Hi,

My problem is slightly off topic from pic's, but can anyone explain to me
(or point to any source of info) the idea of a CRC error detection algorithm,
which I've stumbled across ? I've thoroughly read the "painless guide to
CRC error detection algorithms" document available on the net, but this
algorithm works another way and is still dark to me.

It's a part of a RDS (Radio Data System) decoding software. It computes
a 10-bit value from a serially received 26-bit block containing 16 bit of data
and a 10-bit CRC checkword. MSB is received first. CRC uses the following
polynomial:

g(x) = x10 + x8 + x7 + x5 + x4 + x3 + 1

The algorithm is very simple:

Clear the Register with zero.
For each received bit set:
  Register = Register XOR table[consecutive bit number]

And here is an equivalent C program:

  r=0;                           /* computed value */
  for (i=0; i<26; i++)
    if (received_bit()==1)
      r^=table[i];

int table[] =
{
  0x8000, 0x4000, 0x2000, 0x1000, 0x0800, 0x0400, 0x0200, 0x0100,
  0x0080, 0x0040, 0xB700, 0x5B80, 0x2DC0, 0xA1C0, 0xE7C0, 0xC4C0,
  0xD540, 0xDD80, 0x6EC0, 0x8040, 0xF700, 0x7B80, 0x3DC0, 0xA9C0,
  0xE3C0, 0xC6C0
};

I've tested this algorithm and it returns 0 if the checksum of the received
block is correct, but how does it work and how was the table created ?
Any help would be appreciated.

  Piotr

Piotr Piatek      pisielek@inet.com.pl