Rolling timer

Franklin · 2006-06-02 15:44

I'm looking for a way to initiate a response when I get three inputs (on same pin) within five seconds. The problem is the timing has to roll with the inputs such that if the span from input 1 to 3 is six seconds the inputs 2 and 3 are still in play if a fourth input would make three in five seconds. Any thoughts or code examples?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Stephen

allanlane5 · 2006-06-02 15:48

I think you either need a pseudocode description of what you're trying to do, or a flow-chart.

cica · 2006-06-02 16:17

This is quick and dirty, and hasn't been debugged yet, but I think it does what you want.

' {$STAMP BS2}
' {$PBASIC 2.5}




Btn             PIN     0
btnWrk          VAR     Byte
Time1             VAR   Word
Time2             VAR   Word
Time3             VAR   Word
Timer             VAR   Word
duration_needed   VAR   Word

timer=0
duration_needed = 5000 'Play with this number to get 5 seconds

Main:
  ' The BUTTON instruction will cause the program to branch to
  ' No_Press unless P0 = 0

PAUSE 5
BUTTON Btn, 0, 200, 20, btnWrk, 0, No_Press
time3=timer
IF time3 - time1 < duration_needed THEN GOTO SUCCESS
ELSE
time1=time2-timer
time2=time3-timer
timer=time2
ENDIF
GOTO Main



No_Press:
  timer = timer + 1
  GOTO Main


Success:
 DEBUG "Success"
END

-Tom

Post Edited (cica) : 6/2/2006 4:53:03 PM GMT

Franklin · 2006-06-02 16:56

check for input
if input start a timer and increment counter
if counter > 2 and timer < 5 GOTO next thing
(this part I don't know
if timer > 5 delete first count and set timer to second event and continue)

So if the events happen as, event1...4 seconds later event2...2 seconds event3...2 seconds event4
Time between event1 and event3 is more than 5 seconds so no output but time between event2 and event4 is 4 seconds so OUTPUT!
Hope this makes it more clear. I would have included a flowchart but I don't have software for that.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Stephen

Franklin · 2006-06-02 16:58

Thanks cica, I'll look at that closer when I can.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Stephen

Tracy Allen · 2006-06-02 17:37

How this is done wll depend on other aspects of your project.
1) How accurate does it have to be?
2) Do you have an external real time clock or timer that you want to use as the reference, or do you want to do it all in PBASIC?
3) Does the program have to be doing other things while waiting for the three clicks, or can it be modal, just sitting there waiting for the button as its only task?

PB   VAR in1   ' pushbutton state, debounce with RC
time1=254   ' initialize times on long side, bytes
time2=254
time3=254
DO
 DO : timer=timer+1 MAX 254 : PAUSE 100 : LOOP WHILE PB   ' count 0.1 seconds PB=1
  DO : timer=timer+1 MAX 254 : PAUSE 100 : LOOP UNTIL PB  ' count 0.1 seconds PB=0
   ' PB has been released 0-->1, this snippet counts timer from release to release
  time1 = time2
  time2= time3
  time3 = timer
  IF time1+time2+time3 > 60 THEN GOSUB stuff2do  ' 6.0 seconds
 LOOP

That is a modal snipppet devoted to monitoring the button with a granualarity of 0.1 second. You can see the problem if the stuff2do takes a long or indeterminate amount of time. While it is off doing the stuff, it is not timing of the button. If the stuff takes a definite amount of time, you can adjust the timer or the PAUSE to compensate. Similarly, if other stuff has to be done by the program while waiting for the buttons, it can be stuffed in as part of the PAUSE in each timing loop, and the PAUSE adjusted. It all depends on precision required and the other stuff. The situation and programming is diffferent if you have an external reference clock source.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

Franklin · 2006-06-02 17:49

At this point... No it does not need to be very accurate, it's to reduce false triggering of a PIR motion detector. RTC? If there is one on the PDB I have it but I don't think I need the accuracy. It would be nice to do other things until it gets the first pulse and after the routine times out. Thanks for the input.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Stephen

Bruce Bates · 2006-06-02 18:03

Franklin -

Just for the record, many PIR detectors have built-in false alarm detection. Even beyond that, some are dual-sensing, in that they sense BOTH I/R _and_ movement, and won't fire off unless BOTH are present.

Regards,

Bruce Bates

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

cica · 2006-06-02 18:05

Franklin-

I used to install alarm systems for a living. There are PIR detectors that require 3 events in order to trigger. I'd lose the stamp and buy one of these $20 detectors.

-Tom

Rolling timer

Comments