Conversion fromBCD(x)

[Guest]

Hello !

Could somone explain me in details what these lines means, please ?
It is supposed to be useful for a conversion on a DS1306.
I works very well, but i would like to understand it.

Many Thanks !

[Ttelmah]

I recognise that code. :-)
Two seperate functions. One to convert a binary number to BCD, and one to convert a BCD number to normal binary.
BCD, uses the top 'nibble' of the binary value to represent the tens, and the bottom nibble to represent the units. Hence to convert 'from' BCD, you want the top nibble (val & 0xF0) divided by 16, times ten, plus the value of the bottom nibble.
So the number '0x89', has to become '89'.
Now fortunately the shift right operation, (>>4), gets rid of the bottom bits as it shifts, and shifts '0' into the top bit (so you don't need to mask with 0xF0). So the line 'fromBCD', takes the value shifted right four times (which is the same as val & 0xF0 divided by 16), and multiplies this by ten, by using the value from the array. This is slightly quicker than performing the actual multiplication, and allows 'illegal' values (greater than 9), to be clipped, and then adds the low nibble (val & 0xF).
Hence with the '0x89' example, you get the total of '80', plus '9', giving the required '89'.
toBCD, has to do the opposite, It does this by comparing the value with 10, 20, 30, etc., from the array, then if it matches, simply returning the index in the array*16 as the BCD number. If the value is below the compared value (so for instance, it is 89, and the comparison gets to 90), then it reduces the index by 1 (we are in the eighties), subtracts the array value from the source (giving us '9'), and returns the array index (8) * 16, plus the remainder (9), giving '0x89'.
There are literally dozens of ways of coding this, the particular version has the advantage of being quick on the PIC 16 chips.

Best Wishes

[Guest]

Thanks a lot that really helps me !