off, the sweep is over and the instrument can begin a new sweep with a new list of beams.

Pulse collisions are avoided by the spacing of pulses during the build-up phase and by the order of the beams. As long as the range (delay) never decreases there will always be enough time between any 2 transmit pulses for the receive window and it can occur at its optimal time. The solution is shown by simulation to average 90-percent efficiency in that the instrument is transmitting or receiving (but never both) 90 percent of the time. This can be shown to be optimal, given the constraint that the number of echoes in flight needs to be constant over a sweep. This timing solution is the heart of an onboard processor/controller board for the second generation of Global Precipitation Mission.

The work is being done by Andrew Berkun and Mark Fischman of Caltech for NASA's Jet Propulsion Laboratory, with cooperation from consultant Ray Andraka. For further information,access the Technical Support Package (TSP) free on-line at www.techbriefs.com/tsp under the Electronics/ Computers category. NPO-30560