Check Sum is a method where we use modulo summation to detect errors in a stream of data. Usually the sum operation is operated on package of bytes where the last byte is the Check Sum number. When you add all the numbers in the package the sum should be zero (in octets the addition is modulo 256, it means that if the sum reaches to 256 it turns to zero), if from some reason it isn't then we know that there is an error.
Check Sum has no ability to correct errors but only detect them.
Let's see an example:
12 40 05 80 FB 12 00 26 B4 BB 09 B4 12 28 74 11 BB 12 00 2E 22 12 00 26 75 00 00 FA 12 00 26 25 00 3A F5 00 DA F7 12 00 26 B5 00 06 74 10 12 00 2E 22 F1 74 11 12 00 2E 22 74 13 12 00 2E 22 B4
In the example we can see a block of octets that are represented in Hexadecimal. The last number in every row is a number that if you sum the whole row, and there is no error, you should get a zero.
It's important to mention that you should build a reasonable row, that is, not very short where the overhead would be too much, and not very long because if you detect that there is an error you should send the whole message again.