comparing input signals in assembly

Comments

• JonnyMac Posts: 9,342

  2011-05-06 12:52 edited 2011-05-06 12:52

  Do you have a pull-down on P15? If not, releasing the button may not cause the pin to change state. You must pull any pin used as an input to a known state (to ground for active-high inputs, to Vcc for active-low inputs). A 10K resistor to ground of Vcc is fine (one resistor per pin).
• kwinn Posts: 8,697

  2011-05-06 12:53 edited 2011-05-06 12:53

  Don't forget that in assembly (PASM) when you read the pins you get all 32 bits. You have to mask off the unwanted pins with an AND instruction.
• latofte Posts: 9

  2011-05-06 13:25 edited 2011-05-06 13:25

  Hmm, I put a pull-Down on both 14 and 15, then made the output from the switch also go to 14 and 15 so when the switch is flipped they have the same input. But it still does not seem to like something?
  
  I also added Bits res 1 at the end and changed the code to
  ....
  Mov Bits, Ina
  And Bits, Pushed
  Cmp Bits, Pushed
  IF.....
  
  But still not accepting that the two are equal
• tonyp12 Posts: 1,951

  2011-05-06 13:28 edited 2011-05-06 13:28

  I think he is already using a mask called Pushed with his bitwize test instruction.
  but 01 and 10 and 11 will both produce Z=0
  
  1st you need a pull-down resistor as stray electricity make keep the input in a unkown state
  as it's pretty much is a antenna if not connected to 3v or ground.
  
  
  And your code can be simplified alot, think statemachine.
• Ariba Posts: 2,693

  2011-05-06 13:29 edited 2011-05-06 13:29

  test Ina, pushed
          If_E            jmp #:LED
          If_NE           jmp #:LED2
          {If both pins are high it should jump to faster blinking LED}
  

  This has several errors:
  1) ina must be the source register, otherwise you read the shadow ram
  2) you nead to specify the flags that you will have affected with wc, wz
  3) TEST does an AND of the INA and the Mask in "pushed". pushed masks both pins, so you can only detect if both pins are low or not. But Pin14 is always high so you will never get the zero flag set:

  Pin 14 is connected directly to 3.3v and pin 15 is connected to a switch which is also connected to 3.3v

  
  If you want compare two input pins, the best way is to read one pin into the Carry flag and the second pin into the Zero flag, then you can use the IF_C_EQ_Z and IF_C_NE_Z conditions:

  test firstPinMask, ina  wc
               test scndPinMask, ina   wz
    if_c_ne_z  jmp #LED   'Equ
    if_c_eq_z  jmp #LED2  'notEqu
  

  Be aware the the Zero flag is set if the pin is low and vice versa,but the Carry flag is set if the pin is high.
  
  Andy
• davidsaunders Posts: 1,559

  2011-05-06 13:37 edited 2011-05-06 13:37

  The trouble is that binary 010_0000_0000_0000 is not equal to binary 100_0000_0000_0000, these are the values for the two pins as active.
• latofte Posts: 9

  2011-05-06 13:45 edited 2011-05-06 13:45

  Awesome thanks guys!
  I was trying to figure out how to use those flags, now hopefully I can get a few more practice problems under my belt before I come begging you guys again
• tonyp12 Posts: 1,951

  2011-05-06 13:50 edited 2011-05-06 13:50

  Do you plan to have two buttons, and fast blink if only both pins are high?
  Or fast blink if both buttons are the same 0=0 or 1=1 ?
  
  Not forgetting wz or wc with the opcode test, is a must.
  
  Inverting the switches, by using pull-up resistor and switch to ground.
  and you can get away with the single: test mask, ina wz
  as z=1 only if all pins in mask=0, if you want that both switches need to be pushed.
• kuroneko Posts: 3,623

  2011-05-06 17:16 edited 2011-05-06 17:16

  Ariba wrote: »

  If you want compare two input pins, the best way is to read one pin into the Carry flag and the second pin into the Zero flag, then you can use the IF_C_EQ_Z and IF_C_NE_Z conditions.

  Not sure about best but test reflects the parity of the result which means you do one test and look at the carry flag.
• Ariba Posts: 2,693

  2011-05-06 19:24 edited 2011-05-06 19:24

  kuroneko
  
  Nice solution!
  I need to be more careful with adjectives like "best", if there is a possibility that you read the thread
  
  
  Andy
• kuroneko Posts: 3,623

  2011-05-06 19:35 edited 2011-05-06 19:35

  Ariba wrote: »

  I need to be more careful with adjectives like "best", ...

  Admittedly best could also mean easy to understand for a beginner(?) so in that case I'd go with your version which makes it obvious where each bit goes and what's done with it (conditional execution).
• kuroneko Posts: 3,623

  2011-05-06 20:10 edited 2011-05-06 20:10

  latofte wrote: »

  Button
  mov dira, output 'sets only 14 and 15 to input
  ...
  output   long %[COLOR="red"]11111111_1111111[/COLOR]1_00[COLOR="red"]111111_11111111[/COLOR]  
  

  I wouldn't do that. Only ever enable outputs for the pins you are going to drive (e.g. LED on pin 16). There could be h/w connected to the other pins which disagrees with the level you're driving those pins with (especially 28..31). So in your case a mov dira, Pin16 is all you need.