buttons.c (demo branch)

[VonSzarvas]

Buttons is broken because some pins are hard-defined.

Suggested update (with original code in comments after each changed line):

old:

low(2);

high(4);
waitcnt(CLKFREQ/4000 + CNT);
tDecay[1] = rc_time(4,1);

high(27);
waitcnt(CLKFREQ/4000 + CNT);
tDecay[0] = rc_time(27,1);
low(2);

new:

high(NAV_TOUCH_L);
waitcnt(CLKFREQ/4000 + CNT);
tDecay[1] = rc_time(NAV_TOUCH_L,1);

high(NAV_TOUCH_R);
waitcnt(CLKFREQ/4000 + CNT);
tDecay[0] = rc_time(NAV_TOUCH_R,1);

---

rc_time leaves the pin as an input when it's done, right?

Extra idea- After the above code block, set the pins low to mitigate interference?

low(NAV_TOUCH_L);
low(NAV_TOUCH_R); 

[PropGit]

@VonSzarvas - Yes, rc_time should be leaving the pins as input when it's done. @AndyLindsay can correct me if I'm wrong.

If we later set the pins low, they'll be set back to outputs. Is that what you intend, so there's no floating button lines (is that the "interference" you mean)?

Question: What's on I/O 2 (that was set low twice in the old code but not at all in the new)?

[AndyLindsay]

I don't think I wrote that. The button code for the non-WX badge was this:

int button( int pad )
{
  char *addr = (char *) tself->p_pinslist;
  int pin = addr[6-pad];
  int ctr = 0;
  for(int i = 0; i < 5; i++)
  {
    high(pin);
    pause(1);
    input(pin);
    pause(5);
    int pb = input(pin);
    if(pb) ctr++;
    low(pin);
    pause(1);
  }
  return !(ctr == 5);
}

int buttons(void)
{
  int pb = 0;
  int mask = tself->pinsmask;
  for(int i = 0; i < 5; i++)
  {
    //high(pin);
    OUTA |= mask;
    DIRA |= mask;
    pause(1);
    //input(pin);
    DIRA &= (~mask);
    pause(5);
    //int pb = input(pin);
    pb |= ~INA;
    //low(pin);
    OUTA &= (~mask);
    DIRA |= mask;
    pause(1);
  }
  char *addr = (char *) tself->p_pinslist;
  int pads = 0, pin;
  for(int i = 0; i < 7; i++)
  {
    pads <<= 1; 
    pin = addr[i];
    int bit = (pb >> pin) & 1;
    pads |= bit;
  }    
  return pads;
}

[VonSzarvas]

Io2 was reassigned to another sensor. Instead the other side of the passive buttons is permanently low. That links to why I wondered if setting the 2 passive button lines low when not being scanned would remove potential floating problems from the high impedance. Its 470k on the line to gnd.

[VonSzarvas]

@AndyLindsay @MatzElectronics

Andy, yes the code changed for BadgeWX. Maybe it could be improved, but I can confirm it works great as is. Matt might have crafted these edits on the original badge code?

• for now, all I can see needs doing is the hard-coded pin numbers changed to the define names so that it will work with RevA2 and RevA (as some pins changed from RevA1).

Impact note: Removing IO2 low will break compatibility with RevA1 boards for the touchpads, but I'd suggest we are past the point to concern with that. (Especially considering- RevA1 code can be compiled in SimpleIDE with overrides if needed)

[MatzElectronics]

Should be fixed now. This can be closed.