A96G140/A96G148/A96A148 User’s manual
16. USART 2
215
16.14
0% error baud rate
USART2 system of A96G140/A96G148/A96A148 supports floating point counter logic for 0% error of
baud rate. By using 8-bit floating point counter logic, cumulative error to below the decimal point can be
removed.
Floating point counter value is defined by baud rate error. In the baud rate formula, BAUD is presented
in the integer count value. For example, If you want to use the 57600 baud rate (fXIN = 16MHz), integer
count value must be 16.36 value (BAUD+1 = 16000000/(16×57600) = 17.36). Here, the accurate BAUD
value is 16.36. To achieve the 0% error of baud rate, floating point counter value must be 164 ((17-
16.36) x 256 ≒ 164) and BAUD value must be 17. Namely you have to write the 164 (decimal number)
in USART_FPCR and 17 (decimal number) in USART_BAUD.