Yokogawa Data Acquisition with PID Control CX2000 User Manual
Page 196

7-4
IM 04L31A01-17E
Calculating the Sum Value
If you set the parameter of the CS command to “1 (enabled),” the checksum value is
output only during serial communications. The check sum is the same as that used in
the TCP/IP and is derived according to the following algorithm.
Buffer on Which the Sum Value Is Calculated
• For the header sum, it is calculated from “data length + flag + identifier” (fixed to
6 bytes).
• For the data sum, it is calculated from “BINARY data.”
1 byte
0
(1)
(2)
(3)
(4)
(6)
(5)
Padding
If the data length of the buffer is odd, a “0” is padded so that it is even. (1) through
(6) are summed as unsigned two-byte integers (unsigned short). If the digit
overflows a “1” is added. Finally, the result is bit-wise inverted.
Sample Program
The sum value is determined using the following sample program, and the calculated
result is returned. The sum determined by the sample program can be compared with
the header sum of the output BINARY header section and the data sum of the output
BINARY footer section.
/*
* Sum Calculation Function (for a 32-bit CPU)
*
* Parameter buff : Pointer to the top of the data on which the sum is
calculated
* len : Length of the data on which the sum is calculated
* Returned value : Calculated sum
*/
int
cksum(unsigned char *buff, int len)
{
unsigned short *p;
/* Pointer to the next two-byte data word in the
buffer that is to be summed. */
unsigned int csum;
/* Checksum value */
int
i;
int
odd;
csum = 0;
/* Initialize. */
odd = len%2;
/* Check whether or not the number of data points
is even. */
len >>= 1;
/* Determine the number of data points using a
“short” data type. */
p = (unsigned short *)buff;
for(i=0;i /* Sum using an unsigned short data type. */ csum += *p++; if(odd){ /* When the data length is odd */ union tmp{ /* Pad with a 0, and add to the unsigned short data. */ unsigned short s; unsigned char c[2]; }tmp; tmp.c[1] = 0; tmp.c[0] = *((unsigned char *)p); csum += tmp.s; } if((csum = (csum & 0xffff) + ((csum>>16) & 0xffff)) > 0xffff) /* Add the overflowed digits */ csum = csum - 0xffff; /* If the digit overflows again, add a 1. */ return((~csum) & 0xffff); /* bit inversion */ } 7.1 Response Syntax