The TMS320C32 was specifically designed to bring the benefits of floating-point processing to high-volume DSP application markets where cost considerations have typically dictated the use of fixed-point processors. With pricing at U.S. suggested resale of $9.95 in high volume, product engineers can use the same floating-point technology for device prototyping, eliminating the customary time-consuming switch to lower-cost devices for commercial deployment.
Featuring a 144-pin plastic quad flatpack (PQFP), the 32-bit 'C32 uses unique memory interface and data-packing features to allow execution of code from 16-/32-bit-wide memories and access of 8-/16-/32-bit data. The 'C32 also features faster instruction cycles (33, 40, 50 ns), edge- or level-triggered external interrupts, and two power management functions (IDLE2 and LOPOWER), plus on-chip, dual DMA channels with configurable DMA/CPU rotating priority.
Because the 'C32 is object code compatible with the 'C3x family, product designers can take full advantage of the widespread development support, including high-level language compilers as well as third-party boards and software. The 'C32, a register-based processor, is much easier to program than fixed-point, accumulator-based DSPs, and also offers higher performance with multiple address generators and multiple-bus structure. Because the 'C32 incorporates features to handle high-end fixed-point and high-volume, low-cost floating-point designs, it is well-suited for applications such as speech recognition, bar code readers, audio compression, embedded control, and telecom.
The device is available for limited sampling today in 40- and 50-MHz versions. A 60-MHz version is planned for 3Q95 during full-volume production.