Matrix Keypad routine

[mister_e]

i has the advantages to be 'user configurable' and support many matrix Keypad type... up to 8X8.

Can be used in a Timer interrupt as well, so you can read from it when you want, etc etc.

Enjoy!

[T.Jackson]

It looks quite comprehensive.

I tend to base most of my algorithms around direct logic - boolean algebraic structures - rather than continuous mathematics, i.e calculations. PIC's are not as well equipped for calculations.

[mister_e]

If that was true... ASM additions, substraction, bit shifting functions and few STATUS register bits wouldn't exist...(carry/Borrow, Z)

[T.Jackson]

If that was true... ASM additions, substraction, bit shifting functions and few STATUS register bits wouldn't exist...(carry/Borrow, Z)

So, you believe that a PIC performs on par with: A = (100 / I) + (J * 2) Vs IF B = 1 THEN A = 120 ?
Huge amount of work for the PIC to do something like this: A = (100 / I) + (J * 2)
100 x the num of clock cycles perhaps?

[mister_e]

Trent, If your example was right i could agree with some part of it... programming method and code optimization is an endless story... i'm not going to argument on which is better, which is worst... it's not what the actual thread is about anyway.

100X ??? NAH, i don't think so. Depending the Value of I, few times maybe... but not 100 for sure. and * 2, is just a Byte shifting, one position to the left... not much... 1-2 cycle not much.
-----------------------------

Charudatt,
before going in sleep mode, make sure those PINs are set to input (with pull-up) and the others are set LOW and OUTPUT (or input but with pull-down resistor). Once you'll press a key, RB<7:4> should see the LOW signal... which should wake your PIC, then you call the keypad scan.

[paul borgmeier]

I tend to base most of my algorithms around direct logic - boolean algebraic structures - rather than continuous mathematics, i.e calculations. PIC's are not as well equipped for calculations.

Huh? Lookups have their place (and I love them when needed) but to say PICs are not equipped for calculations is just plain silly.

For all your projects (that you have posted), was time of the issue or code space? (answer = the latter). With your example equation, which one takes less code space? Which one is faster for larger I? (answer = calculation based approach).

' Calculation based equation A = 100/I +J*2
If I > 100 then
A = J<<1
ELSE
A = 100
FOR X = 1 to 100
A = A – I
IF A < I then OUT
NEXT X
OUT:
A = X+ J<<1
ENDIF

or

' lookup based equation A = 100/I +J*2
IF I = 1 then A = 100+J<<1
If I = 2 then A = 50 +J<<1
If I = 3 then A = 33+J<<1
If I = 4 then A = 25+J<<1
...
' (a bunch more  if-thens)
...
if I =100 then A = 1+J<<1
If I > 100 then A = J<<1

[CluckShot]

I know the Pull up Resistor attaches to Vss but where does it go else wise? Also... What happens to other test programs etc by its presence?

Paul

[T.Jackson]

Huh? Lookups have their place (and I love them when needed) but to say PICs are not equipped for calculations is just plain silly.

I didn't say not equipped - I said not as well equipped. And, I'm more so referring to something that's more so "hard coded" over calculations. When I say hard coded I'm talking about something that's direct as in (a = 100) Vs (a = b x 10 / 2) No prizes for guessing which is faster and uses less resource.