the bottleneck of the system. In order to speed up the
operation of the multiplier, radix-4 modified Booth
encoding was utilized [5]. Since the adders and subtractors
have only minor contribution to the critical path, they
were implemented using basic ripple carry structure. The
complete signal synthesizer architecture is given in Figure
3. It consists of the complex multiplier presented
above, two register files (one for the feedback values, and
one for the coefficients), and a FSM to control the operation.
The signal synthesizer architecture was synthesized
for a 0.35 mm 3.3 V 4-metal n-well CMOS-process.
According to the results the maximum system clock frequency
is 80 MHz. The total area is 12,000 logic gates.
The extra logic required for the implementation of the
above presented amplitude control method results into an
area increment of approximately 13%.
The synthesis results for a 16-bit digital synthesizer are
presented in Table 1. The table also gives the performance
of two other angle rotation based signal synthesizer
methods for comparison purposes. As can be seen from
the Table 1. the performance of the complex multiplier
based synthesizer is competitive with the other designs.
A clear advantage of this approach over the conventional
ones is that the architecture is most suitable for resource
sharing. If the complex multiplier is not required for signal
synthesis, it can be used for general complex computation
purposes.