Quad Encoder Problem

[HenrikOlsson]

Hi,

What's weird is the 'MotorRPM' as displayed on the LCD jumps by 3-4 (i.e. current setting is '120' and then one click of the encoder makes it '124').

Not weird, as I wrote earlier - turning the encoder from one click/detent to the next results in one quadrature cycle which is four "edges", ie four interrupts so, four counts:



That's why you can't have any delay in the ISR because turning the encoder one "click" results in four interrupts, if you have a delay (Pause or other) in there it misses one or more of the four edges in the cycle and gets confused.

You could setup a timer for the EEPROM save, or you could just set a variable to some value in the ISR, the Main routine then decrements this value and when it hits zero you save the value to EEPROM. That will allow the user to turn the knob and the value be saved some time after the change is complete, psuedo code

TimeToSave VAR WORD
ValueDirty VAR BIT

Main:
   'Code...
   'Code...

   If ValueDirty THEN
     TimeToSave = TimeToSave - 1
     If TimeToSave = 0            ' Time to save ?
       Write.....
       ValueDirty = 0               ' Clear the flag
     ENDIF
   ENDIF

   'Code
   'Code
Goto Main

ISR:
  TimeToSave = 5000
  ValueDirty = 1
  'More code....
@ INT_RETURN

Here the interrupt routine sets a flag indicating that the value has changed, it also sets a "counter" variable which the main routined counts down. If the Main routine executed 5000 times without the value being changed it saves the value to EEPROM. Obviously this relies on a somewhat constant execution time thru the Main routine but you get the idea.

/Henrik.

EDIT: Cross posting.... As for the other "weird" thing. It seems that DT-Ints does NOT handle clearing hte interrupt request flag for the IOC interrupt on these devices so you SHOULD to do that manually before exiting the interrupt (normally NOT needed when the ClearFlag switch is set to Yes). If the interrupt request flag is NOT cleard the interrupt refires over and over and over again which is why the button doesn't seem to work - that code never runs because the CPU spends all the time servicing the interrupt. At least that's what I think is happening, I know it didn't work here if I didn't clear the flags manually.