Substraction

[Joe]

Hello,

how to do a substraction of 0x50 - 0x45 to have the result 0x05 and not 0x0b ?

Thanks a lot.

[JohnKennedy]

Simply you can't as the answer is correct at 0x0b if you convert the hex to decimal and do the maths you will see this (80 - 69 = 11 which is 0x0b in Hex)

HTH

JFK

[Ttelmah]

You won't.
Decimal 50-45, gives 5. Hexadecimal, 0x50 (decimal 80), minus 0x45 (decimal 69), gives 0xB (decimal 11).
If you want five as a result, start with the numbers in decimal, not hex.
I think in fact the numbers you have, may be binary coded decimal (BCD), not hex. If so, there have been routines posted here on the forum in the past for doing BCD conversions, and what you need is to convert the incoming numbers to decimal, before performing the arithmetic, and then convert them back (if the values have to go back to the same chip).

Best Wishes

[Joe]

Ok thanks, i hope it is ok, in fact it is bcd exactly how to convert BCD to decimal ?

Is that correct

[Ttelmah]

Yes. This is a routine I posted to do this. You need the array 'conv' defined as well before it'll work.

Best Wishes

[Joe]

Ok thanks i think i figured it out !

Thank you very much !

[Joe]

I forgot to ask something !

If i want to make operations on BCD i can't, isn't it ?
I have to convert it in decimal and then make operations on it and then re convert in BCD ? Is that right ?

I can't substract with BCD ?

[Ttelmah]

You could potentially write BCD maths functions. You would have to seperate out the digits, do the arithmetic on these, and then rebuild the final number. The other way of doing this, is the 'ADD adjust' operation, where you look at the result of the decimal arithmetic, and the carry bits, and work out how to adjust the result to give the required BCD value. This operation is a standard function on some processors (the Z80 for example), but does not exist on the PIC. BCD, has fallen from favour in recent years, so processor support is not common on modern chips. The simplest solution, would be something like: