This filter closely approximates the desired wave shape, regardless of the bit rate.

A digital solution for the problem of filtering serial digital data prior to modulation of a telemetry transmitter has been developed by engineers at the Dryden Flight Research Center. The solution is described in a patent application entitled "DIGITAL APPROXIMATION PREMODULATION FILTER FOR PULSE-CODE-MODULATED SIGNALS."

Figure 1. The Digital Approximation Waveform is a slightly degraded version of the output of an analog premodulation filter.

When transmission of serial digital data for telemetry is required, care must be taken when modulating the transmitter to avoid any spurious radiated signals outside the assigned radio frequency band. A premodulation filter is used to eliminate undesired harmonics, thereby limiting the frequency bandwidth of the serial digital signal.

The use of a digital approximation (as distinguished from an analog) premodulation filter entails some degradation of the infinite output resolution of a traditional analog filter (see Figure 1), but this degradation should be tolerated in the interest of keeping out-of-band radiation below the maximum allowed by regulations.

Figure 2. An N-Segment DAP Filter includes a shift register, a summing amplifier, and an output buffer.

The digital approximation premodulation (DAP) filter concept was tested in transmission of a pulse-code-modulation (PCM) system non-return-to-zero-level output via an L-band transmitter. Normally, 16 segments per bit are used, but in this test, the radiated frequency spectrum obtained when using the DAP filter with only 8 segments was essentially identical with that obtained when using an analog filter.

The DAP filter (Figure 2) comprises three main parts: (1) a part that divides each bit into a number of segments, (2) a part that converts each segment into a voltage that approximates the output of an ideal premodulation filter, and (3) an output buffer amplifier. An integer frequency multiple of the bit-rate clock signal is used to shift the serial digital output through a parallel output shift register to segment each bit. The conversion to a voltage level can be accomplished by a resistor array and summing amplifier to approximate the rise or fall of a half cosine wave for each bit transition. If the integer frequency multiple of the bit-rate clock signal is not available from the signal source, it can be generated from the source bit-rate clock signal by use of a phase-lock loop.

Because the DAP filter is synchronized with the bit-rate clock as described above, it always produces the proper wave shape, regardless of the bit rate of the unfiltered data signal.

This work was done by Harry Chiles and Rod Bogue of Dryden Flight Research Center.

This invention is owned by NASA, and a patent application has been filed. Inquiries concerning nonexclusive or exclusive license for its commercial development should be addressed to the Technical Information Specialist, Dryden Flight Research Center, (805)258-3720. Refer to DRC-95-28.