Gerhard,

thanks for your time and code fix. I think you can be a good teacher
I liked the way you explained and it became lot more clear after your
explanation what I was doing wrong. I changed the code the way you 
recommended
and parity started to work. Now my next step is to use single counter 
instead of
2 counters like you said. i will post updated code later today.

once again thanks for time and effort

Andre




Gerhard Fiedler wrote:

>Andre Abelian wrote:
>
>  
>
>>I got basic spi running now I am having hard time to get odd parity bit  
>>9-th  bit  working
>>here is the code any help will appreciate.
>>    
>>
>
>Below I extracted the part of the code that seems to create parity. A
>couple general comments:
>
>- Try to use a consistent formatting style. It's not that important how
>exactly it is, but consistency helps a lot with spotting errors when
>looking at the code. With formatting style I mean placement of braces,
>spaces etc.
>
>- Try to use helpful variable names. "temp" is actually "parity" here; it
>doesn't do anything "temp". 
>
>
>Now a few specific comments:
>
>- Is there a reason why you use i as index in the first for loop and count
>as index in the second for loop? Seems to me that both loops could use the
>same index. (But see below regarding i and regarding the two loops.)
>
>- When debugging, it sometimes is awkward to use variable names with one
>letter (for example, it's difficult to place the cursor over them for the
>editors that show the type and the debuggers that show the type and value).
>I use e.g. "ii" instead of "i". Just as quick to type, and avoids the
>one-letter name problems.
>
>- Is there a reason you need two for loops? Both loops loop over the same 8
>bits. You could keep track of the parity while you are shifting out the
>bits, in the same loop.
>
>- Is there a reason you restrict the parity to 7 bits, even though you seem
>to send 8 bits? (Your data format seems to be PS2, but I'm not familiar
>with this protocol.) Usually you would either restrict already what gets
>sent to seven bits and make sure the 8th bit is in fact 0, or you'd
>calculate the parity over the full 8 bits. 
>
>- I'm not quite sure about the data format you want to generate (again, I'm
>not familiar with PS2). Are you sure you don't need a high/low clock pulse
>after setting the data output to the parity?
>
>- Last but not least (actually, that's now the real issue with the parity
>:), it seems that the parity generation is kind of odd. You write three
>times to the first bit of parity: the first time when initializing it to 1,
>the second time with the xor before the second for loop, the third time
>with the first loop execution of the second for loop. I didn't follow it
>through, but it just doesn't "feel right" :) -- and this would be what I
>suggest you go through and clean up. The xor before the second for loop
>seems superfluous, and the initialization of parity seems wrong.
>
>
>    int parity = 1;      // 1=ODD, 0=EVEN         <<< 1st time
>    int data_copy, count;
>    data_copy = data;
>
>    data_copy &= 0x7f;   // trim to 7 bit ASCII
>    parity ^= (data_copy & 1);                 // <<< 2nd time
>    for (count = 0; count < 8; ++count) 
>    {
>        parity ^= ((data_copy >> count) & 1);  // <<< 3rd time
>    }
>
>Gerhard
>
>
>  
>
>>void ps2_out(BYTE data)
>>{
>>      BYTE i;
>>      int temp=1,data_copy,count;
>>      data_copy=data;
>>      output_low(PS2_CLK_OUT);  // start bit
>>       delay_us(35);
>>      for(i=0;i<8;++i)
>>      {
>>      output_bit(PS2_DATA_OUT, shift_left(&data,1,0));
>>      OUTPUT_HIGH(PS2_CLK_OUT); //
>>      delay_us(35);
>>      output_low(PS2_CLK_OUT);
>>      delay_us(35);
>>      }
>>      data_copy &= 0x7f; //trim to 7 bit ASCII
>>      temp^=((data_copy) & 1);
>>      for (count = 0; count < 8; ++count) {temp ^= ((data_copy >> count) 
>>& 1);}
>>      if (temp==0){output_high(PS2_DATA_OUT);} // even
>>      if (temp==1){output_low(PS2_DATA_OUT);} // odd
>>      output_high(PS2_CLK_OUT);  // stop bit
>>      delay_us(200);
>>     
>>}
>>    
>>
>
>  
>
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist