Algorithm for single error correction using H codes 1

[trevorsync]

I have been able to generate an algorithm to randomise a bit position where an error is likely to occur.Which basically generating an algorithm to select different bit positions of the transmitted codeword for different integer values obtainable from the algorithm.This scheme is employed in the single bit error correction and single bit error detection.

I am currently faced with a problem of randomly selecting any two different bit positions in such a way that every bit position is selected for different integer value obtainable from the algorithm

I urgently need help on this.Thx

 

[VEBill]

For the ultimate in flexibility and speed, prepare a table (list) in advance. Then you can randomly point into the table to select one of your pre-calculated errors. For 8-bit words, the list would be 8x7=56 elements long (trivial). This approach might even be more compact than the code to choose two different bits on the fly.

PS: Watch out for flakey Random routines. Personally, I'd let it run for a while and make a histogram to confirm that it's hitting all the values.

 

[nbucska]

What is the origin of your data (disk, tape, serial line,
CPU to CPU, keyboard, sensors etc) ?

Tharsmission speed? max. tolerable delay ?



----------------------------
Please read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[trevorsync]

Thank you VE1BLL and Nbucska for the posts.I have started work with the syndrome table as an alternative.It is a far better approach to ecc and edc.Thank you ve1bill

As regards the origin the data...i am not too particular about that.I would like for us to assume a 4 bit binary data.

 

[trevorsync]

VE1BLL, please could you give me links which would help with generating a syndrome table for different data lengths

 

[VEBill]

MS-Excel is your friend.

 

[trevorsync]

For SECDED, i understand a hamming code of distance 4 is to be used.For a 4 bit data word, we would be needing 4 check bits to implement SECDED.

Normally, for a 4 bit data implementing SECSED we would be using the 7,4 matrix which is of the form

1110100
1101010
1011001

The above code of of hamming distance 3.Now,the problem i am facing is constructing a code with hamming distance 4 for implementing SECDED?

 

[nbucska]

I have written for the PC in assembly a fairly fast
program to encode different ECC-s and detect the error
It is easy to add more polynomials. Presently it has only
three 32th order and one of each 16,48 and 56th
order polynomials.






----------------------------
Please read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[trevorsync]

I finally solve the problem.Thank you engineerig forum.I need someone to give me tips that would be helpful in taking this project to another level.