This paper presents a neural circuit using PWM technique based on an analog
-digital merged circuit architecture. Some new PWM circuit techniques are p
roposed. A bipolar-weighted summation circuit is described which attains 8-
bit precision in SPICE simulation at 5V supply voltage by compensating para
sitic capacitance effects. A high performance differential-type latch compa
rator which can discriminate 1 mV difference at 100 MHz in SPICE simulation
is also described. Next, we present a prototype chip fabricated using a 0.
6 mu m CMOS process. The measurement results demonstrate that the overall p
recision in the weighted summation and the sigmoidal transformation is 5 bi
ts. A neural network has been constructed using the prototype chips, and th
e experimental results for realizing the XOR function have successfully ver
ified the basic neural operation.