Simple Button Press on P1.1 to count on LCD using a msp430fg4618

Randy Adams said:

dispChar(pos, index);

if (!(P1IN & BIT1)) // if P1.1

{

index=index+1;

if (index > 9)

{

index=0;

dispChar(pos, index);

pos = pos+1;

index=1;

 Hi Randy, this is not question of MSP430 but method, think what happen in your code at processor speed and how many times processor execute your lines when you press button.

 You need develop a way to do action just when button is pressed and never repeat till this is released.

 this is MSP430 forum, question is unrelated and can be found on C programming basis:

 use a variable to store status of button and a small delay so:

 int oldbuttonstate = 0;

 int buttonstate;

// change if(!(P1IN & BIT1)) // if P1.1 to

buttonstate = !(P1IN & BIT1);

if( (buttonstate != oldbuttonstate) && buttonstate) // if button changed and button pressed

{ // button get depressed so take action and set old status

 oldbuttonstate=buttonstate;

..... code was on if clause

} // this is where old if finished

 else

   if(!buttonstate) // if button get released

oldbuttonstate=buttonstate; // reflect new state to old button state

this can cure partially your issue but need a delay to avoid another issue, think about what happen when you press button, it is not immediately closed but get closed and open few times in a few mS, this phenomena known as bouncing repeat some times and is called bouncing...

 so try yourself or read about...

Randy Adams said:

I will post my question in the C programming basis

 Hi Randy, C forum is not on TI platform, your trouble can be instead an MSP related so .. try isolate why it require a lot of times, it need time to initialize lcd or a lot of time just when you press button?

 From increasing numbers something I think can be better, Other issue are probably present.. think what is in your code and continue execute at processor speed...

 What you think happen when you press button? Now it is locked and don't continue execute code but still doesn't work...

 Why?

Hints:

 what do pos variable and what happen to?

Is not this program destroying all memory by some mechanism?

Katie Pier said:

I'm trying to make sure I understand what is happening with your code - are you wanting your display to increment the digit by 1, each time that you press the button?

 Hi Katie, I like too much your proposal but I fear this is like using a cannon to shut pigeons.

 Some part like display is not present but I suppose you can point what can happen.

 Just a doubt about xtal18pF I remember this was a too high value or not?

ISR is a difficult to grasp topic to who is troubled of initial learning, I think from my experience is better start guide learner how to walk than proposing the best rocket thruster.

 Yes debouncing is what it is missing and need some learning too. First close the way to button edge detect then some way to debounce and problem was learned from root. Some other issue can be present but we have to dig in.

 Randy,  can you explain better the delay how and when it happen? So you have to wait then it work, what are you seeing on display? is display high contrast or it is grayed?

Katie I proudly wait your opinion, and Randy too of course ;)

Randy Adams said:

Katie,
I have the MSP430 Experimenters board. I started with the LDC example program and removed the extra printed functions. I then added in an simple loop to count when a button is pressed. I can send you my entire code if you would like. I think that I am not understanding the interrupts of this chip. I teach C-Programming and the Atmel Chip and can do anything I want with this chip. I hope to do the same with this chip. All I need is to get pass this program and I think I can take it and run with my other programming. Thanks
Randy

 Hi Randy, sorry, I read about how and why you written this program, please accept apologizes about, so can be if you still are in trouble can help:

 I wrote some short notes about LPM and interrupt but  I think this can help you point to solution and introduce where trouble can b:

 load code thru debugger, then locate timer interrupt service, at end there is an instruction similar to this:

__bic_SR_register_on_exit(LPM3_bits);  // Clear LPM3 bits from 0(SR)
place a breakpoint just before this instruction then start single stepping program over function, so you reach the first instruction after for(;;)

here all stop, just wait for a while and you get at interrupt BP, so from there step over and I am sure you immediately grasp what happened when you loop over.

 after for(;;)

_BIS_SR(LPM3_bits + GIE); // LPM3, enable interrupts

 this instruction enable the interrupt but also enter immediately LPM3, this stop cpu and never get key read before timer interrupt wake up cpu,

 move to end of loop to avoid long wait for cpu wake, or if you like avoid interrupt and LPM  for now just comment and use loop.

also check variable pos = pos+1;

 I dont know if in dispChar(pos, index); can do some damages to memory if overflow to number of digit.

Another hint, MSP is a word machine and is not so efficient with byte sized variables, use as much as possible integer or 16 bit values if you need speed and leave byte just if you are in a tight space or for character buffer.

 About delay it can be on interrupt of timers, probably at end of irq service code  get some decision on LPM status.

 If you wish post interrupt code I can help if you like.

About this:

_BIS_SR(LPM3_bits + GIE); // LPM3, enable interrupts

 this set bit of status register: GIE General Interrupt  Enable

but also bit of LPM state -> Low Power Mode, LPM3 stop CPU so it stop execution from here.

if you use on interrupt it can enable interrupt to nest higher priority care about  LPM  can stop execution.

  A special instruction is at end of service to set not status register but saved image of status register on stack to wake up or go to LPM on return.

this instruction manipulate image of status register on stack pointer, effect on status register is applied on return from service.

__bic_SR_register_on_exit(LPM3_bits);  // Clear LPM3 bits from 0(SR)

this manipulate status register and execute immediately, program stop immediately if LPM is entered. Many level of LPM are available and deep one stop also clock.

_BIS_SR(LPM3_bits + GIE); // LPM3, enable interrupts

Roberto Romano said:

Get from here the software examples, they where addressing all feature one at time for this processor not specific to board.

 Randy, about C style, today I remembered one thing about this board:

 Software if not rewritten was one of worst I seen from TI (ages 2006 till 2008), so check if on interrupt service there are some asm clause, this case the best way before get crazy about is to completely rewrite code from scratch.

 I remember it was manipulating stack to force return to different routine and then performed some stack corruption too. If code was not clean and you still have old CD on box it is hard to modify and get something useful, crash of code are quite guaranteed, this manipulation was on key pressure interrupt so be careful about.

 I don't know why you selected this board but if you intend use on a class evaluate launchpads, series 2G has advantage of dil, can be easily  fit on a breadboard with just one resistor on reset and decoupling capacitor on power rails, price are a fraction of this one.

 One other  is  the 5529 for computing power and USB too.